Researchers present Entropy-Valley, a training-free technique for selecting target canvas lengths in masked diffusion machine translation that scores candidate lengths…
Alibaba researchers describe DREAM, an agentic control architecture that adds a perception-aware policy layer above existing recommendation pipelines, comprising an…
Researchers investigate how intermediate language artifacts in multi-stage LLM agent workflows can inadvertently weaken operational constraints, evaluating 1,296 synthetic episodes…
Researchers propose MoTE, a decoder architecture that converts large language model feed-forward networks into task-specific experts while keeping the multimodal…
Kai Zhao presents TorchMorph, a lightweight PyTorch extension implementing 22 morphological operators as fused CUDA kernels that operate directly on…
Researchers describe the Station, an autonomous multi-agent mathematical discovery system where AI agents from different model families collaboratively pursue research…
Researchers developed LAWA, a world action model architecture that represents future intentions using compact latent actions rather than generating explicit…
Researchers developed EchoWM, an omnimodal world model that generates synchronized 720p video, environmental sound, music, and speech while responding to…
Researchers developed TLive-Omni, a multimodal model that processes image, video, audio, and text inputs into a unified representation space for…
Researchers introduce ConceptEdit-12M, a 12-million-pair dataset covering more than 1,000 fine-grained edit concepts to improve image editing models, alongside a…
Researchers introduce Block3D, a text-to-3D generation framework that partitions discrete shape-token sequences into blocks, generates them sequentially, and jointly denoises…
Ant Group researchers propose the User Behavioral Densing Law, which quantifies how tokenization capacity should scale with data size for…
