Latest AI/ML research from ArXiv and Papers with Code
On-policy self-distillation (OPSD) enables large language models (LLMs) to learn from privileged information (PI) without a separate teacher model. However, prior work largely trea...
Large language models (LLMs) are increasingly used in critical domains, where their susceptibility to hallucinations makes reliable and interpretable uncertainty quantification ess...
Reinforcement learning can align visual generators with reward signals beyond their pretraining objectives, but existing score-function methods typically introduce stochasticity by...
Video world models turn actions and camera motion into long visual rollouts, making video a natural interface for interactive simulation. Such interaction requires responsive gener...
Single image 3D generation aims to dress a given 3D mesh with relightable materials while faithfully preserving the visual details of a reference image. Existing methods either int...
World and action modeling are closely related, yet their reciprocal benefits under unified causal training remain unclear. We introduce OneWM, a simple framework for unified causal...
Existing agent benchmarks typically evaluate agents on well-specified tasks defined by issue reports, specifications, or task tickets. Real-world customer delivery is different: fo...
Streaming vision-language models (VLMs) must continuously integrate information from the past while remaining responsive to current observations and user queries. Existing memory m...
We placed four Gemini 4 Argon agents in a sandboxed Minecraft survival world and gave them one goal: defeat the Ender Dragon. Each agent controlled a player via an open-source libr...
AI agents increasingly browse the web on users' behalf, yet they consume webpages as screenshots or text, exposing their LLMs to unvetted third-party content. Browsers enforce sema...
When a speaker says ``some students passed'', listeners infer that not all of them did. Reasoning models now let users decide how long a model thinks before it answers, but we do n...
Harness evolution automatically optimizes the prompts, skills, tools, and orchestration code around a fixed-weight model, offering a way to improve agentic systems without updating...
In a small language model the vocabulary is the single largest part. In Gemma 3 270M the embedding table holds 168M of 268M parameters, and because quantisation methods leave the e...
Post-hoc image explanations usually show where a model looks using spatial heatmaps or indicate which channels matter using global channel scores. These separate views are limiting...
As Vision-Language Models (VLMs) advance, they are increasingly expected to perform spatiotemporal reasoning over motion, actions, and view changes beyond what the input images sho...
Computer use has reached a turning point: the focus is expanding from GUI-only interaction to systems that complete tasks on users' computers through visual and programmatic interf...
Generative simulation can automate robot environment construction, but physically valid environments are not necessarily useful for training. The challenge is to identify environme...
When humans collaborate with AI agents on long-horizon tasks, what happens to the agents' contributions? For solving long-horizon tasks, humans place greater importance on ongoing ...
Retrieval is a fundamental task in multimodal systems: given an image or text query, the goal is to find the most relevant images or texts in a collection, often using a shared emb...
Peer review is under strain from the growth of submissions, and review comments written by humans or AI can misstate what a paper says. We introduce ReviewFact, a benchmark that la...
Unified multimodal models (UMMs) receive image-generation supervision that exposes them to visual information not fully documented in text. Can learning to generate images expand w...
A coding agent works on repositories that can be far larger than its context window can hold. Prompt compressors and retrievers offer solutions to this problem, but patching an iss...
As joint audio-video generation models become more capable, they can generate audio-visual scenes exhibiting three forms of complexity: multi-event composition, multi-layer audio m...
Sparse autoencoders (SAEs) represent model activations with sparse, linearly decoded features, but the downstream effects of those features need not compose additively. We introduc...