Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
7
0
11
Spatial intelligence can be solved 5.6x faster than autoregressive VLMs by treating candidate answers as classes and using native attention matching for direct scoring.
A single MLLM can now perform temporal grounding and video QA with significant inference speedups by treating them as structured decision tasks.
A unified proactive generation framework enables LLMs to form reusable factual memories from streaming video without compromising real-time perception.
TimeLens2 outperforms models with up to 397B parameters by effectively grounding temporal evidence in videos, redefining expectations for multimodal LLM capabilities.
Current MLLMs fail to provide adequate support for visually impaired individuals, particularly in anticipating navigation-critical events in real-time.
VideoChat3 achieves unprecedented generalization in video understanding while maintaining high efficiency, outperforming larger models with just 4 billion parameters.
DOPD reveals that intelligently routing supervision based on advantage gaps can significantly enhance capability transfer in distillation, outperforming conventional methods.