Search papers, labs, and topics across Lattice.
Zhejiang University
6
0
10
LLMs can dramatically improve MIMO controller tuning by reasoning about complex interactions, achieving optimal performance with far fewer evaluations than traditional methods.
Transferable causal discovery across diverse time series is now possible thanks to a new pretraining paradigm that learns causal structure from synthetic interventions.
Despite advances in AI-driven surgical skill assessment, reliably tracking hands and tools in open surgery videos remains a surprisingly difficult problem, hindering motion-based analysis.
Forget relying solely on numbers: VoT unlocks richer time series forecasts by fusing LLM reasoning over event-related text with multi-level data alignment.
LLMs can now better understand time series data by explicitly modeling trends and seasonality, leading to improved question answering performance.
Forget static agent communication structures: ST-EVO dynamically reshapes multi-agent dialogue flows in both space and time, boosting accuracy by up to 25%.
