Tabular foundation models (TFMs) learn to fill in tables the way language models fill in text, and tables are arguably the format in which most physical measurement arrives. Did they learn any physics in the process? They are Bayesian by construction, so the question is what their prior contains. We probe it directly, evaluating four of them (TabPFN-3, TabICLv2, TabDPT and Real-TabPFN-2.5) against six baselines on datasets sampled from 316 physical equations, in and out of domain. TFMs dominate, out of the box and after tuning. But we show that their prior can represent neither a noiseless mechanism nor physical units, which is why they interpolate physics without yet being able to act as physical models.
Interpretability research increasingly asks when concepts emerge during training and whether linear probes recover real structure, but in language models these claims are hard to validate because language offers little ground-truth ordering of concepts or relationships among them. We propose the use of astronomical gro...
UniverseTBD Kshitij Duraphe, Aman Kumar, Michael J. Smith et al.· 0 citations
PhysElite is presented, a large-scale bilingual multimodal benchmark for Olympiad-level physics reasoning that benchmarks 18 open-source and closed-source MLLMs, and finds that even the strongest model reaches only 33.7% answer accuracy.
Ruoran Xu, Wending Gao, Liyunfeng Chen et al.· 1 citation
Machine learning is now a central tool for solving inverse problems in particle physics and astronomy. Models are trained on simulation and deployed on real data, raising the question not just of whether they fit, but of whether they are wrong in ways we did not anticipate: the unknown unknowns. This challenge of model...
J. Cruz-Martinez, C. Cuesta-Lázaro, Alexander Held et al.· 1 citation
The first application of sparse-autoencoder-based mechanistic interpretability to particle physics suggests that mechanistic interpretability can reveal learned latent physics encoded within a model's internal representation and help design downstream tasks that exploit it.
Raphaël Bonnet-Guerrini, Johann Ioannou-Nikolaides, I. Timiryasov et al.· 0 citations
The integration of large language models (LLMs) into scientific workflows is accelerating, yet their ability to reconstruct the reasoning underlying published research remains unexplored. Papers specify explicit procedures while leaving many methodological dependencies-data selection, calibration corrections, priors, a...
Yue-Hui Wang, Xin-Yu Qi, Gui-Rong Xue et al.· 0 citations
Exploring how generative AI could make machine vision more accessible to businesses. The post GenEye in a Box: Making Machine Vision Something You Can Just Ask For appeared first on GPT-Lab.
MIT News · Artificial Intelligence· news.mit.eduOct 7, 2026
Students in MIT’s Concourse program delve deeply into the human condition, debate challenging questions, and learn to develop judgment about issues that can’t be quantified.
Training AI agents with reinforcement learning can be challenging because their tools, context, and decision-making are managed by complex frameworks. Agent Lightning connects existing agents to RL training, making it easier to improve them without rebuilding them. The post Agent Lightning v1.0: A 3,500-Line Lightweight Agentic RL Framework for Training Agents with Real Harnesses appeared first on Microsoft Research.
MIT News · Artificial Intelligence· news.mit.eduOct 6, 2026
