Qianli Ma

This work proposes an exploration-guided prompt scaffolding framework that adapts the training prompt distribution dynamically throughout RL post-training of multimodal large language models (MLLMs), and uses a lightweight rollout-based proxy for prompt utility derived from KL-regularized policy improvement theory.

LightNav-0 achieves state-of-the-art navigation performance by unifying spatial reasoning and action generation in a single model, eliminating the need for task-specific components.

AEGIS can effectively defend against Indirect Prompt Injection attacks while maintaining low latency and high utility, a breakthrough for LLM safety.

By explicitly exposing the model's reasoning process during SVG generation, CTRL-S achieves higher task success rates, superior SVG code quality, and exceptional visual fidelity compared to existing methods.

Ditch the resampling: ReIMTS forecasts irregular time series 27% better by recursively splitting data and preserving original timestamps, unlocking crucial sampling pattern information.

A 1000x larger video reasoning dataset reveals early signs of emergent generalization, offering a new foundation for training and evaluating spatiotemporal AI.

VLMs can be taught to understand the physics of image degradation well enough to control diffusion models for zero-shot image restoration, without fine-tuning the generative backbone.