Track new results across pure and applied mathematics.
Search papers, labs, and topics across Lattice.
Salesforce, Stanford University
6
0
7
68
Language supervision slashes visual generation error by 14 points, outperforming pixel-only losses.
Video LLMs can significantly improve their QA performance by integrating spatio-temporal evidence, bridging the gap between accuracy and visual perception.
Unlock linear-time video VLMs without accuracy loss: StateKV matches full self-attention while crushing sliding-window methods, all without finetuning.
Training generative models just got a whole lot easier: GPIC offers 100M permissively licensed, captioned, and safety-filtered images.
The report reveals that as AI capabilities surge, our systems for governance and evaluation are lagging dangerously behind, raising urgent questions about management and oversight.
A unified Vision-Language Model and Diffusion architecture unlocks surprisingly effective optical flow forecasting from noisy web data, enabling language-conditioned robot control and video generation.