Jian Wang

SlimSearcher cuts tool-call rounds by up to 58% without sacrificing accuracy, redefining efficiency in web agent training.

Co-evolving world models with agent policies leads to a 16.75% boost in performance, revolutionizing how language agents navigate complex tasks.

Forget static user profiles – LATTE forecasts where a user's preferences are *going*, not just where they've been, boosting personalized LLM generation.

LLMs struggle to grasp the nuances of cross-cultural aesthetic stylistics, often mistaking surface-level linguistic features for genuine cultural understanding.

Achieve high-fidelity 4D mesh generation from video by cleverly repurposing existing positional encodings for temporal information, sidestepping the usual trade-offs between expressiveness and pretraining compatibility.

On-device LLMs can now drive real-time recommendation improvements, unlocking faster adaptation to evolving user intent without cloud reliance.

SiPeR reveals how integrating scene dynamics with Bayesian inference can dramatically enhance the relevance of conversational recommendations in real-world contexts.

Active belief intervention can drastically improve embodied agents' decision-making by overcoming the pitfalls of belief inertia.

A unified benchmark reveals the trade-offs between pixel-wise accuracy and perceptual realism in state-of-the-art image super-resolution techniques.

LLMs can learn to anticipate their opponents' moves and make better decisions in strategic games by explicitly modeling the other player's behavior during training.

Open-source CPUs can now get near-AMX-level AI acceleration with a matrix extension that's both configurable and decoupled from the core pipeline, slashing design overhead.