On-policy learning is central to post-training: training on a model's own generations limits distribution shift and helps preserve existing capabilities. On hard problems, however, on-policy RL provides no reward-based learning signal when all attempts fail. On-policy distillation provides next-token supervision from a...
Shripad V. Deshmukh, Christian Bock, Kangrui Ruan et al.· 0 citations
Vector-Borda Empirical-Gap Elimination (VB-EGE) is presented, a fixed-confidence adaptive sampling algorithm for efficiently identifying models on the Pareto frontier in arena-style evaluation and established a worst-case lower bound showing that VB-EGE is minimax-optimal up to logarithmic factors.
Solving reasoning tasks supported on graphs typically requires providing a Language Reasoning Model (LRM) with a serialized description of the graph. This approach results in the input length growing with the number of edges, which limits the size of the problems that can be solved. Furthermore, even when graphs fit in...
Javier Porras-Valenzuela, Arushi Arora, Alejandro Ribeiro· 0 citations
This work introduces $MLIPS: a training-free method that shares candidate search while retaining each consumer's own scoring and selection, and characterize this shared working-set structure and use it to motivate Top-$G, an exploratory bounded pool construction for prefill.
Samkit Jain, Saikishan Pampana, Shlok Agarwal et al.· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
Mixture-of-experts (MoE) models already make a sparse choice for every token: which experts should process it. We ask whether a sparse dictionary should reuse that choice. NR-SAE gives each routed expert a small dictionary and opens only those the frozen router selects. At block 15 of Gemma 4 26B-A4B, after ten million...
We study trail-weighted pruning of evidence for Bayesian-network prediction. A traversal enumerates the active trails from each evidence node to the outcome, scores each trail by the product of its edge strengths, and drops evidence whose score falls below a threshold. For collider-free networks (chains and forks) over...
The predictive closure characterization is proposed, which accounts for the relationships between features and shows that unpredictable features can become better represented when they help predict other features, and that changing one feature's predictive relationships can affect how well others are represented.
Mosh Levy, Aly Lidayan, Yoav Goldberg· 0 citations
Byzantine-robust aggregation shields federated learning from malicious updates, but the rules that filter out suspect updates do so by comparing clients' parameter vectors. That comparison requires access to every client's update, which secure aggregation exists to withhold, costs $O(n^2 d)$ in the parameter dimension...
Saheed A. Tijani, Frank Jiang, Ming Liu· 0 citations
Large language models are deployed globally, yet safety alignment and evaluation remain concentrated in English: harmful requests refused in English can elicit unsafe completions when translated into low-resource languages. We evaluate ten widely deployed LLMs across 79 languages (English plus 78 low-resource) under a...
Jalil Alva, David Harper, Abigail E. Thornton et al.· 0 citations
This work presents VPFlow, a value-predictive planner that pairs the world models with a one-step flow policy and plans in its noise space via evolutionary search, and demonstrates that value-predictive representations, not exact state prediction, are what offline planning requires.
Preference labels align model training, but the judgments behind them are difficult to inspect or verify. In this paper, we examine how preference construction can be checked when intermediate decisions are recorded, and which of them remain uncertain. We present Symbolic Preference Labeling, a method in which a pretra...
Modern language models are increasingly asked to handle several tasks within a single prompt. Doing so requires solving each task correctly while integrating the resulting answers into a coherent response. In this paper, we identify a fundamental limitation in this process, where models make substantially more errors w...
Chenpan Li, Thies Brockmoeller, Stefanie Jegelka et al.· 0 citations
A machine-learning approach called CheMeleon provides a practical way to develop new drugs or other molecules with much less detailed, expensive experimentation.
Research Manager Mike Walker reflects on a career spanning hacker culture, cybersecurity, and AI, sharing lessons about trust, teamwork, leadership, and why solving difficult problems requires curiosity, judgment, agility—and an understanding of the rules. The post Learn the rules: The hacker mindset appeared first on Microsoft Research.
Learn how Google Antigravity and our Cheminformatics Skills accelerate drug discovery workflows, search compounds, and propose novel rigidified macrocycles.
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
