MIST separates scene setup from change execution to isolate memory failures in video world models, and M3-Agent improves ...
🤖 GroundingPI: What foundation model does embodied AI need? 4B parameters. SOTA overall across 34 grounding benchmarks, surpassing GPT-6 Astra! By training precise perception as a native capability, ...
Author here. WebFovea is a vision-based web agent that placed 2nd in the WebRetriever Challenge 2026 (57.0/100). It works on live websites through their own UI: screenshots in, clicks and keystrokes ...
Remote KV caching for long-context LLM serving can severely degrade latency unless the store is sharded by parallel rank, ...
OneWM unifies causal training of world and action modeling, showing that learning action-conditioned dynamics improves ...
Velox accelerates long video rollouts by routing attention via geometric frustum slabs, reducing latency while preserving ...
The method converts pre-trained transformers into elastic models that adapt continuously to varying computational budgets by ...
Timestral is a family of time series foundation models for zero-shot forecasting, available in 5M, 28M, 123M, and 209M parameter versions. This repository contains the 209M checkpoint. Timestral ...
Abstract MAI-Gym is an open platform that unifies GUI agent evaluation, data collection, and online reinforcement learning training across mobile, computer, and web environments, enabling scalable ...
Timestral is a family of time series foundation models for zero-shot forecasting, available in 5M, 28M, 123M, and 209M parameter versions. This repository contains the 28M checkpoint. For more ...
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
Join the discussion on this paper page You Changed Your Mind, The Model Didn't: Demystifying Intent in Multi-Turn Dialogue ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results