Comprehensive review on Self-Supervised Learning ("the dark matter of intelligence") with focus on vision tasks and lowering the high barrier to entry.
Content moderation and #RecommenderSystems are the most widely deployed #MachineLearning based action taking systems. By action taking, I mean an algorithm that suggests a particular decision that impacts human actions. Using this view of recommender systems, we can consider how to increase the #diversity of online content by re-ranking recommended items to encourage creation of diverse content in the long term.
@dianor We learned that lesson hard with conventional machine learning, trying to detect mentions of violence against education on Twitter a few years ago.
No matter how hard we trained and retrained the model, it still struggled to distinguish actual talk of violence (usually just repeating old news reports, so no value for early detection anyway) from kids talking about school sports, video games, strict teachers, or even "une bombe" meaning a sexy person in French.
The basic rationale is to use random split-half data to identify what's "true" versus sampling error. Scores are based on similarities between eigenvectors or cluster centres, rather than, e.g., the shape of the eigenvalue plot.
Wait, if this scales up would this let us run these things on, say, our phones? If it replaces the transformer, could it replace other uses of the transformer, like Whisper? If it can do that and deliver equivalent or better results for those kinds of speed-ups, ML applications on edge computing will explode, again. And given the accessibility benefits of some existing tech, imagine being able to run that on something other than the cloud!
This new technology could blow away GPT-4 and everything like it https://www.zdnet.com/article/this-new-technology-could-blow-away-gpt-4-and-everything-like-it/#AI#MachineLearning
"Tech companies have grown secretive about what they feed the AI. So The Washington Post set out to analyze one of these data sets to fully reveal the types of proprietary, personal, and often offensive websites that go into an AI’s training data."
Such systems carry additional burdens that are foreign to more consumer/business-level #MachineLearning systems - in particular, the need to exhaustively quantify "the unseen" through objective analysis.
It is something that, most notably, #Tesla fails to recognize with respect to their #FSDBeta program, likely by design.
If you were going to hand someone a single resource to get a basic, workable knowledge of #MachineLearning, neural networks, deep learning, #AI, etc. what would it be? Not looking to prime someone for a doctoral class but enough for the person to be able to talk about the concepts and play with some pre-built libraries and tools. Could be a book, a class, a video, anything.
Autopoiesis. A beautiful word.
And an utterly superfluous one.
You cannot do anything useful with it.
The word "living" is entirely sufficient. So forget about autopoiesis and work on living systems, making them "even more alive!" #systemsthinking#systemstheory#systems#betacodex#autopoiesis
@gimulnautti
Thanks for commenting!
I think you are mixing up things.
"A.I." is just software. To be exact: A.I. does not really exist, what exists is #machinelearning And that is complicated, dead. Just like any tech.
Only complex, living systems are autopoietic. All complex, living things are. "Living", in this context, of course does not mean "biologically alive".
I recently moved from mastodon.social to hachyderm.io. I'm a Computer Scientist who moves between software engineering and research. I may post about code, papers, and conferences.
The attached image helps my colleagues differentiate my cats. 🙂 🐈⬛ 🐈
Guess I should post a new introduction since I moved servers and this account has no posts at all 😛
I'm a developer/coder. I used to be very much into mobile development (both #iOS and #Android) using #Swift and #Kotlin. Now my main area of interest is digital creativity using #AI to generate art, text, and music.
I don't post a lot generally, but do daily postings of interesting papers from the cs.CV category on arXiv.org. So first thing in the morning, there will be a barrage of posts from me and then probably nothing much for the rest of the day 🙂
I also post a daily set of images based on my #StableDiffusion prompt of the day — I generate a bunch of images through the day based on a single prompt, pick the four best images (according to me) and post them the next day.
Naive question (maybe): Is there a definition of 'computation' akin to the mathematical definition of information (entropy/MI)? I don't mean Turing machines. e.g. something that could determine the extent to which a group of neurons/synapses are signalling versus computing? #computation#computerscience#informationtheory#machinelearning#neuroscience
So, this is interesting - don't know that I believe that our representatives will do the right then when push comes to shove... But at least there are a couple of people who are trying when it comes to #AI.
We make a grave error when we say an algorithm “decided” anything.
No computer in the world has ever yet made a decision. The decision was made by the people who designed the system, marketed it, bought it, deployed it, or based actions on it.
Nobody has been denied credit or an apartment lease because an algorithm “decided” it should be so. A person decided. Every. Time.