Danyal Ahmed @danyallah_
ml hobbyist Joined February 2024-
Tweets42
-
Followers6
-
Following801
-
Likes116
We are the original scaling lab, scaling up GPs since 2002
Major surprise by 3 @Berkeley_EECS grad students: Meghal Gupta, @HongxunWu, Mihir Singhal. *Deterministic* eps-approximate quantiles in O(1/eps) mem. The previous best was the KLL sketch, which was randomized and used O(lglg(1/p)/eps) mem, p = fail prob arxiv.org/abs/2404.03847 1/
After years of work, we can now move on from GP regression to universal functional regression. OpFlow: Just do GP regression, without Gaussian prior assumption. Learn prior from your data. 🔥🔥 arxiv.org/pdf/2404.02986… Jw, Yaozhong Shi, @af_gao, @zross_ @Caltech @nvidia
I have put online the content of 10 lectures of 2h on the basics maths of AI. It contains raw transcripts, so probably not super useful, but it might be helpful to give some ideas. mathematical-tours.github.io/maths-ia-cours…
I often mention in my talks that we should worry less about nonconvexity and more about non-smoothness. We can get surprisingly far without worrying about all those saddle points, but ignoring non-smoothness quickly leads to methods that don't work in practice.
I often mention in my talks that we should worry less about nonconvexity and more about non-smoothness. We can get surprisingly far without worrying about all those saddle points, but ignoring non-smoothness quickly leads to methods that don't work in practice.
I have n consecutive integers, how many permutations exist such that when moving from left to right there is only one move that goes from higher to lower eg. (1,5,2,3,4) for n = 5 is a valid permutation @littmath. This feels like this should be standard.
Cooking up something special! Can't wait to get a paper out so everyone can try it out. An optimizer with no extra overhead, no additional parameters. Stay tuned!
We live in such strange times. Apple, a company famous for its secrecy, published a paper with staggering amount of details on their multimodal foundation model. Those who are supposed to be open are now wayyy less than Apple. MM1 is a treasure trove of analysis. They discuss…
🎉 Curated list of textbooks on pattern recognition, machine learning, AI and deep learning: links to online PDFs when available by their authors franknielsen.github.io/Books/CuratedB…
With OpenAI, Figure 01 can now have full conversations with people -OpenAI models provide high-level visual and language intelligence -Figure neural networks deliver fast, low-level, dexterous robot actions Everything in this video is a neural network:
This week, @xai will open source Grok
A pleasant surprise to see Stephen Boyd revisiting teaching Optimization online: youtu.be/kV1ru-Inzl4?si… In addition to Gilbert Stang’s series I’m wondering what other people believe to be timeless applied math lecture series on YouTube?
Incredibly excited to announce Hawk and Griffin (arxiv.org/abs/2402.19427), two recurrent language models with 1) finite sized state + fast inference 2) efficient training on device 3) excellent performance:
ชัยวีร @ngRf6JyUt76D4
52 Followers 1K Following คุณต้องการนัดเดทกับสาวไหมคะ เพิ่ม https://t.co/xaRRs1bcpY
Jonathan sillah☦️ @SillahJonathan
140 Followers 3K Following Psalm 40:17 As for me, I am poor and needy, but the Lord takes thought for me. You are my help and my deliverer; do not delay, O my God!
Oussama Zekri @oussamazekri_
59 Followers 276 Following @ENS_ParisSaclay Maths department + MVA and @Huaweifr Research Intern | Interested in Optimization and core Machine Learning.
firstnamebunchofnumbe.. @urusualskeptic
76 Followers 461 Following
Prestige Aesthetics a.. @MyPrestigeAE
14 Followers 220 Following Prestige Aesthetics: Where elegance meets innovation. Discover personalized beauty treatments for radiant results. https://t.co/uzE6MT0swW
Finley Hernandez @FinleyHern82785
69 Followers 3K Following
Mike Dusenberry @dusenberrymw
403 Followers 761 Following Research Engineer @GoogleDeepMind. I work on Bard research. Focus: reliable A(G)I at scale (large neural nets, probability, decisions).
Basil Mustafa @_basilM
1K Followers 130 Following researching ML @ google brain ZRH | no strong opinions about AI, but very strong opinions about why herbal infusions are awful and should not be called teas
Metod Jazbec @MetodJazbec
113 Followers 393 Following Machine Learning PhD student with @eric_nalisnick at @AmlabUva
Shreyas Padhy @shreyaspadhy
260 Followers 553 Following PhD student at the University of Cambridge. Ex @GoogleAI Resident, @jhubme and @iitdelhi. I like the math of machine learning & neuroscience. Also DnD.
Vincent Fortuin @vincefort
2K Followers 601 Following PI @helmholtz_ai, Faculty @TU_Muenchen, Fellow @zuseschoolrelAI, Branco Weiss Fellow. Prev: @CambridgeMLG, @stjohnscam, @ETH_en, @GoogleAI, @MSFTResearch.
Joan Puigcerver @joapuipe
861 Followers 376 Following Software Engineer in Research at Google DeepMind, Zürich.
Carlos Riquelme @rikelhood
2K Followers 2K Following head of language models team @StabilityAI previously researcher @GoogleBrain @Stanford @Oxford & @UAM_Madrid amateur writer at https://t.co/Qa9neIqmg8
rodolphe_jenatton @RJenatton
354 Followers 272 Following
Ulrich Paquet @ulrichpaquet
1K Followers 208 Following Research Scientist, Google DeepMind Director, African Institute of Mathematical Sciences (South Africa) Co-founder, Deep Learning Indaba
Stephan Gouws @sgouws
1K Followers 1K Following AI Researcher, fascinated by communication across all modalities, wine lover, traveller, golfer. Prev: Google, DeepMind.
Avishkar @avishkar58
2K Followers 592 Following Research Engineer at @Deepmind. Strengthen African machine learning through the Deep Learning Indaba @DeepIndaba. All opinions my own.
Michael Aerni @AerniMichael
48 Followers 116 Following AI privacy and security | PhD student @CSatETH | Ask me about coffee ☕️
Ryan Adams @ryan_p_adams
34K Followers 1K Following Machine Learning Researcher, CS Professor (@PrincetonCS), Dad, Woodworker
Stratis Tsirtsis @stratis_
420 Followers 429 Following PhD student @mpi_sws_ @maxplanckpress. Interested in machine learning, decision making, and causality. Previously at @Meta, @ntua.
Christopher Yau @cwcyau
4K Followers 3K Following Senior Research Fellow, University of Oxford. Turing AI Fellow and Director, HDRUK-Turing Wellcome PhD Programme in Health Data Science.
Fanghui Liu @Fanghui_SgrA
175 Followers 219 Following Assistant Prof. @WarwickDCS🇬🇧on kernel methods, learning theory | previously at @EPFL🇨🇭@KU_Leuven🇧🇪 @sjtu1896🇨🇳
Kostas Tsampourakis @KTsampourakis
76 Followers 650 Following Stats PhD @EdinburghUni. Work on Prob ML, focus on SSM, Bayesian state estimation/filtering & learning dynamics. https://t.co/2xjHceuRKq
Sébastien Darses @DarsesSebastien
3K Followers 3K Following Assoc. Prof. (he/him 🏳️🌈), Visiting @CRM_Montreal @CNRS_INSMI @CNRS 📚 Probability, Statistics, Analytic Number Theory 👨🎓Teaching project @highkholle
Sirio Legramanti @SirioLegramanti
708 Followers 519 Following Assistant Professor in #Statistics @UniBergamo Main research interests: #Bayesian inference, dimensionality reduction, #networks, ABC, #privacy
Athanasios (Thanos) V.. @vlontzos
500 Followers 377 Following Research Scientist @SpotifyResearch - Causality in ML | PhD from @BioMedIAICL | prev. @apple, @nasa_fdl, @zeitmedical, @GEHealthcare
FoundationsAISeminars @FAIS_Warwick
14 Followers 13 Following Foundations of AI Seminars @ Warwick | Department of Computer Science | University of Warwick| Organising Team: @parisgiampouras @Fanghui_SgrA @ltt_hal
Theo Damoulas @TDamoulas
45 Followers 145 Following Professor of Machine Learning in CS/Stats @uniofwarwick | UKRI Turing AI Acceleration Fellow | Warwick ML Group https://t.co/Cq197vuidm
Jakub Tětek @JakubTetek
178 Followers 270 Following I do research in theoretical computer science. I'm especially interested in differential privacy and beyond-worst-case analysis.
Matthew Jackson @JacksonMattT
490 Followers 446 Following PhD student @whi_rl & @flair_ox | Incoming RS intern @wayve_ai (GAIA team) | 🎓 MSc @ucl_dark, BA @cambridge_cl | Meta and Model-Based Reinforcement Learning
Louis Kirsch @LouisKirschAI
2K Followers 779 Following PhD at IDSIA with @SchmidhuberAI. Working on self-improving AI that generalizes (MetaGenRL, VSML, GPICL). @DeepMind @GoogleAI intern, @UCL, @HPI_DE alumnus.
Krueger AI Safety Lab @kasl_ai
251 Followers 48 Following We are a research group at the University of Cambridge focused on avoiding catastrophic risks from AI.
George Cantwell @GTCantwell
282 Followers 238 Following Neural network interested in statistics, physics, and computer science. Asst prof @Cambridge_Eng
Subbarao Kambhampati .. @rao2z
16K Followers 29 Following AI researcher & teacher @SCAI_ASU. Works on Human-Aware AI. Former President of @RealAAAI; Chair of @AAAS Sec T. Here to tweach #AI. YouTube Ch: https://t.co/4beUPOmMW6
Karl Tuyls @karl_tuyls
2K Followers 334 Following Ex: team lead @ DeepMind,@GoogleDeepMind - still working on AGI in a Multi-Agent world. CS professor (Liverpool/Leuven) and LFC fan.
Lucia Cipolina Kun @LuciaCKun
1K Followers 4K Following Maths, Reinforcement Learning and Game Theory. NYU Courant, University of Bristol. Alternative account on art restoration: @ArtRestoreAI 🇺🇾
Michael Beukman @mcbeukman
79 Followers 64 Following PhD Student at @FLAIR_Ox, previously at @raillabwits. Interested in Open ended learning, unsupervised environment design, procedural content generation
Aryeh Kontorovich @aryehazan
4K Followers 470 Following probability, statistics, metric spaces, Markov chains, freedom (social & academic), Israel, Jew stuff. opinions represent my employer & all other groups I'm in
Suvrit Sra @optiML
5K Followers 351 Following Currently @MIT. Interested in Optimization, ML, and Math 🙃 On the non-professional side: Punjabi (ਪੰਜਾਬੀ) language, music, and culture. ⬛️
Elias Bareinboim @eliasbareinboim
14K Followers 561 Following Professor of Causal Inference, Machine Learning, and Artificial Intelligence. Director, CausalAI Lab @ Columbia University.
Giovanni S. Alberti @GSAlberti
381 Followers 323 Following Professor of Mathematical Analysis @malga_center, @UniGenova. Previously at ETH Zurich, École Normale Supérieure Paris, Oxford University.
Luigi Carratino @luigicarratino
230 Followers 278 Following Designing fast ML algorithms because life is too short to wait.
Adrien Taylor @TaylorAdrien
340 Followers 164 Following Researcher at @inria_Paris in the @Sierra_ML_Lab team
Alessandro Rudi @AlessandroRudi
407 Followers 147 Following Researcher at INRIA & École Normale Supérieure
Yorgos Felekis @yfelekis
323 Followers 757 Following PhD student in Machine Learning and Causality @uniofwarwick | Member of Warwick Machine Learning Group | 🧪🖥️ | https://t.co/NDLyaDbOXb
Nicola Branchini @Branchini_Nic
616 Followers 2K Following 🇮🇹 3rd yr Stats PhD @EdinUniMaths 🏴.🤔💭 about importance sampling, uncertainty quantification. In search of explanations, simple ideas, clarity
Fanny Yang @FannyYangETH
915 Followers 87 Following Assistant Professor @CSatETH Zurich and @ETH_AI_Center, heading the Statistical Machine Learning group. Postdoc @Stanford. PhD @Berkeley_EECS. Violinist.
Austin Huang @austinvhuang
3K Followers 1K Following General intelligence as personal computing. Past: @GoogleDeepMind, MIT, Harvard, Berkeley.
Karsten Kreis @karsten_kreis
2K Followers 444 Following Senior Research Scientist at @NVIDIA | Former Physicist | Deep Generative Learning. Opinions are my own.
Stan @mcmc_stan
12K Followers 45 Following Making Bayes fast since 2011. Also at https://t.co/5A7QRcd1AJ.
Yu-Xiang Wang @yuxiangw_cs
2K Followers 280 Following Faculty @ucsbcs, director of ML lab. Previously @ScSatCMU and @awscloud. Researcher in #machinelearning, #reinforcementlearning, #differentialprivacy
An elegant and soon-to-be indispensable tool for the modern AI researcher.
Heyo, @francoisfleuret wants to have a platform to post links to arxiv papers with max 200 characters TL;DRs and voting system, so I just built it tldr-ai.org It's an unpolished MVP yet, but still, I can't wait for you to start using it and sharing your feedback!
Missed our last talk with Prof Po-Ling Loh - Differentially private M-estimation via noisy optimization? You can watch it here! #MachineLearning in #AI youtu.be/KYZ2Gt81G1I. .
This line of @PyTorch code fascinates me every time I come across it: y = x_backward + (x_forward - x_backward).detach() As @ThomasViehmann explained: "It get’s you x_forward in the forward, but the derivative will act as if you had x_backward"
If humans and chimpanzees split 6 million years ago, that's roughly only 300,000 generations of grandparents.
@francoisfleuret Only one I don't feel comfortable saying I know the *precise* definition of is VC dimension. But there is a past version of me who did :p Something something shatter, something something number of ways you could fit
@ylecun ...As a (Bayesian) inference problem. :)
Only if academic institutes were less classists and snobist! (among many other things)
I think the hardest thing for me the last few years has been seeing so many talented scientists who obviously belong in the academy turn into tech company middle managers or startup founders.
@francoisfleuret Yeah, constant valued neurons. I think those, exploding/vanishing gradients and so on etc. were important problems that shaped what we do today. Likewise for curse of dimensionality. No Deep RL e.g. REINFORCE, PPO, MCTS etc? ML: kernel smoothing, PCA, (un)supervised learning?
@KLdivergence I explain it to Bayesians as inferring the posterior predictive p(yt|ypast, theta-hat) using a function call. Since ML doesn’t care about inferring unidentified parameters, and latent variables are out of fashion, what else is there to infer?
Michael Jordan giving a talk at KAUST on a good topic 😉
@natolambert AE-DPO (active exploration) is simple enough arxiv.org/pdf/2312.00267…
Snorkelling in the Red Sea with Michael I Jordan. Mike is visiting KAUST this week. cemse.kaust.edu.sa/events/event/a… cemse.kaust.edu.sa/events/event/d…
@PlatoLemma @PlatoLemma Your proof of Exercise 13 in Section 1C of Linear Algebra Done Right is not correct. The error occurs when you conclude that W + X is a subset of U. You have assumed that u + (w + x) is in U for all choices of u in U, w in W, and x in X, but this may hold only for…
@RanaAgastya Well, it’s still search, just likelihood-based. It also has drawbacks - seems prone to reward hacking after some point.
@thegautamkamath I grind for my students. And for the love of science and knowledge:) It's not rational, but I can't help it. I am not sure whether this sound honest, but I really never cared about anything but my students and the joy I get from learning new things and connecting to others
I deanonymized it. Some care needs to be taken to disambiguate names, but I think this is mostly correct up to ties. Levine submitted 22 papers more than half were rejected. A lot of other expected names.
In case some of you were (like me) curious about this stat for AI conferences: here it is for ICLR2024
@avt_im I’m guessing you’ve Not been on a hiring committee.
@jxmnop Things look Linear in the middle of a sigmoid - they look exponential earlier
Trends for United States
You might like
