DreamActor-M2 is a new way to make a still picture move by copying motion from a video while keeping the character’s look the same.
OCRVerse is a new AI model that can read both plain text in documents and the visual structures in charts, webpages, and science plots, all in one system.
The paper shows how to make AI think faster and smarter by planning in a hidden space instead of writing long step-by-step sentences.
The paper shows a fast, training-free way to boost an LLM’s step-by-step reasoning by smartly reusing the model’s own probabilities.
Hyper-Connections (HC) make the usual single shortcut in neural networks wider by creating several parallel streams and letting the model mix them, but this can become unstable when stacked deep.
The paper shows a simple way to teach AI models what not to learn by removing only the exact words (tokens) related to unwanted topics during pretraining.
ASTRA is a fully automated way to train tool-using AI agents by making both their practice stories (trajectories) and their practice worlds (environments) without humans in the loop.
MemOCR is a new way for AI to remember long histories by turning important notes into a picture with big, bold parts for key facts and tiny parts for details.
ConceptMoE teaches a language model to group easy, similar tokens into bigger ideas called concepts, so it spends more brainpower on the hard parts.
This paper introduces PLaT, a way for AI to think silently in a hidden space (the brain) and only speak when needed (the mouth).
This paper teaches language models to be safer, more factual, and higher quality during pretraining, not just after, by using reinforcement learning with a stronger model as a helper.
Qwen3‑ASR is a family of speech models that hear, understand, and write down speech in 52 languages and dialects, plus they can tell you when each word was spoken.