6 papers
Persistent Memory in Multi-Agent LLM Inference: What It Costs, What It Buys, and When You Can Tell
Hochan Son, Kyungdoe Han, Jaehan Koh +3
Decomposing long-context inference across cooperating agents bounds the active KV cache per call rather than total evidence, which matters when KV-cache memory binds. Many such sys…
Prediction-Powered Conditional Inference
Yang Sui, Jin Zhou, Hua Zhou +1
We study prediction-powered conditional inference in the setting where labeled data are scarce, unlabeled covariates are abundant, and a black-box machine-learning predictor is ava…
Uncertainty-Aware Multimodal Learning via Conformal Shapley Intervals
Mathew Chandy, Michael Johnson, Judong Shen +4
Multimodal learning combines information from multiple data modalities to improve predictive performance. However, modalities often contribute unequally and in a data dependent way…
ALIGN: Aligned Delegation with Performance Guarantees for Multi-Agent LLM Reasoning
Tong Zhu, Baiting Chen, Jin Zhou +3
LLMs often underperform on complex reasoning tasks when relying on a single generation-and-selection pipeline. Inference-time ensemble methods can improve performance by sampling d…
Surrogate-Powered Inference: Regularization and Adaptivity
Jianmin Chen, Huiyuan Wang, Thomas Lumley +2
High-quality labeled data are essential for reliable statistical inference, but are often limited by validation costs. While surrogate labels provide cost-effective alternatives, t…
Learn then Decide: A Learning Approach for Designing Data Marketplaces
Yingqi Gao, Wenlu Xu, Jin J. Zhou +3
As data marketplaces become increasingly central to the digital economy, it is crucial to design efficient pricing mechanisms that optimize revenue while ensuring fair and adaptive…