Bahaeddin ERAVCI @beravci.bsky.social · 06/05/2025Who would deny this world's transience, so clear to see? Who does not know, though long we live, that death will surely be? Who would deny three measures of cloth (the shroud) are the final legacy? Who does not know this very grave is our true homeland. 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 06/05/2025Rough English translation: O, people of wisdom, look upon our state, This is, for certain, human's ultimate fate. A troubled, ephemeral dominion is this earthly place, Where every soul mistakes its hardship as some grace and comfort. 100 1
Bahaeddin ERAVCI @beravci.bsky.social · 06/05/2025Started my NLP lectures today exploring the fascinating levels of natural language. This slide features an interesting example: Turkish, using Greek letters (orthography) on a historic tombstone from Istanbul. A poetic and meaningful common lesson transcending barriers... 110 1
Bahaeddin ERAVCI @beravci.bsky.social · 07/04/2025We still know very little about complexity theory and the emergent behaviors of self-organizing processes across micro-to-macro systems. I guess, today’s highly compartmentalized mainstream scientific tradition hinders progress toward a holistic understanding. 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 07/04/2025The universe, in a baffling sense, creates local pockets of complexity (decreasing entropy through biological/physical self-organized structures) while relentlessly advancing toward a state of maximum global entropy. www.quantamagazine.org/why-everythi...quantamagazine.orgWhy Everything in the Universe Turns More Complex | Quanta MagazineA new suggestion that complexity increases over time, not just in living organisms but in the nonliving world, promises to rewrite notions of time and evolution. 110 1
Reposted by Bahaeddin ERAVCIBahaeddin ERAVCI @beravci.bsky.social · 15/03/2025Don's main distinction for a CS mentality: - ability to jump very quickly between levels of abstraction, between a low level and a high level, almost unconsciously - deal with non-uniform (he means mathematically dis-continuous, discrete IMO) structures 001 1
Bahaeddin ERAVCI @beravci.bsky.social · 15/03/2025Don's main distinction for a CS mentality: - ability to jump very quickly between levels of abstraction, between a low level and a high level, almost unconsciously - deal with non-uniform (he means mathematically dis-continuous, discrete IMO) structures 001 1
Bahaeddin ERAVCI @beravci.bsky.social · 09/03/2025Came across a book (actually a transcript of lectures at @mitofficial.bsky.social) from a CS legend Donald Knuth, the author of The Art of Computer Programming. Not nearly as popular as TAOCP. Love the line "Computer God talks about God" in the foreword, we'll see where it leads... 110 1
Bahaeddin ERAVCI @beravci.bsky.social · 02/03/2025VLDB is a good example with a monthly cycle and accepted papers getting published concurrently in PVLDB each month. 130 3
Bahaeddin ERAVCI @beravci.bsky.social · 02/03/2025The origins of aesthetics is really fascinating. Why deem this scene utterly spectacular and even tie it to the "long-tailed mountain lady"? What mechanisms shaped this "taste" and how? 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 27/02/2025#Severance isn’t a typical tv show. It’s a sharp dive into philosophy of mind, probing identity, memory, and mind-body duality with surprising depth. Highly recommend... 050 5
Bahaeddin ERAVCI @beravci.bsky.social · 18/01/2025There appears to be a striking correlation between ignorance on a topic and the confidence with which people make bold statements about it. Can easily use this principle as a de-noising filter... 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 12/01/2025Self (and its counterpart the other) is a very handy abstraction to make the most of our limited processing power. Illusion of free will is a beneficial yet erroneous causal explanation we created after we observed the self interacting with the other(s) for some time. m.youtube.com/watch?v=_Ig9...m.youtube.comThe illusion of self and the illusion of free will, explained | Annaka HarrisYouTube video by Big Think 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 01/01/2025People with absolute no theoretic or practical knowledge/experience (not a single call to nvidia-smi) about deep learning seems to easily predict the future of AI. Their self-proclaimed prophetic confidence -still- amazes me. 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 29/12/2024Incentives influence choices and mass choices create the Zeitgest. Unrestricted social media ecosystem favors the quick and superficial, seducing everyone —from professors to everyday individuals— to contribute to content that appeals to our lower faculties, much like primal instincts. 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 29/12/2024Some reflections and insights after 1993 NIPS by Leo Breiman known for developing CART, bagging, and random forests. Always find less formal writings of the pioneers more insightful. 020 2
Bahaeddin ERAVCI @beravci.bsky.social · 29/12/2024Intersection of info theory and complexity theory has always been very interesting. While entropy quantifies global uncertainty (potential information), observer-dependent entropy brings observer’s view (its world model) to define subjective uncertainty. www.quantamagazine.org/what-is-entr...quantamagazine.orgWhat Is Entropy? A Measure of Just How Little We Really Know. | Quanta MagazineExactly 200 years ago, a French engineer introduced an idea that would quantify the universe’s inexorable slide into decay. But entropy, as it’s currently understood, is less a fact about the world th... 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 22/12/2024AGI isn't around the corner and scaling auto-regressive LLMs won't get us AGI. I argue, while AR-LLMs are great improvements, we need some very important paradigm shifts with lessons from the past. open.substack.com/pub/beravci/...open.substack.comBridging Generative AI and Truth: Ancient Lessons for Modern TechWhen a seasoned lawyer last year filed a brief citing six precedent-setting cases, he trusted the AI chatbot that assisted him. 040 4
Bahaeddin ERAVCI @beravci.bsky.social · 14/12/2024The infamous event of "cultural generalization made by a keynote speaker" shows how bias and miss-generalization are hard problems even for humans (even if a MIT professor). So, we maybe more compassionate with LLMs trained on our data. 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 14/12/2024Hallucination, especially somehow grounded ones, is not a bug but rather a feature. Artists, for centuries, have been the "hallucinators of human society" challenging the common sense, i.e. the common daily patterns in the society. 011 1
Bahaeddin ERAVCI @beravci.bsky.social · 12/12/2024#NeurIPS and other major conferences should consider making presentations, at least important keynotes/highlights, publicly available. I could easily make an argument with public fundings for research presented. Funding agencies can also support this for more open science. 010 1
Bahaeddin ERAVCI @beravci.bsky.social · 09/12/2024Feels like the beginning of 1900s with huge discoveries each year but this time huge strides in tech. Exciting to be a witness of the tech revolution ranging from AI to quantum compute... blog.google/technology/r...blog.googleMeet Willow, our state-of-the-art quantum chipOur new quantum chip demonstrates error correction and performance that paves the way to a useful, large-scale quantum computer. 011 1
Bahaeddin ERAVCI @beravci.bsky.social · 07/12/2024GPU poor man's home setup ready for a long night... 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 29/11/2024Great insights, thx @tiziano.bsky.social. Especially like it because its prospective and interventional experiment. One question: Any observed/estimated bias because of the "install browser extension" constraint? 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 28/11/2024This saga reminds me of the access modifiers I teach in my Java OOP course. Maybe we need something similar in the generative AI age: - Public: Content accessible to both AI and humans - Protected: Human-only consumable public content Easier said than done with a lot of technicalities though... 010 1
Bahaeddin ERAVCI @beravci.bsky.social · 27/11/2024Test-of-time awards are the **real impact** metrics akin to revolutionary science in Kuhnian sense. Congrats to @ian-goodfellow.bsky.social and Ilya with GANs and Seq2Seq. blog.neurips.cc/2024/11/27/a...blog.neurips.ccAnnouncing the NeurIPS 2024 Test of Time Paper Awards – NeurIPS Blog 010 1
Bahaeddin ERAVCI @beravci.bsky.social · 26/11/2024:) Not an easy task to match 1kw magnetron power with a beamformer in 12cm wavelength (2.4GHz) region in the usual oven size. Maybe with a roomsized oven... BTW, guessing some EE background Durk? 120 2
Bahaeddin ERAVCI @beravci.bsky.social · 26/11/2024Yes, this is from the book but I think it is originally from one of Tenenbaum's previous papers on game engines for learning physics. 010 1
Bahaeddin ERAVCI @beravci.bsky.social · 25/11/2024This we need definitely @bsky.app bsky.app/profile/bera... 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 25/11/2024Image credits: Griffiths, Chater, Tenenbaum 2024 100 1
Bahaeddin ERAVCI @beravci.bsky.social · 25/11/2024When we're talking about learning (machine/biological) we should not forget about the giant feedback loop which we are trying to model and infer. Folks who think AGI can be achieved from internet text with scale alone are either: - Hyping for their personal gains - Don't have a clue whatsoever 120 2
Bahaeddin ERAVCI @beravci.bsky.social · 24/11/2024Wondering how a global/local context shift can be integrated to these network models. This effect is called Zeitgest in social sciences, "spirit of the age", an invisible agent which dominates the herding behavior. 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 24/11/2024Along with the great lectures on MIT OCW. RIP Gilbert. Fun fact: his favorite matrix was a basic differentiator. 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 23/11/2024A disciple of Golden Mean waiting to be included :) 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 23/11/2024Swiss church installs AI-powered Jesus: Two-thirds of the users found it to be a “spiritual experience” "I think there is a thirst to talk with Jesus." www.theguardian.com/technology/2...theguardian.comDeus in machina: Swiss church installs AI-powered JesusPeter’s chapel in Lucerne swaps out its priest to set up a computer and cables in confessional booth 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 23/11/2024But AI research, for the first time in history, is providing us tools and analysis methodologies to quantify the possible solutions and perspectives. This is why I find it very exciting to be within this space. 010 1
Bahaeddin ERAVCI @beravci.bsky.social · 23/11/2024A lot of contemporary discussions in AI/ML have a corresponding philosophical discourse. For example, use of copyrighted material in LLM is linked with the Ship of Theseus paradox. We frankly don't have answers for these problems and maybe will never have an objective answer. 120 2
Bahaeddin ERAVCI @beravci.bsky.social · 22/11/2024Intelligence is hard to define but is related to in-domain vs out-domain generalization. Trivial case being k-NN in-domain retrieval. AR-LLMs have better interpolation performance but usually fail catastrophically out-domain. Defining "in-out domain" for any task isn't easy as it looks though. 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 22/11/2024Sincerely support this, diverse perspectives from which I can learn and no recsys with a centralized agenda... 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 21/11/2024Starter packs were very instrumental in getting the conversation going. But is there any way we can save a starter pack as a list @support.bsky.team? If not would really appreciate if we could. 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 21/11/2024When the ancients, whether rationalists or mystics, urged 'Know thyself,' they were pointing to our faculty of reason and cognition. Just started and it feels like it is going to be a classic. Grateful for the dopamine hits with intellectual pleasure... 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 21/11/2024Not sure if we can call it smarter but certainly with less confidence in priors leading to more exploration with less risk aversion. 000 0
Bahaeddin ERAVCI @beravci.bsky.social · 20/11/2024I know where this goes :) We'll soon be sharing bluesky dotfiles for config on GitHub. 160 6
Bahaeddin ERAVCI @beravci.bsky.social · 20/11/2024Thank you Costa, saved the day. Working on multimodal BTW... 010 1
Bahaeddin ERAVCI @beravci.bsky.social · 20/11/2024Spinning up #Bluesky to increase signal to noise ratio. I do hope their recommendation system doesn't mess up @bsky.app. Frankly **anything** is probably better than X right now. 010 1