Kavli Affiliate: Wei Gao| Summary:Decades of cognitive science establish that humans navigate environments by forming cognitive maps, defined as allocentric and topology-preserving representations of 3D space. While modern Vision-Language Models (VLMs) demonstrate emergent spatial reasoning from 2D egocentric inputs, it remains unclear whether they construct an analogous 3D internal representation. In this paper, we demonstrate […]

Kavli Affiliate: Wei Gao| Summary:Agentic reinforcement learning (RL) has emerged as a key driver for improving the multi-step reasoning and tool-use capabilities of LLMs. However, its efficiency is bottlenecked by long-tail rollouts with multi-turn environment interactions, making static GPU provisioning a poor fit: overprovisioning wastes GPUs on stragglers, while underprovisioning increases contention and slows training. […]

Kavli Affiliate: Joseph Falson | Summary:The response of ultra-thin superconducting materials under parallel magnetic fields is often leveraged to obtain insight into the nature of the condensate, including features attributable to unconventional forms of pairing. Despite there being multiple competing mechanisms responsible for suppressing superconductivity, it is common for these analyses to overlook certain depairing […]

Kavli Affiliate: Wei Gao| Summary:Hybrid parallelism underpins large-scale LLM training across tens of thousands of GPUs. At such scale, hardware failures on individual devices lead to performance skew across devices, diminishing overall training efficiency. Existing resilient systems overlook sequence length variability in datasets and device performance skew under hybrid parallelism. As a result, (1) iteration […]

Kavli Affiliate: Anqi Zhang| Summary:We study the problem of approximating the total variation distance between two mixtures of product distributions over an $n$-dimensional discrete domain. Given two mixtures $mathbbP$ and $mathbbQ$ with $k_1$ and $k_2$ product distributions over $[q]^n$, respectively, we give a randomized algorithm that approximates $d_mathrmTVleft(mathbbP,mathbbQright)$ within a multiplicative error of $(1pm varepsilon)$ […]