Search papers, labs, and topics across Lattice.
8
0
11
Memory consolidation that adapts to temporal relevance could revolutionize how autoregressive video generators maintain coherence over long sequences.
Superficial reasoning in video temporal grounding can be transformed into high-quality, time-aware insights with the right optimization framework.
Tactile feedback can transform robotic manipulation, enabling real-time adaptation that boosts success rates in challenging contact-rich tasks.
WLA models can learn complex tasks from egocentric videos without requiring action annotations, achieving unprecedented success rates in multi-task learning.
AI research agents, despite their potential, currently excel at incremental improvements within established research areas, but struggle to generate truly novel scientific directions.
Overcoming perceptual uncertainty in vision-language navigation is now possible by explicitly modeling geometric, semantic, and appearance uncertainty with a novel Uncertainty-Aware Gaussian Map.
Today's best AI agents can only solve 55% of real-world academic tasks that university students find challenging, revealing a significant gap between current AI capabilities and the demands of academic workflows.
Advanced image editing models may look good but often miss the mark on logical consistency, revealing a critical gap in current AI capabilities.
