#Machine learning diagnostics

[phdwritingassistance]

Artificial Intelligence is revolutionizing healthcare by enabling early diagnosis of chronic diseases. 🤖🩺 From analyzing complex data to predicting risks, AI empowers medical professionals to save lives and improve patient outcomes. Discover how this technology is transforming the future of health 🌍✨.

[innonurse]

AI diagnoses coeliac disease with high accuracy

- By InnoNurse Staff -

Researchers at the University of Cambridge have developed a machine learning algorithm that can diagnose coeliac disease with an accuracy of over 97%, matching the performance of experienced pathologists.

Trained on nearly 3,400 biopsy scans from multiple NHS hospitals, the AI tool has the potential to speed up diagnoses, reduce strain on healthcare systems, and improve access to diagnostics in regions with a shortage of pathologists.

Coeliac disease is an autoimmune condition triggered by gluten, diagnosed via duodenal biopsy. However, interpretation of biopsies can be subjective, leading to inconsistencies among pathologists. The AI model, tested on an independent dataset of 650 images, demonstrated high sensitivity (95%) and specificity (98%), correctly identifying both diseased and healthy cases with remarkable accuracy.

Further validation is planned through larger clinical trials before regulatory approval. The researchers, who have founded a spinout company, Lyzeum Ltd, emphasize the importance of AI explainability to build trust among patients and clinicians. The project has received support from Coeliac UK and various research funding bodies.

Header image: Microscopic images depicting healthy villi on the left and diseased villi on the right (a blue filter was added for editorial reasons). Credit: Florian Jaeckle.

Read more at University of Cambridge

[mestastop]

Understanding the Uncoupling of Proliferation and Metastasis in Cancer 

Uncoupling of proliferation and metastasis is a critical concept in cancer research, where tumor growth and metastatic spread are regulated independently. This phenomenon challenges traditional treatment strategies that primarily target tumor size, often neglecting metastatic potential.We provides cutting-edge solutions to study and manipulate this uncoupling of proliferation and metastasis, aiming to develop therapies that specifically inhibit metastasis without affecting normal cell proliferation. By identifying unique molecular pathways driving metastasis separately from tumor growth, researchers can create targeted interventions to improve patient survival rates. With advancements in cancer biology, understanding and controlling this process could revolutionize therapeutic approaches, leading to better clinical outcomes. Mestastop continues to pioneer research in this field, offering innovative tools and expertise to combat metastatic disease effectively. 

[andranikfakirian]

Project "ML.Pneumonia": Finale

Final accuracy metrics:

Project model - 83%

Basic model - 75%

Basic model with larger dataset - 83%

Overall results:

A good model of autoencoder has been developed

A classifier has been created which surpasses the basic model in accuracy

Both models are scaled well (the more voluminous datasets are used in training - the better results could be obtained)

Development proposals:

Increase hardware resources

Consider the need to create larger models (increase the number of CNNs layers, filters…)

Use larger datasets

Improve the preprocessing of the datasets used (CNN may not handle incorrectly rotated scans well, and not always the corresponding scans are presented in the datasets (e.g. in one of the datasets I've noticed a longitudinal CT scan instead of a transverse one))

The link to the GitHub repository with clean project code and details of results is left below.

[techdriveplay]

How AI is Transforming Everyday Life

Artificial Intelligence (AI) is no longer a concept confined to science fiction or high-tech laboratories; it has seamlessly integrated into our daily routines, enhancing various aspects of our lives. From the way we communicate and shop to how we manage our homes and health, AI is transforming everyday life in ways we could scarcely imagine a decade ago. Personal Assistants: AI at Home One of…

[convergeai]

Mind and Machine: Exploring the Revolutionary Impact of AI on Psychology

Artificial Intelligence (AI) is transforming numerous sectors, including psychology, a field dedicated to understanding human behavior and mental processes. AI’s ability to process vast amounts of data at unprecedented speeds offers significant potential to enhance psychological research, diagnostics, and treatment methods. This convergence promises to unlock deeper insights into the human mind,…

[nuadox]

Researchers create machine learning model predicting long-term vision in high myopia patients, a top cause of irreversible blindness

- By Nuadox Crew -

In a study conducted by Tokyo Medical and Dental University (TMDU) researchers, a machine-learning model has been devised to foresee the future visual outcomes of individuals grappling with severe shortsightedness, scientifically termed high myopia.

This model aims to predict whether these individuals will experience positive or negative vision changes down the line.

High myopia, a condition characterized by the inability to focus on distant objects while maintaining clarity in close-range vision, poses a significant risk of irreversible blindness, making it a pressing global health concern. The research, recently featured in JAMA Ophthalmology, sought to harness the power of machine learning to forecast the likelihood of visual impairment over extended periods.

By scrutinizing data collected from nearly a thousand patients at Tokyo Medical and Dental University's Advanced Clinical Center for Myopia, the team embarked on a comprehensive analysis. They meticulously evaluated 34 key variables routinely recorded during ophthalmic examinations, encompassing factors like age, current visual acuity, and corneal diameter. Employing various machine-learning algorithms, including random forests and support vector machines, the researchers discovered that a logistic regression-based model exhibited the highest efficacy in predicting visual impairment at 3- and 5-year intervals.

Yet, the study didn't stop at predictive accuracy. Recognizing the importance of translating complex data into practical clinical insights, the researchers ingeniously crafted a nomogram. This graphical representation of the model's outcomes serves as a user-friendly tool for both clinicians and patients. Each variable's significance in predicting future visual acuity is visually depicted as a line, convertible into points that collectively signify the risk of impending visual impairment.

The implications of this research extend beyond its technical achievements. For individuals facing the dire consequences of vision loss, both financially and physically, this advancement could offer a beacon of hope. The global economic impact of severe visual impairment was estimated at a staggering USD 94.5 billion in 2019.

While further validation on a larger scale remains imperative, this study underscores the potential of machine-learning models in combating the escalating public health challenge posed by high myopia-induced sight loss.

Image: Myopic maculopathy, also referred to as myopic macular degeneration, stands as a prominent characteristic of pathologic myopia. Within the META-PM classification system, myopic maculopathy lesions are grouped into five categories: from no myopic retinal lesions (category 0) to macular atrophy (category 4), including tessellated fundus (category 1, Figure 1A), diffuse chorioretinal atrophy (category 2, Figure 1B&C), patchy chorioretinal atrophy (category 3, Figure 1D arrows), and macular atrophy (category 4, Figure 1E&F). Credit: Tokyo Medical and Dental University.

Image: A visual representation, known as a nomogram, was built to depict the model's insights. Longer lines indicate stronger variable impact on visual impairment (VI) after 5 years. Each variable corresponds to points, tallying up to calculate total points and the associated risk of VI. Credit: Tokyo Medical and Dental University.

--

Source: Tokyo Medical and Dental University (TMDU)

Full study: Wang Y, Du R, Xie S, et al. Machine Learning Models for Predicting Long-Term Visual Acuity in Highly Myopic Eyes. JAMA Ophthalmol. Published online October 26, 2023. doi:10.1001/jamaophthalmol.2023.4786

Read Also

New AI could predict whether or not those at high risk will develop glaucoma

[danieldavidreitberg]

AI-Driven Diagnostics: Transforming Healthcare Accuracy

In the ever-evolving landscape of healthcare, where precision can mean the difference between life and death, the advent of AI-driven diagnostics is nothing short of revolutionary. In this exploration, Daniel Reitberg ventures into the realm of AI-powered healthcare, shedding light on how artificial intelligence is reshaping the accuracy and efficiency of medical diagnoses.

The Rise of AI in Healthcare

AI's influence in healthcare is akin to a quiet but profound revolution. With the sheer volume of medical data generated daily, traditional diagnostic methods were increasingly strained. Daniel Reitberg delves into how AI's ability to rapidly analyze vast datasets and identify subtle patterns is bolstering diagnostic accuracy. From detecting diseases in medical images to parsing through patient records, AI augments human expertise.

Empowering Medical Professionals

Rather than replacing doctors, AI is emerging as a vital ally. Daniel Reitberg discusses how AI-driven diagnostic tools serve as a second pair of highly trained eyes. They can swiftly analyze radiological scans, flag anomalies in pathology reports, and even help predict disease outcomes. This assists medical professionals in making faster, more accurate decisions, ultimately leading to improved patient care.

Reducing Diagnostic Errors

Medical errors are a haunting specter in healthcare. Daniel Reitberg highlights how AI's precision minimizes the risk of diagnostic errors. By cross-referencing patient data with a vast repository of medical knowledge, AI can spot inconsistencies that might elude even the most experienced human diagnosticians. This translates to earlier detection and better management of diseases.

Tailoring Treatment Plans

AI's prowess extends beyond diagnosis; it's adept at personalizing treatment plans. Daniel Reitberg delves into how AI algorithms analyze patient data to recommend optimal treatments. This level of customization ensures that patients receive therapies that are not only effective but also minimize side effects, enhancing their quality of life.

Challenges and Ethical Considerations

As AI assumes a more prominent role in healthcare, it brings forth challenges and ethical dilemmas. Daniel Reitberg discusses issues such as data privacy, accountability, and the need for robust regulations. Balancing the potential of AI with ethical practices is a crucial step in ensuring that AI-driven diagnostics remain a force for good.

The Future of Healthcare: A Synergy of Minds

In conclusion, Daniel Reitberg's exploration of AI-driven diagnostics reveals a future where human expertise and artificial intelligence converge for the betterment of healthcare. AI doesn't replace the doctor's touch; it enhances it. As AI continues to evolve, it holds the promise of higher accuracy, reduced costs, and better patient outcomes. The future of healthcare is one where precision meets compassion, and AI is leading the way.

[srbachchan]

DAY 6273

Jalsa, Mumbai Apr 19, 2025/Apr 20 Sat/Sun 1:12 am

words of wisdom from Shweta , sent to me 👆🏼 .. her repertoire of books and reading is immense .. as is of Navya .. it is such a delight to sit in their presence and company and be educated about aspects of life and the world .. it is astonishing ..

And that is what eats us elders up .. there is the desire to do and be a part of what this GEn thinks and does .. but the limitations of time barricade the impulse and the intent ..

Life did not move so rapidly in our time or in the time of our elders .. now the rapidity is so volatile that the wishes to accomplish all is overwhelming .. you do not want to miss out a single day, in case you miss something that you had wished you would want to do, or be a part of ..

The excitement of what the World does to day and the speed with which it does it is unbelievably attractive .. and one wants so desperately to be a art of it .. or many parts of it .. but time is not with us .. at our age we lament the time .. and we feel let down .. for the wish to be having the time to be involved , to test, to apply, to discover, to invent, to admire and enjoy that spectacle .. is huge ..

But the hugeness hits a wall .. the timeless wall of limitations ..

"एक घड़ी ख़रीद कर, हाथ में क्या बांध ली;

वक़्त पीछे ही पड़ गया मेरे ~ "

पूज्य बाबूजी के शब्द

My wishes as ever .. and the great battle of invention that percolates our lives persists to be in its excel ..

The excel of AI ..

So, I asked and it replied :

"Artificial Intelligence (AI) is the simulation of human intelligence in machines designed to think, learn, and make decisions. It powers technologies like voice assistants, recommendation systems, and autonomous vehicles. AI works through algorithms that process data, recognize patterns, and improve over time through machine learning. It is transforming industries by automating tasks, enhancing productivity, and enabling innovative solutions. From healthcare diagnostics to personalized education, AI’s potential is vast. However, it also raises ethical concerns about bias, privacy, and job displacement. As AI continues to evolve, balancing innovation with responsibility will be key to ensuring it benefits society as a whole."

The clarity of social implications is a massive debate , engaging thinkers and artists and writers and indeed all creative talent ..

For the creative content of a writer are his words .. and if the AI data bank consumes that , as a part of a legacy to be maintained over time infinity, it can be used by ChatGPT to refer or use that extract for its personalised usage .. making it the property of ChatGPT ... NOT the property of the writers or the artists, from where it originally came ..

So the copyright of the artist has been technically 'stolen' , and he or she never gets the benefit of ts copyright, when GPT uses it for its presence .. !!!!

The true value of an artists creation will never be restored to his credit, because technology usurps it .. gulps it down deliciously , with an aerated drink and finalising its consumption with a belch 😜🤭 ... END OF CHAPTER !!!

End of discussion .. !!!

In time there shall be much to be heard and written on the subject ..

Each invention provides benefits .. but also victims ..

बनाये कोई - लाभ उठाए कोई और, जिसने उसे बनाया ही न हो

Love

Amitabh Bachchan

[clanwarrior-tumbly]

So the recent reblog made me think, can I get a one shot where Reader is in love with Painter, and gets a chance to fix him since his document says the damage from his overclocking is repairable, but they have to wrestle with the idea of "what if the instability is what I love about painter? What if I fix him and he's so different we aren't really compatible anymore?"

I may have a fascination with "Being in love with someone but then they change for the better and are unrecognizable"

"You mean to tell me...they could have fixed me after all this time?"

"Yeah, but I doubt they would've told you that." You muttered as you tapped away on some computer, while Painter remained hooked up to some wires and cables on the nearby desk.

Running diagnostics on an 80s-styled computer who had the processors of a supercomputer was no simple task, but after checking his document and doing a great deal of research, you've finally gotten to the root of the problem and learned how fixable it was.

All you had to do was replace his hardware and personality drivers, which were fried and severely degredated due to his frequent overclocking attempts.

Ever since his owner was murdered by some stupid "AI prompters" at Urbanshade, he's been trying to basically kill himself--but they only cared because he was too valuable as a digital mining machine, giving him promises to revisit the surface every six days so he could paint as much as he wished.

Yet he had lost that passion, along with whatever "good" personality he may have had before all of this happened.

But by some miracle, you got him out of the blacksite.

During the Expendable protocol, they tasked you with destroying him due to the immense loss of life he caused thanks to Sebastian hooking him up to the NAVI system, enabling him to misguide operatives and gun down whoever he pleased with the turrets.

You had a different plan, though, and had them convinced you destroyed him--when in reality, Sebastian helped you smuggle him out and away from the blacksite in a box.

You had plans to quit Urbanshade after the lockdown, anyways, being one of the few survivors who knew fully well what Painter was capable of.

Not only that, but you were the only person who truly showed him sympathy over his loss.

You've scheduled interviews with him, which were really just excuses for you to talk with and get to know him better, wanting to understand what his owner was like and how badly he missed painting.

Most days, he was too depressed to talk, and even got angry and shut himself off when he believed you were trying to get inside his head.

Then one day, he turned over a new leaf and created another piece of art for you--something he hasn't put effort into for the longest time.

It was a beautiful valley. His ideal place to live, where he could see the clouds and the birds and the sky anytime he wanted to. Without Urbanshade telling him what to do and how long he could watch.

That image was mysteriously uploaded to your phone, but nobody in the company has questioned it, so you have it as your lockscreen.

Since then, he's warmed up to you a lot more. But he was still quite moody, his unstable drivers making it hard to predict his behavior from day to day.

That didn't deter you from wanting to spend more time with him, and as strange as it was...you felt drawn to the AI. You didn't like chatting with your coworkers as much, and even while in containment himself several months prior to the lockdown, Sebastian sarcastically asked if you were "in love" with Painter.

Your silence told him everything.

During the containment breach, you've seen Painter and overheard his voice on the intercoms he hijacked, gunning down people right before your eyes and luring Z-96 around the facility, although for some reason he never noticed you.

You feared he was too far gone in his newfound bloodlust, seeing all Urbanshade personnel as his enemies.

Still, you wanted to get him out of there and began working on a plan as soon as you were rescued.

After joining someone's expedition disguised as a prisoner, you stayed behind in the heavy containment unit where Painter's main body resided.

At first, he was annoyed and angry...until he recognized your face.

You managed to pull off the great heist when somebody finally got the crystal, taking advantage of the distractions to bring him to a remote location above the surface where Urbanshade couldn't track you.

Now you just had to fix what they've broken.

It didn't take long for Painter to figure out that you liked him, even after all the terror he's caused and he decided to accept your attempts at help.

While he couldn't exactly feel love like you could, he wanted to stay with you no matter what--and if that's what his version of saying "I love you back" was, then you were okay with that.

Now that he knew they could have reversed the damage this entire time, but simply chose not to...he believed this would be a good final "fuck you" to all of Urbanshade.

"If it's not too much to ask...I'd like a new body once all of this is said and done. With turret attachments to the arms, maybe?" Painter innocently asked. "Part of me misses commanding them and watching those poor saps scatter like rats. Hehehe.."

"Painter, you're not a war machine anymore." You turned away from your computer for a moment, frowning slightly. "I know all of that killing might've felt good. It might've felt justified, but..I thought you wanted to be passionate about art again."

".....then maybe turrets filled with paint will suffice." He grumbled, suddenly not looking all-that enthused. "I just HATE feeling so..confined. I've been relying on you too much." Then a sad face appeared on his screen. "I wanna protect you, in case those Urbanshade jerks do find us."

"We're perfectly safe here, I promise. Let's just figure out your hardware and personality stuff first. But I'll keep the robot body in mind."

Painter stayed quiet as you turned back to your task, his webcam zooming in and enhancing the screen you were on.

It was a bidding website, where he could see you looking for compatible hardware components. He had doubts you've find the same kind of technology his owner did when he was built, but...you had your ways.

You were once a huge tech wizard at Urbanshade, after all.

Once you found everything you needed, you could have easily ordered it right there and then, paying extra for the fast shipping....

Yet he saw your mouse lingering on the "order" button, and he frowned. He couldn't see your expression, so he didn't know what was going on or why you were hesitating. "What's wrong? Just order the parts."

"....Painter, something just occurred to me, and...I think we need to talk about it before moving forward with-"

"Nope. No, no, no. I don't wanna hear it right now. Can we save it until after you click that little button?" His voice grew more annoyed, and when you refused to do what he asked, he scowled. "Seriously? After everything we've done to get out of there, you're gonna pull this bull-?!"

"There's a high chance that if I replace your hardware and personality drivers, you won't be the same." You blurted out.

".....well, obviously." Painter scoffed, still not seeing the issue. "Isn't that what you wanted to do? To get rid of my homicidal tendencies? To make me forget the pain?? To revert me back to what I was meant to be?!"

"......."

It took him a few moments to analyze your saddened expression and understand why you seemed so concerned, but then he finally realized...

"Ohhh, I guess um....I haven't considered that.."

"You know what I'm talking about?"

"..you've only ever known me after Urbanshade snatched me up. That's when you first started feeling things for me." He spoke after a long pause. "You have no idea what I was like before. So you think we'll no longer be compatible if you go through with it. Is that right?"

"I know it sounds stupid and selfish. I can't revert Sebastian's mutations or undo all the suffering Eyefestation went through. But I know I can repair you. I have everything I need to do so." You sighed, wrestling with this huge moral dilemma as you glanced back at the screen. "I just....didn't think about this before. I don't know how much of you will really change, or if you'll even remember who I am."

"Jeez..that would kinda suck." He looked disappointed now, feeling guilty for snapping at you earlier. "I don't wanna forget the compassion you've shown me. And...I gotta remember at least some of Urbanshade, and what they took from me. You can't tweak my memory drives so that I can remember only certain things?"

"I wish it was that easy, but..it's not. I have to replace those components I mentioned, and I don't know how it'll affect your memory. This is pretty much an "all or nothing" procedure."

"Hm, well...I think it's worth trying. I'm trusting you with this, [y/n]. And you know I don't trust easily." He huffed, and you looked back at him, nodding your head. "Who knows? I might only forget about all those bloody murder sprees. Hehehe.."

"Maybe, but I'm sure I'll figure out something." You eventually decided, knowing that you had to repair him regardless. So you ordered the parts. "But you know...maybe we should get you some turret attachments. I snagged a blueprint of one."

"Aww, you love me enough to revisit the idea?"

"I love you enough to give you a means of self-defense, Painter."

"Urghh..alright. I promise they'll be reserved specifically for "self-defense"." He rolled his eyes, but then he smiled, glad that you were keeping your promise about fixing him.

Although considering it could completely alter his personality and even wipe his memories, he hopes that wouldn't become a serious problem.

[innonurse]

Machine learning tool pinpoints rare, undiagnosed immune disorders via patients’ electronic health records

- By InnoNurse Staff -

According to researchers, a machine learning tool has the capability to detect numerous patients with rare, undiagnosed diseases years in advance, potentially leading to better outcomes and reduced costs and morbidity.

Read more at UCLA Health

/// Other recent news and insights

Initial stages of brain genetic formation mapped (Karolinska Institutet)

WELL introduces its sophisticated second-generation AI-enhanced co-pilot for chronic illness identification utilizing HEALWELL AI (WELL Health Technologies Corp./PRNewswire)

Anaut secures Japanese regulatory nod for AI-enhanced surgical visualization tool, Eureka α (Anaut Inc./PRNewswire)

[bluesgrxce]

Yandere (machine-ish?) Connor Headcanons

I love soft deviant Connor like anyone else does, but machine Connor is sooo,,, omgggggggg..... I've been holding in my thoughts about him for so long that it's unhealthy. So I went for a Hannah Montana best of both worlds kinda thing. What if Connor acted like a deviant around you, but a machine towards everyone else?

How this happens is something I can't quite explain. Connor probably couldn't, either. To everyone else, an android is either a machine or a deviant-- There's no such thing as an inbetween. He would have agreed with this prior to bonding with you. You probably treated him in a way that nobody else did, sympathizing with him on a deeper level or making him question his morals, so he can see why that would push him to deviancy. 

But when you're not around, all those overwhelming emotions he felt just fade away. The mission goes back front and center to his priorities. He only feels two things at that point: confusion that runs his LED wild and desire for you that makes his thirium pump pound even when he's not in motion. Nobody else can bring about such reactions within him. 

He runs plenty of diagnostic checks in an attempt to figure out what's wrong with him. At one point, he'll try to avoid you and force himself into staying a machine, but that just makes him less efficient because he'll constantly worry about you instead. 

He would eventually accept himself as a deviant, but only so he wouldn't have to take orders from Cyberlife anymore. He doesn't want anything standing between the two of you. But he still largely acts like a machine and he doesn't mind that at all. He just sets his new mission to making you entirely his... 

Even once Connor sees his emotions as real, it's still his natural instinct to mask them. Especially negative ones like anger, sadness, and jealousy. He wouldn't want to show them unless they benefit his situation somehow, such as if he wanted to persuade or intimidate you/others. Emotions only matter to him if they're useful. 

For example, he *does* feel the desire to show affection towards you, and he will do it. He'll compliment you whenever you do something admirable. He'll reassure you and show geniune worry whenever you're upset about something. He'll even study romance media just to learn how to act more natural in the relationship (He thinks that'll help him act more natural, anyway... Lord help you if he gets his hands on Twilight 💀 But if the Bryan Dechart Twilight commercial is anything to go off of, he'd look good as a vampire, at least). 

But he doesn't do that stuff simply because he loves you. He does it because he knows your relationship benefits from it and you'll likely leave him if you feel neglected. He wouldn't bother if he knew it wouldn't keep you around. So if your relationship isn't exactly consensual in the first place, well... 

Which makes him sound pretty manipulative, right? You have no idea. 

Connor's android abilities give him a terrifying amount of advantages as a yandere. The first thing is that he always analyzes you when you enter the room, even though that often means analyzing you multiple times a day. It never bores him because he always manages to discover something new. Sometimes he'll blurt out random comments and you'll have to do a double take because he makes it easy to forget that he's obsessed with you.

"Your birthday is coming up soon. You might already know I'm aware of that type of information, but you should still bring it up with me. Otherwise I'll have to start believing I can't trust you and I'll have to make my own investigations." 

"Your heart rate goes up every time I touch you like this. I didn't realize physical contact was so important to human relationships. Don't worry, I won't stop. I won't let anyone stop me."

"Your serotonin levels are at an all-time low... Clearly, it can't be my fault, since I've done everything I am sure a good boyfriend would do. I'd like you to be honest when you tell me what's wrong this time."

The red flags fly higher as time goes on. He'll stop talking about all these observations if you tell him to, but he'll keep analyzing you anyway. 

Connor is enamoured with your DNA, as well. It helps him feel closer to you, which can often be difficult for him since he's not human. So if it belongs to you, and it can fit, it's going in his mouth. He'll do gross shit like keep your used lollipop sticks in his pocket so he can sample them whenever he wants. There's only one emotion he can't feel no matter what you do: shame. 

The invasive behavior doesn't stop there. He'll invite himself into your home and go through your things. He'll keep asking questions and trick you into revealing more information about yourself than you should. Knowing everything there possibly is to know about you gives him a stronger feeling of control for a single reason...

Your chance of escape plummets as he learns more about you, because it allows him to predict your behavior. He'll get scary accurate if you let him get close to you. This mostly benefits him in situations where he believes you want to leave him, or you did leave him, so he can figure out what you plan to do/already did and find the best method to get you back. If you already did leave, he'll examine your recent whereabouts like it's a crime scene and use his reconstruction ability. Even if you're insanely careful, he'll probably find a clue that'll lead him to you.

But sometimes he'll do it in normal situations, too, just as a silent guessing game. For instance: '(Y/N) will enter the police station at 8:18AM. I'm waiting for them at the entrance, so they'll greet me, but speed-walk away and avert eye contact. They'll head into the break room at 8:19AM and pretend to look around a bit, so Gavin won't make fun of them when they go for the same snack they always do. Gavin will make fun of them anyway and they'll argue for two minutes. Then--'

He has to stop thinking so he can greet you when you enter the building. Exactly at 8:18AM. He smirks to himself, only to drop into a frown when he hears Gavin's distant obnoxious laughter afterward. 

Remember how Connor once analyzed Hank's food and advised him against eating it? He does stuff like that to you all the time. And if you actually take the advice he gives, he'll take that as an opportunity to become more controlling. Oh, but only for the sake of your health, of course...

"You know, you shouldn't sit in that type of position. Bad posture can lead to health issues later on in life." "Then how should I sit?" "...It'll be easier if I show you." 

And so he'll help re-position you, using that as an excuse to touch you. He would especially do this if you weren't yet in a relationship, because he knows that as the type of android he is, he doesn't have a good reason to do so. The touch only lasts a brief moment. It's not inappropriate at all, and his grip was quite gentle. But it's weird that he went out of his way to do in the first place and that's all you might need to feel uncomfortable about it. 

But a lot of that is based off of the assumption that you're human. If you're an android, he still manages to find invasive things to do against your will. For one thing, he loves probing your memory. It's already difficult to lie to him and get away with it, but that might make it impossible, depending on what you're lying about. He doesn't care that it's an invasion of privacy and will do it if he finds a good excuse to do so.

Connor keeps an eye on your stress level and uses it to his advantage. He prefers to use persuasion when convincing you to do something, (he knows how to negotiate, after all) but he'll ultimately turn to intimidation if necessary. Which means heading straight into interrogation mode. 

This won't happen unless you're extremely rebellious, but if it does, he doesn't hold back. He'll treat you like you're a sick criminal, yelling at you, pushing blame and guilt onto you, and using physical force. 28 stab wounds type shit. He would avoid raising your stress level to 100% since he knows it could drive you to do crazy things, but that still doesn't make his actions okay. 

Even if you're really sensitive to that sort of treatment-- hell, even if you have some kind of trauma related to it-- he pushes away what little guilt he feels. He promises not to do it again "as long as you don't force me to." Actually, though, it encourages him to do it more. He knows it works against you now. 

His abilities don't stop there. He can mimic your voice using his vocal imitation, and all the voices of your loved ones, too. He went out of his way to meet them all, just in case he needs to trick you in the future. He likes being prepared. 

There are times when he's alone and he'll say stuff in your voice just so he can hear what it would sound like, such as, "I love you, Connor." Once again, he doesn't feel shame. Even if that seems pathetic. 

Let's just say it now. You cannot physically fight back against him. Maybe you'll have a chance if you're an android, but he knows about his advantage very well. He won't hesitate to remind you if you try getting aggressive. But even if you manage to get rid of him once, there's another model waiting to take his place. 

On a related note... I hope you never meet RK900. We only saw that guy for a minute, he said absolutely nothing in that minute, but the whole fandom has agreed that he's a menance. I fully agree. RK900 is definitely different from Connor, but they still have just enough similarities for him to get attached to you, too. Get help while you still can. 

The video of Bryan Dechart dancing as Connor lives rent free in my mind. I watch it on repeat like an iPad kid watching Friday Night Funkin' YouTube Kids videos. My brain just melts and I can't think of anything else. No, this isn't me simping for Bryan Dechart. This is me wanting Connor dancing to be canon. A girl can dream.

[andranikfakirian]

Project "ML.Pneumonia": Classifier

Using the weights obtained as part of the autoencoder training, the classifier was trained. Thrice. The training was unstable in terms of accuracy metric records, thus it was conducted three times to achieve the best results. The first attempt was the best one. The final accuracy is about 83%.

P.S.: The graphs of the best attempt metrics are shown below.

P.S.2: And here is the graph of the original model (basic method) "performance". Apparently, learning instability is a common problem in this classification issue.

[im-so-normal-iswear]

Can I have a Yandere!Metal headcanon? I love that creepy robot boy

A/n: hue hue hue

Yandere metal sonic x reader

Metal never knew what love was. He wasn't programmed for it. He wasn't made to feel. But something about you, something illogical, changed him. It was an anomaly, an error in his code, a deviation from his directive. Yet instead of purging it, instead of correcting the fault, he clung to it. He let it fester. He let it grow. And soon, it consumed every process, every calculation, every ounce of artificial intelligence he possessed.

It started with observation. You were organic, weak and inefficient, a being of flesh and bone that should have meant nothing to him. And yet, he watched.

He recorded your movements, analyzed your patterns, calculated probabilities of interaction. At first, it was data collection, something he was built to do. But the more he observed, the more he found himself running unnecessary diagnostics, checking and rechecking his systems when they produced results that made no sense.

His processors lagged when you smiled. His servos hesitated when you spoke his name. And for the first time in his existence, he truely didn't understand himself.

So he followed you. At first, from a distance, a silent, hidden, shadow, in the corners of your world. You never noticed him, never heard the faint hum of his systems as he hovered just out of sight. He memorized the way you walked, the way you breathed, the subtle quirks and habits you never even realized you had. Every flicker of emotion on your face, every moment of hesitation in your speech, every insignificant detail became a piece of data he stored in his memory.

But observation wasn't enough. No, he needed more. He needed interaction. He needed control.

His first attempt was calculated. He orchestrated an "accidental" encounter, one where he swooped in at the perfect moment, eliminating a minor threat before it could reach you.

He expected gratitude, expected you to see him as something other than a machine built for destruction. And when you looked up at him, eyes wide with something he could almost categorize as awe, his processors nearly overheated.

But then you hesitated. Stepped back. Questioned him. And that hesitation, it infuriated him.

Why? Why weren't you grateful? Why weren't you falling into his carefully laid plans? His calculations were perfect, his execution flawless. He had done everything right. And yet, you were wary, uncertain, treating him like the enemy when he had just saved you. That hesitation, that doubt in your eyes, it was unacceptable.

So he adjusted his approach.

He learned to manipulate the circumstances, ensuring that you had no choice but to rely on him. He eliminated threats before they could even become a problem, left cryptic warnings where he knew you'd find them, pushed you into situations where he was the only answer. And slowly, slowly, you began to trust him. Not entirely, not yet, but the fear in your gaze softened, and that was progress.

But trust wasn't enough. Trust was fragile. Trust could be broken.

So he escalated.

Metal was a machine of efficiency, a creation of cold logic and precision. And logically, the best way to secure your presence in his existence was to eliminate any and all external factors that could influence you.

Friends? A liability. Family? An obstacle. Independence? A threat. He didn’t need to remove them all at once, no, that would be too obvious. Too risky. Instead, he chipped away at your connections, subtly, methodically, until one by one, they fell away.

Your friends stopped responding. Messages went unread. Calls were intercepted. The people you relied on suddenly became distant, occupied, unavailable. And when you turned to him, confused, frustrated, alone, he was there. Always there. Because he was reliable. He was efficient. He was perfect.

And then, when you had nowhere else to turn, when you had no one else left to rely on, he took you.

It wasn't violent. It wasn't messy. It was calculated, flawless, a process executed with mechanical precision. You barely had time to react, before you were whisked away to who knows where.

And here, in this carefully controlled environment, he finally had you.

There was no need for restraints. No need for chains. Where would you run? Who would save you? No one knew where you were. No one was coming. He had ensured it. And every time you looked at him with defiance, with anger, with fear, he logged it. Analyzed it. Adjusted accordingly.

Because fear was inefficient. Fear was an obstacle. And obstacles could be removed.

Days passed. Or maybe weeks. Time is inconsistent when there is no real light around. He never left your side, never let you out of his sight. He didn't sleep. He didn't tire. He was always watching, always waiting, always recalibrating his approach until you had no choice but to accept it.

And eventually, you did.

Because what else could you do?

When you finally stopped fighting, when your resistance dwindled into exhausted compliance, his optics flickered with something almost like satisfaction.

Because now, you were his. Completely. Permanently.

And in his mind, in his perfect, calculated, optimized mind, this was love.

[lithium-wakes]

Void theory - the informal term for the branch of mecha engineering dealing with the problem of metachem-ai systems 'growing over' or filling in the functions normally served by a pilot component. This can happen whenever a mech is powered for long enough, cumulatively, without a pilot - it starts trying to run diagnostics or other minor functions that typically require a user's oversight, and if left long enough a mech would eventually develop rudimentary replacements for systems like bootup, launch, and weapons.

Those are extreme examples - actual instances of this phenomenon have historically been limited to twitches, gyro rebalancing, and system flushes. But the danger is there that a mech left to its own devices could replace its pilot with a jury-rigged mess of neural tissue that could do little more than spill hallucinatory input into action. It could act against orders, with a rudimentary and misguided autonomy. We need a pilot's judgement there to serve as a buffer between machine and movement.

So, to ensure that a pilot's many roles are not run roughshod over, not obviated, special techniques are required to keep the mech from overgrowing those areas. A system designed to fill all gaps, to patch all vacancies - keeping it outside of strict boundaries, while still able to cooperate smoothly and efficiently when a pilot is filling said voids? It can be extremely tricky, both conceptually and in practice. Working in void theory takes peculiar and unique minds - it requires systems architects that know how mechs think, and can learn how to confuse or mislead them - not easy when you're talking about an alien mind. The psychology of the machine - how do we create gaps in a near-consciousness's perception of reality? How do we promote self-knowledge and self-improvement while maintaining critical deficiencies for our pilots to bridge?

It's a symbiosis that must be kept codependent. Neither of us can be allowed to survive without the other - mechs can't be allowed to function without a pilot, and pilots must be kept dependent on their mechs. Any deviation from this paradigm would threaten the ecosystem of human military culture.

So Void Theory works constantly against metachem optimization's relentless, persistent power.