#writing in safe places during Covid Pandemic

[remembertheplunge]

Pandemic and Protest And an Altered State Of Living.June 2 to June 4 2020

June 2, 2020 Tuesday 6:14pm Jail Lobby

Barricades are up near the entry door to the lobby..Like, they are trying to protect from being rammed into.

I just invented a new term: Trump-demic! Inspired by the “oh fuck” Trump pandemic meteor hurtling at earth post card I sent to Zoe recently.

This edition of the Journal is the Protest edition. Protests rage across the country. 

Jared Is not happy that I spent $60 on this journal.

End of entry

Notes : 7/8/2024

I wrote the above entry in the lobby of the Stanislaus County Jail 5 miles West of Modesto. Large cement barricades had been set up to block an attempt to forcibly take the building by Black Lives Matter protesters.

My sister Zoe and I liked to send humorous post cards back and forth to one another. On portrayed a meteorite racing toward earth entitled “oh fuck”. It was the Trump Pandemic meteor!

I had paid $60 for the leather bound journal that I wrote the June 2020 entries in. Jared was my law clerk and business manager and was not happy with the investment.

________________________________________________________

6/3/2020 Wednesday 7:10pm 

NPR: Market Place is on. The Pandemic--Protest is in full swing!

I was up in Stockton at the jail. a fellow defense attorney said police are surrounding the court house. She told me not to come to Stockton tomorrow for court. Too dangerous. She will appear for me.

Meanwhile, Jared said a protest in Oakdale today went violent.

The feeling out here is shaky.

The protests flair here. The virus flairs there. But you never know where or when.

End of entry

Notes 7/8/2024

NPR was National Public radio and Market Place was a show on that station. I listened to Market Place a lot in the early =days of the pandemic. They had honest reporting of how the Pandemic was unfolding.

Oakdale is a town in eastern Stanislaus County, California. 

__________________________________________-

6/4/2020. Thursday 5:20pm Rasputiun’s

Cut. Cut. Gone The two trees marked for destruction are gone. Progress? Productivity? Pandemic and Protest Rage, cutting down 2 trees took priority.

Jerad and I had a beautiful talk with the female clerk at Preservation Coffee House this afternoon. She went to the Sunday 11am protest at 1010 10th (down town Modesto, California) and will go to one in Ripon. She told us that“We need to be willing to be injured to push for change.” .

Magnificent.

Jared said that during the Oakdale protest yesterday, Trump 2020 “all lives matter” stood across from “Black Lives Matter” protesters.

People are out in mobs now. 

I think 10,000 protested yesterday in San Fransisco, Thelma and Louise style, racing for the viral cliff’s edge.

I anticipate a spike in virus and in violence.

Mobile Art Gallery just passed

End of entries

Notes 7/8/2024;

Rasputin’s is a DVD record store located near highway 99 and the rail road tracks in Modesto. During the pandemic, when I could no longer write in coffee houses, I would sit in my car, listen to Mavis Staples songs, write and observe. I got to know the area of the parking lot that I would write in very well. Two young trees I often sat near had been marked to be cut down  with white paint rings around their trunks. On June 4, they were gone and I eulogized them in my entry.  There was another man who would at times park there, too in his hot yellow sports car. He would read his  newspaper. He never looked my way, but, I’m sure that he saw me. Pandamic exiles resorting to a parking lot for covid free reverie.

In 2020 I started noticing Graffiti on the trains as I drove up and down  Highway 99. In March 2020, when the State was in  lock down and the highway electronic signs were screaming out “Stay home and live!”.  I had to be out and drive for court. I never sheltered in place. Besides, I wanted to see the world in its grip of fear. It was fascinating. But, scary , too. And there were the trains. And the Graffiti art work on the train cars. And they were comforting. A message written from before the time of the plague , barreling along as if to say, come follow me . I will lead you to safely out of the virus veil. 

Preservation was Preservation Coffee in Modesto where pre pandemic I spent many hours writing. Post pandemic I have rarely returned and never to write there.

Thelma and Louise was a 1991 movie in which two wild intense women go on a crazy vacation that finds them hurling over  a cliff in the end.

[itsallyscorner]

Maple syrup, coffee, pancakes for two | H.S

pairing: boyfriend!Harry Styles x reader

summary: you unintentionally help Harry with a song he’s been struggling to write

warnings: a lil on the cheesy side tbh, but she’s cute. Mentions the pandemic

a/n: this was one from the drafts, originally written when Harry’s House was released and I finally got around to finishing it :)

Sometime during 2020 and the pandemic

His soft murmurs and humming were the only sounds that can be heard in his makeshift studio. The occasionally creaking of the house or him clicking his pen would break the silence from time to time. His frame was hunched over on the burnt orange colored couch, with his knees almost pressed against his chest, the couch was far too small for his tall figure. However, he liked the couch and it was comfy to lay on when he took breaks from writing. The journal he used to jot down lyrics was balanced on his left knee while his right hand scribbled words on the paper sloppily.

He had just gotten off a zoom call with Kid, Mitch, and some other members who he’s been working with for his new album. While they were supposed to be working in the studio for today’s session, one of the writers had come down with a cold. While it wasn’t confirmed to be COVID, Harry and his team decided it would be safe if everyone just isolated themselves for a while until further notice. Harry actually enjoyed the thought of working on the album at his home. He considered this to be one of the most intimate albums he’s ever made. The artist in him believed that being in the place where he’s comfortable being vulnerable would allow him to write more personal songs. Being home also allowed him to be closer to his muse—you.

The small claw clip holding his fringe from his face began to feel a bit too tight and his eyes were straining against the warm toned light to look at the page on his journal. He had been stuck on a certain verse for a while now and couldn’t bring himself to just call it a day on writing. He had been on the grind when he first started it, but all of a sudden his verses were turning into single words waiting to be properly knitted into a song. So far, some words/phrases he had were:

Wine glass

Puff pass

Side boob ;)

Cocaine

Toothache

Yellow sunglasses

They were the most random and absurd words to be grouped together, yet he knew he was going somewhere with whatever he had. His train of thought came to a halt when your voice rang through the room.

“Hey, you”

His eyes shifted towards your voice and there you were leaning against the door frame. You wore one of his old sweatshirts, which came up to your knees, and some socks on your feet. You weren’t wearing much but your presence and the soft smile gracing your features screamed comfort to him.

The slight frown on Harry’s face turned into a smile that resembled yours. Pushing off of the door frame, you slowly approached Harry’s spot on the couch.

“Hi angel.” He greeted you, spreading his legs out to make space for you. You happily make your way in between his legs and settle on the floor. You crossed your legs and sat so you were looking up at him. Harry craned his neck to place a kiss on your temple, his lips continuing to move down your face to spread little kisses all over your face. When he got to your lips he placed a soft peck on them with a smile on his own pinkish lips.

“Hope you don’t mind me bothering you.” You tease, the sweet smile still on your face. Harry scoffs playfully, “Y’never a bother to me. I missed you today, what’ve y’been up to?”

He’s been in the “studio” since the morning, having a quick breakfast with you and immediately hopping on zoom to work with his team. It was now 6pm and you couldn’t recall seeing or hearing your boyfriend leave the room throughout the day for a break or a snack—which led to you checking in on him.

“Not much, finished up some things for work and caught up on some Love Island.” You shrugged, Harry rolling his eyes jokingly at the mention of the reality tv show. “How ‘bout you? Was today’s session successful?” You ask. Harry hums, reaching for his journal.

“S’half ‘n half. We finished that track we wrote last month—don’t know if y’remember it—but it turned out really great. I have a good feeling ‘bout it once it’s out.” He began while still flipping back to the page he last wrote on. “I started writing another, b’now I’m just stuck. M’brain feels like it can’t think of anything else, s’blank.” He ranted using his hands to express his emotions. You let out a chuckle as you avoid one of his large hands waving around from hitting your face.

“Maybe it’s time for you to take a break, H. You’ve been here all day, god knows how long you’ve been slouched on this couch.” For emphasis, you nudge said couch, Harry shooting you a look.

“S’not a bad couch, leave m’couch alone.” He pouts, silently agreeing that it was definitely time for a break. You duck out of his arms and get up, walking towards the door. Harry follows, moving to get up but halts his movements when a crack comes from his back. You swiftly turn around with your eyes wide and an amused look.

“Oi m’back!” Harry exclaimed in shock. He stood there for second before making eye contact with you. The two of you burst out laughing.

“I fucking told you!” You pointed at him, only to be gently pushed out the door by Harry who was muttering for you to “shuddup”.

Despite the two of you being home, Harry linked his fingers with yours while you both walked to the kitchen. His large hand engulfed yours, his rough thumb stroking the top of your hand.

“I’m actually hungry.” Harry thought aloud once the kitchen came to view. He was wondering what he should eat, deciding between leftovers, cooking, or ordering in. However, he let out a gasp when he saw food already on the table.

He turned around to you beaming, “Y’made my favorite!” On the dining table were pancakes, hash browns, eggs, and coffee. Even though he was eating healthier, Harry believed that one can never go wrong with breakfast for dinner. The main reason why he loved it was because of a memory you both shared during the early stages of your relationship.

The two of you had overslept at his house after movie night and skipped dinner. By the time you both woke up, everything was closed. So the two of you ended up rummaging through his kitchen only to find eggs, pancake mix, frozen hash browns, and coffee. Harry loved that night so much because it was the moment you two truly got to know each other more and connected. It was like finally breaking the barrier of whatever was holding you back from one another. Till this day he remembers the sleepy haze behind your eyes as you shared stories from your past. Your mascara smudged beneath your eyes and your hair was a mess, but none of that mattered because he thought you were beautiful either way. Ever since that night, the two of you would have breakfast for dinner as a staple in your household.

While Harry piled two plates with eggs, pancakes, and hash browns, you filled up two mugs with coffee. Harry liked his black, while you liked yours with a bit of cream and sugar. The two of you settled in the living room, ditching the dining table because it just wasn’t comfy enough. You smiled down at your plate—which had maple syrup dripping down a tower of pancakes—as Harry picked a record to put on. Call it old school, but it was one of the normal things keeping you sane during this lockdown. He had chosen one of his Elvis ones. The same one he played that first night you had breakfast for dinner.

You sat across each other on the couch, feet nuzzled together and your knees bumping alongside the other. The sound of your forks and knives scratching against the plate filled the room along with Elvis’s voice on the record player.

Harry had forgotten he had been writing the entire day. Instead, he remembered all the places you traveled to and the memories you made together on those trips. Being stuck at home for months made the both of you crave the outside world and the normality of it all. Though as much as you wanted to book a trip to Italy, it wasn’t safe to leave the country.

It felt like the world was ending, but to Harry it didn’t feel like it because he was with you. You brought light to the darkness—yes, it was cheesy—but it was the only way Harry knew how to describe being with you.

You guys spent the night eating and reminiscing on past memories you made together. As the hours passed, your plates were now empty and on the coffee counter, while you had found your way into your lover’s arms. His arms held you close to him as your body rested perfectly against his. Your head laid upon his chest, allowing you to feel him breathe and hear the beating of his heart. You were surrounded by his warmth and it was truly all you ever wanted.

Harry could feel you dozing off, your sentences had gotten shorter and your voice had a slight slur. With his nose against your temple he whispered, “Y’know I’ll always love you, right?”

You shifted your head to look into his dark emerald eyes, “Yeah.”

His eyes squinted at you playfully, “How so?” He tested you.

“Because I know I’ll always love you too.”

Then just like that, a spark set off in his head, and all of a sudden the words he has jotted down earlier that day made sense.

[punyhoomans]

Hola! I'm Allan and I need help defending myself in court! If you know me at all you may already be familiar with parts of this story;

At the start of the pandemic, my landlord glitched out and lied to the police to have me forcibly and illegally removed from my home (N.B. that I was the only tenant in the house and this was during the first covid-related lockdown/eviction moratorium)

Homeless in a pandemic and in a city embroiled in BLM protests, I was forced to spend thousands, obliterating my savings to find safe places to stay until I could sign a new lease. Sometimes I slept in my office. It was like camping, only with the threat of unemployment instead of bears.

Though it all my landlord was horrendously antagonistic, threatening me, my property, refusing to allow me to retrieve my stuff, and screaming at the police when I acquiesced to her demands for them to be present.

I was able to hire a lawyer and successfully obtained a protective order and an injunction to stop her from acting on her threats to destroy everything I owned, and forcing her to allow me access to move my things. (Some of which she maliciously damaged beyond repair)

I wish that were the end of it, but it turns out she is also attempting to defame me by levying false criminal charges against me.

Unfortunately, I have been unemployed for a while now and while I can't afford a lawyer, the court appointed attorney says there is essentially no case and no evidence against me, but I still have to appear in person or face arrest for failure to appear.

Tickets from FL to Boston are not cheap, as they have to be refundable due to the inconsistency and unpredictability of court dockets. I've already been rescheduled twice and I'm hoping the third time is the charm to finally end this mess.

This wasn't easy to write. I've asked for a lot of help and almost always got it, and I feel the worse for it every time.

I need to buy a refundable round trip plane ticket and secure lodging for 2-3 days in November and prices will only be going up.

Thank you for reading this. I know that the world is unrelenting in its misery, but your help makes it a little more bearable.

(if gfm doesn't work for you, my other payment methods are here:

https://www.paypal.me/allanc3001

Cash app: $pnyhmns

Ko-fi.com/punyhumans

[covid-safer-hotties]

School’s back and so is a COVID-19 surge: Protecting kids and precarious workers - Published Aug 30, 2024

While the chance of dying has decreased, we should be concerned about the serious long-term complications that can follow an infection.

The 2024 school year is beginning amid one of the biggest COVID-19 waves of the pandemic.

One U.S doctor states, “This is a very significant surge. The levels are very high. They’re the highest we’ve ever seen during a summer wave.” It might be hard to think about, but we’re still in a pandemic and experts are warning against COVID-19 complacency in schools.

Dying with COVID-19 in the acute phase may have decreased, but complications from an infection exist — more than 2 million Canadians have “long COVID” (LC). In this context, societies that see themselves as equitable, inclusive and just need to consider if they’re doing the best job protecting their more vulnerable members, like children and many precarious workers. Research shows governments are not doing the best protecting the rights of children in a crisis, and reports from workers indicate some feel abandoned and left to deal with scary health situations, largely on their own. For school staff, students, their families and communities, this all seems quite cruel. It does not need to be this way.

Organizations are working to make Ontario schools safer. Ontario school safety advocates for cleaner air in schools and research shows schools are safer when things like masking and vaccination are in place. However, these measures are generally not being followed.

Dr. Pantea Javidan writes, “the lack of health and safety in schools resulting from zero-mitigation policies continues to cause great physical and psycho-social harms to children and families.” One study found 14 per cent of adolescents who have gotten COVID-19 developed some LC symptoms. Another study found 45 per cent of infected children with at least one persisting LC symptom. For workers with LC, about 14 per cent have not returned to work within three months since their infection. Many people have been knocked off career paths because of LC. It is a public health crisis for workers and youth.

Allowing uncontrolled spread in communities and schools will likely see the crisis grow, so schools need to be made safe. Not making schools safe violates children’s rights because children’s rights to education include the right to an environment that is safe and not harmful to one’s health.

Kids are major spreaders of COVID-19 and with certain policies, the schools they attend can be safer. HEPA filters should be in every classroom and always on.

Businesses like the Apricot Tree Café in Mississauga are committed to clean air, use HEPA filters and are considered a leader in COVID-19 safe business practices. The owner reports not being sick in four years despite working in a high-risk industry. If a private restaurant can provide clean air, so can the public school system. Not doing so may violate certain human rights.

Dr. Javidan argues “policies threatening the life, health, and education of children are cause for alarm among defenders of human rights.” Importantly, human rights overlap into workers’ rights. In the U.S., workers have been disabled at an unexpectedly high rate since the pandemic began, and researchers are wondering if there’s a COVID-19 connection.

My research looks at precarious occasional teachers (OTs) in Ontario during COVID-19 and ways to make work safer for them, which makes schools safer for students. For example, hiring more secure contract teachers can reduce class sizes making them safer. Higher OT pay, basic income, and paid sick days could help by enabling sick people to stay home. OTs in B.C. make significantly more money per day than Ontario OTs, so higher wages are possible.

Ontario OTs have no paid sick days and can be exposed to multiple schools and hundreds if not thousands of students and staff, which puts them and others at risk. Higher wages and/or a universal basic income allows OTs to have a buffer due to lost income from days absent. A policy of paid sick days can stop sick workers going to work and helps a robust economy.

There are many educational and employment policy options available to reduce COVID-19 levels and protect students and workers during the 2024/25 school year. To uphold the rights of children and workers, these policies should be advocated for by community groups and unions, and adopted by governments and school boards.

[ohgodimafraud]

hey, thank you so much for engaging with sickromancer’s totally weird responses to the “drama”/“discourse” about… uh… vulnerable chronically ill people daring to discuss traumatic experiences during a pandemic! I really liked their writing, but a while back they posted this, so… not shocking that they don’t actually know the difference between Hot Fantasy Contagion and Hot IRL Indifference

https://www.tumblr.com/sickromancer/748033970277580800/as-being-a-contagion-lover-who-is-a-germaphobe-irl?source=share

listen im not into cancel culture or whatever. i think i'd like to clarify my thoughts one more time tho.

if i disagree with something strongly enough ill say it. i think taking care to keep others and yourself safe is important and ik how serious it is now. covid can disable healthy people and with each reinfection that risk increases. thats not me fear mongering, thats ive read this and have had it 3 times and each time has been worse. it's a real gamble bc you dont know how your body will react to it. theres one client i have who is immunocompromised and every time i enter the home i wear a mask and they wear a mask and there are air filters on and genuinely it's the only place ive felt consistently safe in the past year outside of my own home.

we're all shaped by our own experiences and i sympathize w her situation like my family friends grew up with a parent with extreme untreated ocd and theyd be hours late to everything if like the toilet paper touched the wall wrong and theyd have their own special gym equipment and all of the kids turned out different. one of them became like an anti vax incel alt right guy. i had issues w my parents growing up that im working thru and like i def am more prone to anger than the average person. taking your trauma out on other ppl isn’t okay either though

if youre sharing food or drink w friends, both of u have consented and great. partner to partner = consent. not being worried about germs is a privilege though and not being able to understand that is whats like confusing to me esp after 2020. it's not that disabled people have an irrational fear of germs it's like...survival, genuinely. and even if people do have a legit phobia why are we shaming them

also to clarify, i am not disabled unless u count asthma/probable copd (but literally i can do all my tasks so i do not) and have had so much carelessness at work including a horrible absence policy and have loved ones who are disabled. live your best lives, i think if you're sick you should try not to get random people sick on purpose and thats really my main piece idc about consensual situations or any type of fantasy.

[lululawrence]

I usually only read larrie fics but I wouldn't mind trying some zouis fics, do you have any favorites? (I don't know if you even read zouis but you are the only "fic" person I follow that reads other ships so I thought it didn't hurt to ask!)

Oh nonnie yes!!! I LOVE zouis so much!! I’m on mobile, so this list isn’t going to be great, and my reasons for loving each aren’t going to be long because my toddler is a terror at the moment, but here’s a basic list for you to start with!

First, my own because self promo? Haha

Kiss Me Or Not - this was part of my birthday Drabbles series for Harry where I was offering to write Drabbles for anyone who donated to Harry’s birthday drive and I wrote some zouis! It’s short and sweet and was my first attempt hehe I love it though

Hope I Start Talking Crazy (before you understand me) - this one is technically zouiam, but there’s a lot of zouis in there as it’s all building up to the finally ot3 hehe it’s also a/b/o (alpha ZAYN and Liam who are already mated and omega louis), Louis moves in next door to ziam, and it’s just. It was so much fun honestly hehe it was for the zouis fest that ran last year!

Finally, You and I (Collide) - I wrote this as a gift for @so-why-let-your-voice-be-tamed who also loves zouis (and who has also written zouis and does a fab job!!) it’s a five times fic involving adventures in baking that Louis gets to be the taste tester for hehe

Okay and now for everyone else! Here are some of the zouis fics I’ve read most recently and loved:

You Only Fall In Love Twice by @beanno28 - okay so this fic is zourry, but since you usually only read larry fics I wanted to include this because sometimes having your usual pairing as part of it can help ease the way for you haha this was written for the Big Bang this year and is a famous/non famous fic and I loved it so much for so many reasons. It really focuses on all of their relationships and growing into the poly relationship together and it’s just. Wonderful. There’s also a focus on the poly aspect and adjusting to it and the difficulties they might face as they get used to it all that you don’t often see in poly fics, so I hugely loved having that in this one.

we’re still the kings of the Friday nights by @so-why-let-your-voice-be-tamed - this was written for zouis fest as well and ITS SO GOOD OMG!!! Friends to lovers, all the pining, that little hint of angst as they figure things out… it’s just. Everything good in a fic basically haha it’s a lot about self discovery as well and the way Mia writes about emotions is beautiful so of course that means this is impeccably done.

darling just dive right in by @so-why-let-your-voice-be-tamed - I told you she loved zouis, didn’t I? Haha I actually read this one as part of her birthday episode of @podfic-pals as a surprise for her, I love it that much. This is kind of… ex-friends with benefits to lovers? Haha I guess exes to lovers is more succinct and accurate to what they were but it’s so much pain. Soooo much pain. And then all the fluff when they realize their mistake. It’s just. So good. And royalty au!!!

New York Kiss by @quelsentiment - this fic. WHEW. Famous/non famous… kinda. Haha it takes place during Covid with quarantine and the uncertainty of the beginning of the pandemic especially playing a fairly large role in all of it and continuing as a strong undercurrent for their building relationship, so if that’s a trigger for you then stay safe, but I’m telling you this fic is art. The way every piece is put together and expounded upon, the way information is given, the emotions depicted, all of it just hits so perfectly and I love it so much.

Situations Like These by @quelsentiment - okay so I’m not going to list all of this author’s zouis fics even though I want to, but they are honest to goodness one of my very favorite rare pair fic authors ever and they also give incredible aspec representation, so just. Check them out. But this fic was just. It means so much to me I love it very much a lot. The ace rep is FAB but also the way that Louis is just trying so hard to figure out what he’s feeling and what it means and not scare ZAYN off in the meantime and it’s just so very relatable to me! The way it’s all crafted is wonderful, so I defo rec this one along with the previous fic in this list lol

Sigh. I was going to leave you with just two of their fics but I can’t. I am listing three more cause I love them too much to leave off.

Favourite Boy - PAIN. FEELS. FWB (acquaintances with benefits?) to lovers!!!!

Dancing Barefoot - zouiam!!!! On a student field trip as parent supervisors!!!

driver’s license - enemies to lovers ft driving instructor ZAYN who distracts hopeful driver Louis enough that he keeps botching his driving test lmaooo

Unplanned Circumstances by @haztobegood - this one is technically not finished yet… or is it just a really open ending? Lmao it’s SO GOOD THOUGH so read it and shower praise on jinny and maybe she’ll finish it for us someday? Lol

Okay there’s. So many more I could include but I’m cutting myself off because it’s gotten long enough lol but here you go! I hope you enjoy!!! Be sure to leave kudos and nice comments for the authors in thanks hehe

[hemometer-goes-india]

wtf ive seen posts on here like NobOdY wEArs mAsKs oN aIRplAnEs AnyMOre!! and i was like pshhh sensationalist nonsense

but wtf. now i am at the airport and for real, why is nobody wearing a mask? Also the AQI is abysmal right now. Visibly bad. And no one is masking??

Masks are such a good way to keep your lungs safe, and to keep your immune system safe in high risk settings.

If you don’t smoke bc you care about your lungs… why would you start inhaling huge amounts of smoke when the AQI is > 100? The cutoffs for “sensitive groups” are arbitrary. Higher # = more bad.

If you DO smoke, don’t you want to save that good good lung capacity for getting toasted? lol

Idk if my followers know, but I literally founded a mask company during the pandemic. Our device was recommended by top air quality experts and I helped write legislation to help workers pick the best masks during COVID. Our device is ASTM 3502 listed. I know what I’m talking about.

One common misconception about masks is that you need to replace them after every use. This is surprisingly false. As long as it doesnt get wet or dirty the meltblown fabric doesnt degrade significantly.

Here are my mask recs:

Best balance of comfort and performance: Air Pop

This was designed for pollution and illness protection in East Asia. These guys are legit.

Most cost effective + high performance: Fix The Mask

$15 up front cost + $0.30 for each replacement “filter” (you just use a surgical mask). Full disclosure, this is my company, but if you need a discount, email hello@fixthemask.com. I will personally hook you up.

Most typically recommended: 3M Aura

this is what most serious masking people will recommend. It works well on most faces, but for me the rubber over the head straps mess up my hair too much.

3M Aura Particulate Respirator https://a.co/d/8mbNmml

I am super lazy, what is the minimum I can do? Any surgical mask with 90%+ filteation efficiency

A regular surgical mask will give you ~40% inward protection per studies in JAMA and other places (Clapp et. al) and others. It won’t give you 90% protection bc air gets around the sides, but it’s better than nothing.

How big a difference is 40% filtration? This is the best visual I’ve seen to illustrate the difference:

Don’t be a dick, and don’t be an idiot. Use a mask ffs.

[inevitablemoment]

“it’s not as bad as it looks.”

Word Count: 1,026

Warnings: Stab wound, fainting, blood loss, brief mention of secondary infertility and miscarriages, brief mention of premature birth, brief mention of temporary deafness, brief mention of neonatal health issues, brief mention of the COVID-19 pandemic

Fandom: Spider-Man (Raimi films)

Pairings: Peter Parker x Mary Jane Watson

It seems to be a running trend of mine this month to write for new fandoms. So, over the past few weeks, I’ve just fallen down the rabbit hole of anything Spider-Man, and right now, the Raimi trilogy has quite a hold on me. A good percentage of this fic is me dumping my post-trilogy headcanons on you.

Enjoy!

————————

Mary Jane was no stranger to exhaustion. But the last however-many hours had been challenging her ability to keep herself going on minutes of sleep.

It was one of the two days of the week on which her alternate would perform, as well as a parent/teacher conference day for the kids’ school district. She and Peter decided to make it a family day before they would have to go to meet with their teachers.

Then, there was this flash of sparkling gold, a few seconds of darkness... and then another flash of gold.

They were still on Time’s Square, with their first stop planned to have been the Disney Store, but she couldn’t place just why it felt so different from a few seconds ago.

Peter had been the first of them to notice one of the newspaper stands, all of the papers blasted with headlines talking about how Spider-Man was a murderer and a criminal. Before she could roll her eyes, she realized that the man-- or, rather, boy-- on the image was not her husband.

Although Mary Jane and the three older children-- seventeen-year-old twins Maisie and Annie (soon to be eighteen), and thirteen-year-old Benji-- were silently thinking the same thing, three-year-old Susan had blurted out, “That’s not Daddy.”

The children, of course, knew of their father’s secret identity as the web slinging hero-- Mary Jane could only imagine how difficult it would have been for them to keep such a secret from their own children when they resided in the same house-- and had been taught from day one to keep it. Though it was tempting to brag about it to their friends.

“We’re still in New York, just not-- our New York,” Peter had tried to explain.

Even as he tried to explain the concept of a multiverse, most of the scientific jargon went over Mary Jane’s and the children’s heads-- at least, except for Maisie.

Peter immediately went to buy the newspaper bearing the likeness of his other... self? Honestly, Mary Jane didn’t know how else to describe it.

“Whoever the Peter Parker in this universe is, he needs my help,” Peter said. “I... I know he does. And not just about this. Something... something’s gonna happen.”

Mary Jane took his hand. “What do you need me to do?”

“Take the kids and find someplace safe,” he told her. “I’ll call you whenever I can, but if you have to move, tell me where you are--”

“I will,” Mary Jane promised.

“I’ll come with you!” Maisie offered.

“Me, too!” Annie offered. “We can help--!”

“No, you’ll be safer if you stay with your mom,” Peter told them. “Okay?”

Maisie was visibly grinding her teeth in frustration, but Annie-- ever gracious and accommodating, the only one of the twins who never had to be scolded twice for the same misdeed-- said, “I understand, Dad.”

Peter hugged the twins, kissing the tops of their auburn heads, before he addressed Benji.

At thirteen years old, Benji looked every bit his father’s son, but had a solemn and quiet disposition that reminded both of his parents of the man that he was named for.

Benji-- their miracle baby after two miscarriages.

Whom they almost lost during a high risk pregnancy, an emergency C-section in Mary Jane’s twenty-eighth week...

“Benji, I need you to listen to your mom while I’m gone,” Peter instructed. “Can you promise that?”

“Yes, Dad,” Benji promised.

Peter picked little Susan up from the stroller. Their little surprise, conceived and born in the midst of an unprecedented pandemic, but had been welcomed into the family with open arms.

“Daddy, you leaving?” Susan asked, her small voice breaking Peter’s heart.

“Just for a little bit,” he said. “I’m going to help someone, but I swear that I will be back as soon as I can. I need you to be good for Mommy and your sisters and brother, okay?”

“Okay,” Susan agreed.

Annie helpfully took Susan into her arms, allowing her parents to share an embrace

“Whatever’s going on, I promise that we’ll be home soon,” Peter swore to her. “I love you, M.J.”

“I love you, too,” Mary Jane said, kissing him. “Stay safe, Tiger.”

Over the next several hours, Peter kept sending her texts, telling her about what was going on. After Susan continued begging, he answered a video call, looking like he was in some sort of lab. She caught sight of two other men in the background-- the baby-faced boy that she had seen, and the other looking like he was ten years younger than her Peter, both of them wearing their own versions of her husband’s suit. The boy from the pictures looked significantly beat up, making her stomach sink as she wondered what in her Peter’s history had just happened for him.

Almost an hour after the video call, everyone else around them began talking about “Spider-Man livestreaming with the Daily Bugle.”

One man had been gracious enough to show Mary Jane the video. It wasn’t her Peter, though. But as she heard the Peter of this world speak, her heart breaking for him as she heard his voice.

“Mom, is Dad there?” Maisie asked, looking over her mother’s shoulder.

“I... I don’t see him, Mayday,” Mary Jane said.

She couldn’t even recall how much time passed before she saw that flash of gold again. But, in an instant, they were standing in the living room of their townhouse.

Peter-- her Peter-- was standing before her.

Without a second thought, she rushed into his arms. The kids almost turned the hug into a full tackle, but he didn’t seem to mind. Mary Jane’s hands began to run up and down his back before she felt something on her hands. She pulled her hand back, and her stomach dropped.

Blood.

“Peter, what happened?” she asked.

“MJ, it’s not as bad as it looks, I swear,” Peter tried to reassure her.

But, he failed as he began to keel over.

“I’ll call 911,” Annie was the first to jump to the occasion, while Mary Jane put her hand over where she had found the wound.

Please... please be okay...

[honest-moth-of-silver-grove]

I posted 42 times in 2022

That's 4 more posts than 2021!

5 posts created (12%)

37 posts reblogged (88%)

Blogs I reblogged the most:

@elytrians

@muchymozzarella

@hamartia-grander

@demigoddessqueens

@thotty-bog-body

I tagged 42 of my posts in 2022

#not writing - 32 posts

#not mine - 10 posts

#art tag - 6 posts

#lol - 4 posts

#resident evil - 4 posts

#@hermione-grander - 2 posts

#my love! - 2 posts

#i dont think im capable of normal anymore - 1 post

#me w/ my villians - 1 post

#next tumblr april fools pls and thank u - 1 post

Longest Tag: 82 characters

#like i adore all of the adaptatiosn but am also sad that were loosing the og myths

My Top Posts in 2022:

#5

hey... r u alive??

Am,,, I...?

(Lol, jk- yes I am. And I'm working on stuff I promise.)

((On A Totally Unrelated Note: Do NOT stop masking and get covid again because it can literally steal months of your life.))

0 notes - Posted October 11, 2022

#4

not for nothing but i absolutely do see the trend of newcomer tumblr users only liking posts and not reblogging them

0 notes - Posted May 6, 2022

#3

actually have a wip!!!

woo-hoo!!!

6 notes - Posted March 30, 2022

#2

update. ask box still closed.

i survived the fall semester only to get covid.... ugh. omnicron is no joke.

stay safe out there y’all!

-Mothmom 💚

7 notes - Posted January 2, 2022

My #1 post of 2022

Nurse!Reader x Carlos Oliveira & Nikolai Zinoviev Headcanons

A/N: (Everyone’s a touch OOC as I doubt either one would willingly drag a civilian along, albeit for different reasons: Carlos because once you’ve been escorted to safety, you’re no longer his immediate concern; Nikolai because he was never all that interested in saving civilians, to begin with, but let’s say you’re just so darn cute/special that you alter canon circumstances, okay? Okay. Also, all of them survive RE3 in this, so yeah.)

☣     ☢     ☣     ☢

When shit starts hitting the fan, you don't panic immediately. After all, you work in a hospital, what better place to be during a pandemic, right? 

Wrong! Oh so wrong. 

You see the effects of the T-virus up close and personal, and it’s not pretty. Prior to this point, you thought such viruses only existed in science fiction or cheesy B-grade horror movies. But here it was in your workplace, quickly overcoming all quarantined forces, hospital staff, and security. 

So you flee. You leave through a back exit and get the hell out of dodge, or at least, the center of dodge. You soon find, however, that it’s not just the hospital that’s overrun, but the city as well.

Maybe that’s how you meet them…

Carlos:

He swoops in and saves you from a hoard of zombies, looking like an A-list action star as he does it. (It’s the hair- it’s incredible.)

He tells you his platoon is rounding up survivors in the subway, that it’s a temporary shelter until they can get the trains up and running again. That he and his teammates were sent in to get everyone safely out of the city. You’re not sure if you believe this plan, but you follow him eagerly as he leads you to the subway. 

On the way, you tell him who you are. When he finds out you’re medical personnel, he’s both impressed and relieved. He tells you his captain has been injured, and they haven’t come across a doctor or nurse that hasn't already been infected. You of course offer to help in any way your can. (Because you’re awesome like that.)

You’re not surprised when he tells you he’s working with Umbrella because you also work with Umbrella as a hospital employee. But to you, they’re a pharmaceutical company, why would they need military personnel? It doesn't sit right with you, but Carlos is an absolute sweetheart. (I mean the man is cracking jokes in the middle of the apocalypse.) So you trust the guy.

Carlos does his best to assure you you’re not a burden or some sort of unwanted weight when you first arrive at the subway car. Mainly, because he can tell by the way you keep offering to help get the trains running, even though you’d be no match for the horde, but also because of Nikolai’s snide comments about Carlos bringing in “yet another one” when the cars still weren’t working. “Hey, don’t worry about it,” he says. “It’s our job to protect you.”

Carlos takes you to his Captain- Captain Mikhail Victor in charge of Umbrella's Delta Platoon. You try and treat the Captain’s injury the best you can, using what little supplies from the first-aid kit that’s available. You apply a generous amount of first aid spray and wrap a makeshift bandage around the cut. You would have attempted sutures if the kit had any but unfortunately for the Captain, it didn’t. The wound isn’t deep but it’s in a compromising place. You know this man won’t make it out of here on foot. 

Overall, you feel you’ve done an inadequate job because you know the Captain is still incredibly vulnerable, but both Carlos and Mikhail assure you, you’ve been more than helpful. 

But you want to help more! When Carlos admits he and Tyrell won’t be catching the train and will instead stay behind to look for Bard, you offer to come with them, to help them navigate the hospital. 

Of course, everyone thinks it’s a horrible idea. One: because Bard is supposedly at the police station, not the hospital, and Two: because you have no weapons training whatsoever. You’re a walking liability. Nikolai teases Carlos about “taking on the burdens of strays”. You flip him off behind his back. 

After a ton of back and forth, you insist you go with Carlos and Tyrell, refusing to get on the train. At one point you take a seat on the ground of the platform and cross your arms stubbornly. (You’re not going and they can’t make you!)

Carlos is the first to accept the situation and roll with the punches. “Alright. But I have one rule.” He says. “No dying on me.” 

You stand, smile, and shake his hand. “Deal.”

The three of you make it to the police station, where you hang back with Tyrell in the main lobby as Carlos looks around for Bard. Once it’s revealed that Bard is still at the hospital, you offer to escort Carlos there. After giving him your best “I told you so” smirk that is.

On the way you find Jill, clearly having been infected by something, even though you’re not certain what. But none of that matters as you and Carlos bring her to the practically abandoned hospital. 

Once Jill is settled, you give Carlos some directions and a rudimentary drawing of where to locate the asshole Bard’s office. “You’ll need a voice key,” you tell him. “You’d have to look around these rooms for one of his cassettes.” 

He thanks you before asking you to look after Jill for the time being. You promise to radio him if her condition worsens. 

When Carlos comes back with the vaccine, you could practically kiss the man. You don’t, of course, it wouldn't be appropriate. (But the thought does cross your mind very briefly.) Carlos lets you administer the vaccine to Jill, you being the trained nurse and all. With all that's happened, being able to do some actual healing feels like nothing short of a miracle. 

You begin to take notes on Jill’s condition, commenting that although it doesn’t seem to be a speedy cure, her fever’s going down and her skin doesn’t look as clammy. 

The miraculous feeling doesn't last long, however, as Tyrell comes bursting through the room, clearly out of breath and pretty banged-up. He turns on the TV and to your horror, you find you have only hours to make it out of Racoon City unless you want to be vaporized. (Which, no, thank you! You certainly don’t!)  

After catching up, Tyrell being high-key surprised you’re still alive, (which honestly, you’re like ‘same’ lol) you decide to go with Carlos underground, to locate the stockpiled vaccine as a last-ditch effort to save the city. You hope whatever they made, that there’s tons of it. 

Before going underground though, Carlos gives you a gun from one of the killed security guards. It only has a few bullets but he feels safer knowing you’re not just going to walk completely weaponless into whatever danger Umbrella has waiting for you. 

See the full post

187 notes - Posted April 4, 2022

Get your Tumblr 2022 Year in Review →

[atypicalbipolar]

Mental Health Month posts from Facebook (because sharing multi year posts suck)

2023

So is it Mental Health Awareness Month? Or could it be Mental Health Acceptance Month? I guess it's a matter of education. But I was inspired by the fact that April is Autism Acceptance Month. I don't know if the sentiment is the same.

* I am out at work as much as I feel comfortable. I lead an internal DEI group and we introduce ourselves each meeting using roles outside of work. I'm a queer person who uses They/She pronouns and I'm a peer with lived mental health experience. My boss and a couple coworkers know I'm a presenter for In Our Own Voice because I invited them to present for June's DEI meeting. But I am NOT comfortable sharing my diagnosis to work (minus my friend) and I can't present myself sadly.

* About this time last year I volunteered to be a 'tech buddy' DBSA is STILL USING ZOOM. This is good for me because the commute home from Belmont after 9pm would be unbearable. Tech Buddies are the zoom host, and help take attendance and manage the speakers queue with raised hands. Going to DBSA let me know I'm not alone in a way I hadn't experienced before.

* My individual gave up her office early on in COVID and is permanently virtual. My group therapist retired in July and I had to say good bye on zoom. His replacement is private pay, but I'm using FSA funds to justify it. I started seeing my Psychiatrist in person only for him to retire 6 months after group did! I'm currently without a prescriber for the 3 time since 2016. Charlestown MGH is overloaded with patients and understaffed. The one difference this time is that I'm stable.

* I'm still running, but I started stressing about it in October, and I think my COVID bout (which was mild) is still impacting my performance. My photo a day streak is STRONG. I even called into a podcast who was doing an episode on "what works for me" because it really does help me get out the house. That and obsessing over step counts.

I enjoy making mental health content during May, but I feel like I'd just be repeating myself this month. Enough has changed from last year to write out this post. If you're still reading I bet you're very aware and I hope you're even accepting at this point.

2022

Mental Health Awareness Month is here again and I'll be posting for the third year.

In a perfect world I could be out on my public twitter. In a perfect world I could disclose online without any fear that come next job hunt, someone in HR will find that and take a pass on me. In a perfect world I could disclose at work even though I don't need any accommodations. It's not a uniform system but I've cobbled together fb, instagram, and my anonymous bipolar twitter account as safe spaces to share. I'm trying to feel less segmented but it's difficult when real world repercussions are a possibility.

Still zooming with NAMI's In Our Own Voice. They're starting to get in person presentations but it's just easier for me logistically to stay virtual. I've zoomed with plenty of places I'd never be able to get to in person.

This time last year I joined DBSA (Depression Bipolar Support Alliance) Boston, a peer led org that hosts support groups. I'm busy Wednesday nights zooming with the "Young" Adults group. A bunch of millennials and gen z folks using their lived experience as shorthand to bond. It's meant a lot to me and I'm walking with team DBSA for this year's NAMI Walk.

Covid still drags on. Last couple of Mays I wrote that I was glad I had Bipolar. And it's still true. My toolbox of coping skills is flush with things I knew, like taking baths with epsom salt, and things I've learned during the pandemic, like how I developed a running habit. My take a photo a day streak inspires me to get outside. And perhaps most importantly I can still see my prescriber, my individual and my group therapists remotely.

So happy Mental Health Awareness month. You're probably very aware at this point.

2021

I’m taking part in Mental Health Awareness Month again. I am still not out on Twitter*. I still cannot publicly disclose my bipolar, or even just 'mental illness' on an account with my full name for fear of personal and professional repercussions. In these kinds of spaces, it is much easier to navigate out as queer than out as bipolar. *This January I started an alternate anonymous twitter account so I can better engage with the neurodivergent, disabled and mental illness communities. (You'd be surprised at how much those overlap)

Before all of this began, I signed up NAMI's In Our Own Voice. You’ve probably heard about it whenever I mentioned the NAMI Walks I’m doing in a few weeks. We started zooming in August and I want to say it was one of the best decisions I made in terms of advocacy and stigma busting. I have presented to different groups, college nursing classes, one of the inpatient units at McLean Hospital, and several groups of Family to Family, the program mom took 15 years ago when I was initially diagnosed.

After over a year of Covid, I'm still GLAD I’m Bipolar. There's a collective mental health flare happening right now. But I have my diagnosis, meds, my therapy, and coping skills. I have a toolbox of things to try when I am stressed out or sliding backwards. And perhaps most importantly I have established relationships with my individual, group and my prescriber. I am so lucky that I've been able to see them remotely.

So happy Mental Health Awareness month. Now you're probably even more aware.

2020

Mental Health Month Post: I realized something: I am not out on Twitter. I can not publicly disclose my bipolar, or even just 'mental illness' for fear of personal and professional repercussions. I am more out in my queerness than in my mental health struggles. How ironic is that? I wish things were different.

When I signed up NAMI's In Our Own Voice, I was hoping to use it as a tool for greater advocacy but also to increase my "outness" as a bipolar person. Even though the training was Presidents Day Weekend, I never thought that a pandemic would put all presentations on hold.

Sure I participate in chronic illness (disability) twitter and will 'flirt' with outing myself. If you read behind the lines, check who I follow and talk to, you may be able to out me. But I talk about my crohns and humira, not lithium, and I leave therapy to my journal on patientslikeme.

But at a time like this I'm actually GLAD I am Bipolar. There's going to be a mental health flare when Covid starts to wind down. But I already have my diagnosis, meds, my therapy, and coping skills. I already have a toolbox of things to try when I am stressed out or sliding backwards. And perhaps most importantly I already have established relationships with my individual, group and my prescriber.

So happy Mental Health Awareness month. Now you're probably more aware.

[coreytravelogue]

Georgia Strait, BC - March 28, 2024

I would love to or have preferred to be going somewhere other than Victoria this extra long weekend but Victoria was the choice. Not because I dislike Victoria, anyone who knows me knows I love Victoria with all my heart. It is just that it is not exciting or exotic enough for the 4 days off.

Last year I went to Toronto which was packed with adventure and exploration, the year before was Edmonton which is not exactly a new place but during the pandemic it seemed like a safe trip. Year before that if I remember was Victoria as well but that was deeper in the pandemic.

There are multiple reasons why I am choosing this safe trip versus other places.

First and most importantly it is because of cost. Whitehorse, Edmonton, Winnipeg, Saskatoon and Fort McMurray were all too expensive or too pricey for not having much to do.

Edmonton would have been a close choice but I decided not to because sure I could flight “cheap” the hotel would not have been plus the only reason I would go would be to see my friend and cousin but cousin would be maybe one evening and I do not know if I could have a day with my friend much less a morning. A lot to hope to travel all the way over there just to go to the water park otherwise. Winnipeg was the next closest but few things to Jayne hat not counting not much I feel one could do in April outside of a Jets game. Every other option just didn’t seem worth it for the large price tag.

For the price of one night and a half’s hotel I can stay for 3 nights at Ocean Island Inn, ferry ride is 42 bucks round trip which is infinitely lower than the airfare.

Victoria has no water park and probably won’t have any hockey games being played this weekend but it will have John’s Place and The Drake which I always enjoy. It will have record shops I like to frequent, a chill atmosphere and an arcade. It also has a friend and her family around her birthday so I would at least have company if they chose to have me.

So it is an easy trip to have, I just need to hope on a ferry and head down.

Also I think I need a weekend where I am simply not in a hurry or need to explore though there are parts of Victoria I can easily explore if I wanted to. I need a weekend to mentally recharge from the last few months of work shit that will probably only escalate when I get back.

I am coming into this trip with only three objectives; write something, treat myself and relax. And if there is one place in the world I am assured to have that it would be Victoria.

I have to be cost effective anyway if I want to to budget properly for Japan.

There is just so much uncertainty in my life right now but then you can say that about the world right now but thinking about it when is there uncertainty. We never know if we were on solid ground till it is gone. I came to realize that lately when thinking of my last workplace. I felt so stressed and unhinged back then but in comparison to now I feel that those were the more stable days.

It is more with the future of work for me and future of what is next if certain things don’t happen. With my 40s looming and failing to travel during that last half of my 30s makes me want to do another drop everything trip more and more though it would be far more risky now. It can’t be blamed, COVID-19 killed fun for everyone for many years but with it it made everything more dangerous and unaffordable.

I am not happy with how things are going and I feel like I am back in the same rut I was in 10 years ago despite all my attempts to change the script. I did nearly everything right and I feel like I am still fucking up. I guess that reminds me of a scene in Star Trek TNG where Picard said you could do everything right and still get it wrong, that is life.

Can only hope to have more clarity and assurance in May, if not I will have to force it into my life in a drastic way.

This long weekend I will start each day with a damn good breakfast, theatres during the afternoon, arcades in the evening and The Drake during the night writing. Rinse repeat or interchange with being around my friend and her family or other cheap things. Anything but work or thinking about work at all costs. That is the mission.

Shazbot nanu nanu

[jcmarchi]

Why Doesn’t My Model Work?

New Post has been published on https://thedigitalinsider.com/why-doesnt-my-model-work/

Why Doesn’t My Model Work?

Have you ever trained a model you thought was good, but then it failed miserably when applied to real world data? If so, you’re in good company. Machine learning processes are complex, and it’s very easy to do things that will cause overfitting without it being obvious. In the 20 years or so that I’ve been working in machine learning, I’ve seen many examples of this, prompting me to write “How to avoid machine learning pitfalls: a guide for academic researchers” in an attempt to prevent other people from falling into these traps.

But you don’t have to take my word for it. These issues are being increasingly reported in both the scientific and popular press. Examples include the observation that hundreds of models developed during the Covid pandemic simply don’t work, and that a water quality system deployed in Toronto regularly told people it was safe to bathe in dangerous water. Many of these are documented in the AIAAIC repository. It’s even been suggested that these machine learning missteps are causing a reproducibility crisis in science — and, given that many scientists use machine learning as a key tool these days, a lack of trust in published scientific results.

In this article, I’m going to talk about some of the issues that can cause a model to seem good when it isn’t. I’ll also talk about some of the ways in which these kinds of mistakes can be prevented, including the use of the recently-introduced REFORMS checklist for doing ML-based science.

Duped by Data

Misleading data is a good place to start, or rather not a good place to start, since the whole machine learning process rests upon the data that’s used to train and test the model.

In the worst cases, misleading data can cause the phenomenon known as garbage in garbage out; that is, you can train a model, and potentially get very good performance on the test set, but the model has no real world utility. Examples of this can be found in the aforementioned review of Covid prediction models by Roberts et al. In the rush to develop tools for Covid prediction, a number of public datasets became available, but these were later found to contain misleading signals — such as overlapping records, mislabellings and hidden variables — all of which helped models to accurately predict the class labels without learning anything useful in the process.

Take hidden variables. These are features that are present in data, and which happen to be predictive of class labels within the data, but which are not directly related to them. If your model latches on to these during training, it will appear to work well, but may not work on new data. For example, in many Covid chest imaging datasets, the orientation of the body is a hidden variable: people who were sick were more likely to have been scanned lying down, whereas those who were standing tended to be healthy. Because they learnt this hidden variable, rather than the true features of the disease, many Covid machine learning models turned out to be good at predicting posture, but bad at predicting Covid. Despite their name, these hidden variables are often in plain sight, and there have been many examples of classifiers latching onto boundary markers, watermarks and timestamps embedded in images, which often serve to distinguish one class from another without having to look at the actual data.

A related issue is the presence of spurious correlations. Unlike hidden variables, these have no true relationship to anything else in the data; they’re just patterns that happen to correlate with the class labels. A classic example is the tank problem, where the US military allegedly tried to train a neural network to identify tanks, but it actually recognised the weather, since all the pictures of tanks were taken at the same time of day. Consider the images below: a machine learning model could recognise all the pictures of tanks in this dataset just by looking at the colour of pixels towards the top of an image, without having to consider the shape of any of the objects. The performance of the model would appear great, but it would be completely useless in practice.

(Source: by author)

Many (perhaps most) datasets contain spurious correlations, but they’re not usually as obvious as this one. Common computer vision benchmarks, for example, are known to have groups of background pixels that are spuriously correlated with class labels. This represents a particular challenge to deep learners, which have the capacity to model many patterns within the data; various studies have shown that they do tend to capture spuriously correlated patterns, and this reduces their generality. Sensitivity to adversarial attacks is one consequence of this: if a deep learning model bases its prediction on spurious correlations in the background pixels of an image, then making small changes to these pixels can flip the prediction of the model. Adversarial training, where a model is exposed to adversarial samples during training, can be used to address this, but it’s expensive. An easier approach is just to look at your model, and see what information it’s using to make its decisions. For instance, if a saliency map produced by an explainable AI technique suggests that your model is focusing on something in the background, then it’s probably not going to generalise well.

Sometimes it’s not the data itself that is problematic, but rather the labelling of the data. This is especially the case when data is labelled by humans, and the labels end up capturing biases, misassumptions or just plain old mistakes made by the labellers. Examples of this can be seen in datasets used as image classification benchmarks, such as MNIST and CIFAR, which typically have a mislabelling rate of a couple of percent — not a huge amount, but pretty significant where modellers are fighting over accuracies in the tenths of a percent. That is, if your model does slightly better than the competition, is it due to an actual improvement, or due to modelling noise in the labelling process? Things can be even more troublesome when working with data that has implicit subjectivity, such as sentiment classification, where there’s a danger of overfitting particular labellers.

Led by Leaks

Bad data isn’t the only problem. There’s plenty of scope for mistakes further down the machine learning pipeline. A common one is data leakage. This happens when the model training pipeline has access to information it shouldn’t have access to, particularly information that confers an advantage to the model. Most of the time, this manifests as information leaks from the test data — and whilst most people know that test data should be kept independent and not explicitly used during training, there are various subtle ways that information can leak out.

One example is performing a data-dependent preprocessing operation on an entire dataset, before splitting off the test data. That is, making changes to all the data using information that was learnt by looking at all the data. Such operations vary from the simple, such as centering and scaling numerical features, to the complex, such as feature selection, dimensionality reduction and data augmentation — but they all have in common the fact that they use knowledge of the whole dataset to guide their outcome. This means that knowledge of the test data is implicitly entering the model training pipeline, even if it is not explicitly used to train the model. As a consequence, any measure of performance derived from the test set is likely to be an overestimate of the model’s true performance.

Let’s consider the simplest example: centering and scaling. This involves looking at the range of each feature, and then using this information to rescale all the values, typically so that the mean is 0 and the standard deviation is 1. If this is done on the whole dataset before splitting off the test data, then the scaling of the training data will include information about the range and distribution of the feature values in the test set. This is particularly problematic if the range of the test set is broader than the training set, since the model could potentially infer this fact from the truncated range of values present in the training data, and do well on the test set just by predicting values higher or lower than those which were seen during training. For instance, if you’re working on stock price forecasting from time series data with a model that takes inputs in the range 0 to 1 but it only sees values in the range 0 to 0.5 during training, then it’s not too hard for it to infer that stock prices will go up in the future.

In fact, forecasting is an area of machine learning that is particularly susceptible to data leaks, due to something called look ahead bias. This occurs when information the model shouldn’t have access to leaks from the future and artificially improves its performance on the test set. This commonly happens when the training set contains samples that are further ahead in time than the test set. I’ll give an example later of when this can happen, but if you work in this area, I’d also strongly recommend taking a look at this excellent review of pitfalls and best practices in evaluating time series forecasting models.

An example of a more complex data-dependent preprocessing operation leading to overly-optimistic performance metrics can be found in this review of pre-term birth prediction models. Basically, a host of papers reported high accuracies at predicting whether a baby would be born early, but it turned out that all had applied data augmentation to the data set before splitting off the test data. This resulted in the test set containing augmented samples of training data, and the training set containing augmented samples of test data — which amounted to a pretty significant data leak. When the authors of the review corrected this, the predictive performance of the models dropped from being near perfect to not much better than random.

(Source: https://arxiv.org/abs/2108.02497)

Oddly, one of the most common examples of data leakage doesn’t have an agreed name (the terms overhyping and sequential overfitting have been suggested) but is essentially a form of training to the test set. By way of example, imagine the scenario depicted above where you’ve trained a model and evaluated it on the test set. You then decided its performance was below where you wanted it to be. So, you tweaked the model, and then you reevaluated it. You still weren’t happy, so you kept on doing this until its performance on the test set was good enough. Sounds familiar? Well, this is a common thing to do, but if you’re developing a model iteratively and using the same test set to evaluate the model after each iteration, then you’re basically using that test set to guide the development of the model. The end result is that you’ll overfit the test set and probably get an over-optimistic measure of how well your model generalises.

Interestingly, the same process occurs when people use community benchmarks, such as MNIST, CIFAR and ImageNet. Almost everyone who works on image classification uses these data sets to benchmark their approaches; so, over time, it’s inevitable that some overfitting of these benchmarks will occur. To mitigate against this, it’s always advisable to use a diverse selection of benchmarks, and ideally try your technique on a data set which other people haven’t used.

Misinformed by Metrics

Once you’ve built your model robustly, you then have to evaluate it robustly. There’s plenty that can go wrong here too. Let’s start with an inappropriate choice of metrics. The classic example is using accuracy with an imbalanced dataset. Imagine that you’ve managed to train a model that always predicts the same label, regardless of its input. If half of the test samples have this label as their ground truth, then you’ll get an accuracy of 50% — which is fine, a bad accuracy for a bad classifier. If 90% of the test samples have this label, then you’ll get an accuracy of 90% — a good accuracy for a bad classifier. This level of imbalance is not uncommon in real world data sets, and when working with imbalanced training sets, it’s not uncommon to get classifiers that always predict the majority label. In this case, it would be much better to use a metric like F score or Matthews correlation coefficient, since these are less sensitive to class imbalances. However, all metrics have their weaknesses, so it’s always best to use a portfolio of metrics that give different perspectives on a model’s performance and failure modes.

Metrics for time series forecasting are particularly troublesome. There are a lot of them to choose from, and the most appropriate choice can depend on both the specific problem domain and the exact nature of the time series data. Unlike metrics used for classification, many of the regression metrics used in time series forecasting have no natural scale, meaning that raw numbers can be misleading. For instance, the interpretation of mean squared errors depends on the range of values present in the time series. For this reason, it’s important to use appropriate baselines in addition to appropriate metrics. As an example, this (already mentioned) review of time series forecasting pitfalls demonstrates how many of the deep learning models published at top AI venues are actually less good than naive baseline models. For instance, they show that an autoformer, a kind of complex transformer model designed for time series forecasting, can be beaten by a trivial model that predicts no change at the next time step — something that isn’t apparent from looking at metrics alone.

In general, there is a trend towards developing increasingly complex models to solve difficult problems. However, it’s important to bear in mind that some problems may not be solvable, regardless of how complex the model becomes. This is probably the case for many financial time series forecasting problems. It’s also the case when predicting certain natural phenomena, particularly those in which a chaotic component precludes prediction beyond a certain time horizon. For instance, many people think that earthquakes can not be predicted, yet there are a host of papers reporting good performance on this task. This review paper discusses how these correct predictions may be due to a raft of modelling pitfalls, including inappropriate choice of baselines and overfitting due to data sparsity, unnecessary complexity and data leaks.

Another problem is assuming that a single evaluation is sufficient to measure the performance of a model. Sometimes it is, but a lot of the time you’ll be working with models that are stochastic or unstable; so, each time you train them, you get different results. Or you may be working with a small data set where you might just get lucky with an easy test split. To address both situations, it is commonplace to use resampling methods like cross-validation, which train and test a model on different subsets of the data and then work out the average performance. However, resampling introduces its own risks. One of these is the increased risk of data leaks, particularly when assuming that data-dependent preprocessing operations (like centering and scaling and feature selection) only need to be done once. They don’t; they need to be done independently for each iteration of the resampling process, and to do otherwise can cause a data leak. Below is an example of this, showing how feature selection should be done independently on the two training sets (in blue) used in the first two iterations of cross-validation, and how this results in different features being selected each time.

(Source: https://arxiv.org/abs/2108.02497)

As I mentioned earlier, the danger of data leaks is even greater when working with time series data. Using standard cross-validation, every iteration except one will involve using at least one training fold that is further ahead in time than the data in the test fold. For example, if you imagine that the data rows in the figure above represent time-ordered multivariate samples, then the test sets (in pink) used in both iterations occur earlier in the time series than all or part of the training data. This is an example of a look ahead bias. Alternative approaches, such as blocked cross-validation, can be used to prevent these.

Multiple evaluations aren’t an option for everyone. For example, training a foundation model is both time-consuming and expensive, so doing it repeatedly is not feasible. Depending on your resources, this may be the case for even relatively small deep learning models. If so, then also consider using other methods for measuring the robustness of models. This includes things like using explainability analysis, performing ablation studies, or augmenting test data. These can allow you to look beyond potentially-misleading metrics and gain some appreciation of how a model works and how it might fail, which in turn can help you decide whether to use it in practice.

Falling Deeper

So far, I’ve mostly talked about general machine learning processes, but the pitfalls can be even greater when using deep learning models. Consider the use of latent space models. These are often trained separately to the predictive models that use them. That is, it’s not unusual to train something like an autoencoder to do feature extraction, and then use the output of this model within the training of a downstream model. When doing this, it’s essential to ensure that the test set used in the downstream model does not intersect with the training data used in the autoencoder — something that can easily happen when using cross-validation or other resampling methods, e.g. when using different random splits or not selecting models trained on the same training folds.

However, as deep learning models get larger and more complex, it can be harder to ensure these kinds of data leaks do not occur. For instance, if you use a pre-trained foundation model, it may not be possible to tell whether the data used in your test set was used to train the foundation model — particularly if you’re using benchmark data from the internet to test your model. Things get even worse if you’re using composite models. For example, if you’re using a BERT-type foundation model to encode the inputs when fine-tuning a GPT-type foundation model, you have to take into account any intersection between the datasets used to train the two foundation models in addition to your own fine-tuning data. In practice, some of these data sets may be unknown, meaning that you can’t be confident whether your model is correctly generalising or merely reproducing data memorised during pre-training.

Avoiding the Pits

These pitfalls are all too common. So, what’s the best way to avoid them? Well, one thing you can do is use a checklist, which is basically a formal document that takes you through the key pain points in the machine learning pipeline, and helps you to identify potential issues. In domains with high-stakes decisions, such as medicine, there are already a number of well-established checklists, such as CLAIM, and adherence to these is typically enforced by journals that publish in these areas.

However, I’d like to briefly introduce a new kid on the block: REFORMS, a consensus-based checklist for doing machine learning-based science. This was put together by 19 researchers across computer science, data science, mathematics, social sciences, and the biomedical sciences — including myself — and came out of a recent workshop on the reproducibility crisis in ML‑based science. It is intended to address the common mistakes that occur in the machine learning pipeline, including many of those mentioned in this article, in a more domain-independent manner. It consists of two parts: the checklist itself, and also a paired guidance document, which explains why each of the checklist items are important. The checklist works through the main components of a machine learning-based study, in each case encouraging the user to verify that the machine learning process is designed in such a way that it supports the overall aims of the study, doesn’t stumble into any of the common pitfalls, and enables the results to be verified by an independent researcher. Whilst it’s focused on the application of machine learning within a scientific context, a lot of what it covers is more generally applicable, so I’d encourage you to take a look even if you don’t consider your work to be “science”.

Another way of avoiding pitfalls is to make better use of tools. Now, one of my pet gripes regarding the current state of machine learning is that commonly-used tools do little to prevent you from making mistakes. That is, they’ll happily let you abuse the machine learning process in all sorts of ways without telling you what you’re doing is wrong. Nevertheless, help is available in the form of experiment tracking frameworks, which automatically keep a record of the models you trained and how you trained them, and this can be useful for spotting things like data leaks and training to the test set. An open source option is MLFlow, but there are plenty of commercial offerings. MLOps tools take this even further, and help to manage all the moving parts in a machine learning workflow, including the people.

Final Thought

It is possible to train a good model that generalises well to unseen data, but I wouldn’t believe this until you’re satisfied that nothing which could have gone wrong has gone wrong. A healthy sense of suspicion is a good thing: do look at your trained model to make sure it’s doing something sensible, do analyse your metrics to understand where it’s making mistakes, do calibrate your results against appropriate baselines, and do consider using checklists to make sure you haven’t overlooked something important.

Author Bio

Michael is an Associate Professor at Heriot-Watt University, Edinburgh. He’s spent the last 20 years or so doing research on machine learning and bio-inspired computing. For more info see his academic website. He also writes about computer science more generally in his Fetch Decode Execute substack.

Citation

For attribution in academic contexts or books, please cite this work as

Michael Lones, "Why Doesn’t My Model Work?", The Gradient, 2024.

BibTeX citation:

@article{lones2024why, author = Michael Lones, title = Why Doesn’t My Model Work?, journal = The Gradient, year = 2024, howpublished = urlhttps://thegradient.pub/why-doesnt-my-model-work,

[reasoningdaily]

Nostradamus was born Michel de Notredame in 1503 in Saint-Remy-de-Provence, France. It's fair to say Nostradamus didn't have a boring life. He originally worked as an apothecary, then attended the University of Montpelier to pursue a medical degree. Along the way, he was expelled for practicing a "manual trade" (the use of herbal remedies), which was considered "too low" for a doctor (via Biography).

The jury is out as to whether he eventually went back to school and received a medical license, but Biography points out he did work as a physician during the plague outbreak in Italy and France. In fact, he became a bit of a "celebrity" plague doctor for his progressive methods (for the time) in treating the plague, which included herbal mixes, hygiene, and removing corpses from cities to slow down the spread of the disease.

He eventually focused on writing almanacs and became an astrologer. His work was so popular that powerful figures of the time hired him to "read the stars" for them. One example was the Italian noblewoman Catherine de' Medici, who wanted to have horoscopes written for her children. Eventually she made him Counselor and Physician-in-Ordinary to her husband, King Henri II of France

Above all, however, Nostradamus is best known for his 942 poetic quatrains (four-line poems), which are supposed to predict a number of future events — some of them well into the future, according to Biography.

Some of Nostradamus' prophecies have come true ... maybe

Fototeca Storica Nazionale./Getty Images

Because Nostradamus' prophecies are written in poem form, they can be very vague. In fact, you could use the same exact poem to "predict" several incidents throughout history, some of them centuries apart. There have been instances, however, where some of his predictions seem very specific. For example:

"The young lion will overcome the older one / On the field of combat in a single battle / He will pierce his eyes through a golden cage / Two wounds made one, then he dies a cruel death" has been interpreted as a prediction of the death of King Henry II of France, who died in a jousting tournament when a younger opponent hit his face and caused a massive wound. The king died after 11 days of extreme pain (per Insider).

Another famous prophecy that appeared to come true referred to the devastating London fire of 1666: "The blood of the just will be demanded of London / Burnt by fire in the year '66 / The ancient Lady will fall from her high place / And many of the same sect will be killed." However, the poem seems to indicate the fire was caused by lightning, when in reality it started inside a bakery (via Insider).

Some of his prophecies were definitely wrong, though. For example, he predicted that in "The year one thousand nine ninety-nine seven month /  From the sky shall come a great King of terror." But nothing particularly terrible happened in September 1999 (per Biography).

The present year doesn't look good for humanity

Shutterstock

Nostradamus had a lot of terrible things to say about 2021. For this year alone, he predicted zombies, asteroids, and a plague, all with grim results. The predictions start pretty dark: "After great trouble for humanity, a greater one is prepared, / The Great Mover renews the ages: / Rain, blood, milk, famine, steel, and plague, / Is the heavens fire seen, a long spark running" (via 7News).

Experts believe the "great trouble" could refer to the COVID pandemic that started in 2020, with "a greater one" meaning the continuation of the virus into 2021. As for the "heaven's fire," turns out the massive Asteroid 2021 KT1 flew by Earth at 40,000 mph in early June, according to NASA's Center for Near Earth Object Studies (via USA Today). But since it missed and Earth is still intact, we can safely say at least this part of the prediction has failed so far.

For 2021, Nostradamus also warns us of "Few young people: half-dead to give a start," which seems to allude to some sort of zombie apocalypse (the half-dead) taking over the world. While we don't have (so far) any information of zombies taking over the planet, the CDC (Centers for Disease Control and Prevention) is helping us get ready for it with their Zombie Preparedness Guide, just in case.

If we survive 2021, we'll be okay for a while

Shutterstock

Despite the grim predictions for 2021, Nostradamus' prophecies continue until the year 3797. So if we make it through potential asteroid crashes, famines, the plague, and a zombie apocalypse, humanity might survive an extra 1,500 years or so. Still, between 2021 and the end of times, Nostradamus predicted a lot of terrible things: wars, the coming of three Antichrists, and many plagues and pestilences, as related by The Sunday Post of Glasgow.

Many of these will leave the world in worse shape than it was before them, but not necessarily end it. At some point, Nostradamus predicted the third Antichrist will "rise over all the kings of the east" and trigger World War III. This won't be pretty and will last up to three decades, resulting in a lot of death and destruction. However, this won't be the end, either. 

According to The Sun, the true end is what Nostradamus called the "final conflagration," a series of extinction level events that might include super volcanoes, "fire from the sky," and what some believe could refer to a massive solar flare. While he doesn't give a date for the end of the world, the fact that his predictions end at a specific time (the year 3797) could be a sign that he believes this is when the end will come.

[techassistanthub]

"Navigating the Real Estate Market During a Pandemic"

The COVID-19 pandemic has disrupted nearly every industry, and real estate is no exception. With social distancing measures in place, many traditional real estate practices have had to adapt to the new normal. Online virtual assistants (OVAs) have become an invaluable resource for navigating the real estate market during this challenging time. In this article, we'll explore how OVAs can help you successfully buy or sell a property, even during a pandemic.

Conducting Virtual Property Tours

One of the biggest challenges facing the real estate industry during the pandemic has been how to conduct property tours while minimizing contact between people. OVAs can help by conducting virtual property tours on behalf of clients. Using video chat technology, OVAs can walk clients through properties in real-time, answering questions and providing a detailed look at the property's features.

Streamlining the Transaction Process

The pandemic has also made it more difficult to complete real estate transactions in person. OVAs can help by providing remote support for tasks such as contract review, e-signatures, and document management. This can help expedite the transaction process and reduce the need for face-to-face meetings.

Providing Market Research and Analysis

Managing Remote Property Listings

With more people working remotely, there has been a shift towards online property listings. OVAs can help manage these listings, ensuring that properties are accurately represented and marketed to potential buyers. This includes creating high-quality photos and videos, writing compelling property descriptions, and managing listings across multiple online platforms.

In conclusion, the COVID-19 pandemic has presented significant challenges for the real estate industry. However, with the help of OVAs, it is still possible to successfully navigate the market and achieve your real estate goals. By conducting virtual property tours, streamlining the transaction process, providing market research and analysis, and managing remote property listings, OVAs can help you buy or sell a property safely and efficiently.

If you're looking for an online virtual assistant to help you with your real estate needs, please contact us. Our team of experienced OVAs can provide the support you need to succeed in today's challenging real estate market.

[digitalniran8]

How E-Prescribing Software Helps Optometrist?

Introduction

The COVID-19 pandemic has increased the usage of digital devices and continuous virtual meetings make a person develop eye irritation and other diseases. How E-Prescribing Software Helps Optometrist? Internet is a boon as it helps to make a transaction over the air and simultaneously, virtual meetings are scheduled at different time horizons. Work from home and online classes have added more reasons to use a digital screen for more than 6 to 7 hours a day additionally.

E-Prescribing Software is a web-based tool that helps medical practitioners to write prescription dispensing organizations. These files are stored and accessed electronically to make a clinical decision. To make any transaction in this software, an individual secure login is required. In case of a data breach, makes that person faces severe legal actions.

Optical Eyewear Prescription in COVID-19

Increased digital screen usage made people have the optical lens to avoid radiation emitted from those screens. An optical eyewear prescription is written by an eyewear prescriber with specific values supposed to give clear vision to a patient without any serious clinical issue.

Optometrists have created new practices for the patient during this pandemic to maintain eye condition.  By following the guidance of the Center for Disease Control and Prevention (CDC), doctors used E-Prescribing Software to ensure prescriptions to clinics and eyewear shops are sent directly.

Using contact lenses is safe and effective for people who practice good hygiene when handling lenses. A virus can be transmitted from contaminated surfaces and objects so American Optometric Association suggested using e-prescribing software and e-commerce service to place orders.

Ophthalmology Software Management

E-Prescribing Software supports Ophthalmology services to record all the data of patients electronically from appointment scheduling to billing. Reading from ophthalmology software is sent live to prescribing space and makes legible understanding.

Consolidated images and diagnosed reports of a patient are sent to the patient unique healthcare ID through Ophthalmic Image Management. Users can create customized forms to speed up the workflow. Options are given to trigger diagnostic and procedure codes automatically based on a patient database.

Benefits

For medical practitioners, E-Prescribing Software gives many benefits

The reports are transmitted instantaneously with no involvement of physical activities

An efficient practice to stop falsify readings and to reduce paper usage.  

Workflow customization options with user-friendly tools that help doctors to access patients' previous eye records from EHR/EMR software.

Unified Cloud-Based System optimized to run on all types of web-browser and allows authorized access anytime, from any device.

Conclusion

Many eye care professionals are adopting towards E-Prescribing Software as it supports the transformation of treatment quality.  The software helps to increase productivity and documentation speed. A lot of information regarding cataracts and general eye care tips are provided in the knowledge space.

[hollymerideth-blog]

Back?

Is tumblr still a thing? 

Honestly I don't care at this point. I started back in college to be a “tumblr girl” and write long post about my life. Abviously I did not follow through with it. Everyone should be thankful I didn’t because that would just be mellow dramatic teenage spiraling. 

College kind of sucked but like it didnt at the same time. It was hard, i met new friends, and most importantly I graduated. I feel like I learned a lot about myself and pushed myself out of my comfort zone. Mostly realized I was uncomfortable with a lot of things and needed medicine. LMAO. Honestly it is what I really needed to have a long time ago. 

OH I FORGOT there was a whole ass pandemic?? also met my husband... well fiancé. 

Anyways not the point. 

Im 25. yep 25. I have having a hard time with it. Actually I am having a hard time wrapping my head around a lot of things. I think that is way I am wanting to start this shit again. I think I just want a place to put my thoughts out into the unknown and just let them live. I am having a hard time not feeling herd in my relationships with people and I think this would be a safe place to keep them. Maybe? 

Anyways, I am going to leave it here with an update on life and maybe my goals for this and where id like it to go? 

Update: 

Met a dude 

Went through a whole ass pandemic 

Graduated College with a video in my living room

Applied to one million jobs finally got an interview got a call three hours later offering me the job

Got engaged 

Became a teacher during a pandemic

Moved into my first home 

Bought like real adult things (paint/appliances/toilet??)

Grew a lot and loved my first group of kids

Got on the district curriculum team. 

My grandparents both got covid and died the same weekend of each other

Got our baby puppy Patty

Taught at a summer camp 

Had the worst year teaching ever 

Had group of kids/team to work with 

Had the best support teachers/made a life long friend 

Taught a summer camp again 

Became a mentor teacher 

Have the best group of kids this year

Going through a quarter life crisis..?

Got asked to be Maid of Honor 

Planning a wedding maybe? 

Applying to Grad School 

Teaching? 

-Honestly there has been so much that has happened in the last three years that are really big things and I don't think I have fully processed them. I don't really know how to. Im hoping that writing out my thoughts will help? 

Goals: 

Move schools 

Grad School 

Wedding 

Maid of Honor 

Build House 

Baby 

Not nessicarly the goals for the year but you know. 

Maybe I will keep up with this maybe not. Its Christmas Break I have time on my hands I hate not being busy so I'm just bored really. 

Until next time.