Want to find out more about Verbalized Machine Learning? Come and talk to Weiyang Liu and I at the ICL workshop tomorrow (Saturday)!😊 #ICML2024 Poster session: 10:25 - 11:15 AM 3:20 - 4:15 PM Github🔗: github.com/timxzz/VML_Exa…

🚀 Excited to introduce our new work: SGP-Bench! Can Large Language Models (LLMs) understand symbolic graphics programs? 🖥️ Imagine giving a model a symbolic graphics program like SVG or CAD and asking it to answer questions about the visual content without actually seeing the

How could online learning apps adapt to learners and improve over time? Even if you're not a machine learning expert, Hanqi Zhou's blog post on our last ICLR paper explains new approaches in simple terms (joint work with 🐘 Álvaro Tejero-Cantero @[email protected] and Charley Wu (@thecharleywu.bsky.social), supported by Theresa Authaler).

Biggest joy and honour leading this project at the intersection of visual neuroscience and ML to a successful finish! Paper: nature.com/articles/s4158…

Researching #AI #MachineLearning #Robotics #HCI? Join our elite doctoral program - a partnership with MPI-IS, @uni_stuttgart & Universität Tübingen! Applications accepted until Nov 15, 2024 at imprs.is.mpg.de/application

A Schulz presenting our recently accepted #NeurIPS #spotlight paper on Latent Diffusion for Neural Spiking Data, check out the poster if you're at Bernstein Network Computational Neuroscience! 🧠

I am super hyped and happy with our recent paper on a ✨VampPrior 2.0✨: Hierarchical VAE with a diffusion-based VampPrior 🦇 We got SOTA VAE results! Kudos to Anna Kuzina, this TMLR paper is the last chapter in her PhD thesis 🤩 📃 tinyurl.com/22rvzc4f 💻 tinyurl.com/3ddrbfcj

How to find all fixed points in piece-wise linear recurrent neural networks (RNNs)? A short thread 🧵 In RNNs with N units with ReLU(x-b) activations the phase space is partioned in 2^N regions by hyperplanes at x=b 1/7

1) With our NeurIPS Conference poster happening tomorrow, it's about time to introduce our Spotlight paper 🔦, co-lead with Jaivardhan Kapoor : Latent Diffusion for Neural Spiking data (LDNS), a latent variable model (LVM) which addresses 3 goals simultaneously:

Latent Diffusion for Neural Spiking Data -- a flexible generative model for realistic neural recordings co-led with A Schulz! This was a very fun and intense project, and my first (and hopefully not the last) foray into #neuro! Check out the thread - lots of pretty plots

𝟏/ 🚨Attention NeurIPS attendees! 🚨 Isotropic weight space priors are sabotaging your Laplace approximation to your neural network’s posterior, and we’ve got the fix. Curious? Read on 🧵

I'm extremely disappointed with the organization of the upcoming German election. Voting is a fundamental right, yet my ballot letter arrived too late in Toronto—now the fastest! delivery won't get it there by Monday, after the election. Unacceptable. Die Bundeswahlleiterin German Consulate TO

✨ New paper: Flipping Against All Odds We found that large language models (LLMs) can describe probabilities—but fail to sample from them faithfully. Yes, even flipping a fair coin is hard. 🪙 🧵 Here’s what we learned—and how we fixed it. 🔗arxiv.org/abs/2506.09998 1/

Great paper by my students Tim Xiao and Johannes Zenn and collaborators that applies ideas from Monte Carlo sampling to (black-box) LLM execution to turn LLMs into better calibrated stochastic samplers.

📢 [Openings] I'm now an Assistant Prof Western University CS dept. Funded PhD & MSc positions available! Topics: large probabilistic models, decision-making under uncertainty, and apps in AI4Science. More on agustinus.kristia.de/openings/

🧠 Two years ago we wrote a blog post exploring an analogy: Large Language Models Are Zero-Shot Problem Solver —Just Like Modern Computers Today, that idea is published on Harvard Data Science Review Harvard Data Science Review ! 🔗 hdsr.mitpress.mit.edu/pub/42vjucmq/

New Recording of Friday Talks @ Tübingen is online! Title: Effortless, Simulation-Efficient Bayesian Inference using Tabular Foundation Models Speakers: Julius Vetter & Manuel Glöckler Manuel Glöckler youtu.be/Wx2p1rrggmM?si…

[FridayTalks@Tübingen] New recording is online! Title: Large Language Models Are Zero-Shot Problem Solvers—Just Like Modern Computers Speaker: Tim Xiao youtu.be/ySHuahiazD4?si…