Institution profile

University of California, Irvine

Academic institutionnorthamerica · us
Official website
Research library921linked papers
Opportunities0open roles
Selected work

Representative Papers

LLM Multi-Agent Systems: Challenges and Open Problems

Feb 05, 2024arXiv.org

This work systematically identifies and addresses four open challenges in large language model (LLM)-driven multi-agent systems: inefficient dynamic task allocation, insufficient robustness in collaborative reasoning, difficulty in hierarchical context modeling, and weak long-range memory coordination. To tackle these, we propose a novel architecture integrating iterative debate mechanisms, hierarchical context encoding, memory-augmented retrieval, and blockchain-based smart contract integration. We establish the first comprehensive challenge taxonomy covering collaborative reasoning, dynamic context modeling, and cross-layer memory coordination—distilling six fundamental unsolved problems. Furthermore, we introduce the first verifiable, scalable, and interpretable LLM multi-agent paradigm tailored to real-world distributed environments (e.g., blockchain systems). Our framework unifies theoretical advancement and practical deployment, providing a principled roadmap for both research and engineering.

36 citations1 influentialRead paper

Structural Nested Mean Models Under Parallel Trends Assumptions

Apr 21, 2022

This paper addresses the disconnect between structural nested mean models (SNMMs) and dynamic difference-in-differences (DiD) in estimating time-varying treatment effects. We propose a novel SNMM framework grounded in the parallel trends assumption—departing from the conventional no-unmeasured-confounding assumption. We establish, for the first time, that SNMMs achieve nonparametric identification under parallel trends alone. The framework unifies estimation of dynamic treatment effects, sustained-intervention effects, direct effect decomposition, and optimal dynamic treatment regimes. Additionally, we develop a sensitivity analysis method to assess robustness when parallel trends are violated. Integrating dynamic causal inference with sequential decision-making modeling, our approach is validated through empirical applications—including Medicaid expansion, flood insurance adoption, and temperature impacts on crop yields—demonstrating its validity and robustness in real-world policy and environmental settings.

9 citations2 influentialRead paper

On Configuring a Hierarchy of Storage Media in the Age of NVM

Apr 16, 2018IEEE International Conference on Data Engineering

This work addresses the joint optimization of media selection, capacity allocation, and data placement (replication vs. tiering) for key-value caching across heterogeneous NVM/DRAM/disk storage under memory budget constraints. We introduce the first systematic modeling framework for multi-level non-volatile cache configurations, analytically characterize the operational regimes where replication or tiering dominates, and propose an adaptive configuration policy grounded in device failure rates and data update frequencies. Our methodology integrates cache access behavior modeling, hierarchical configuration optimization, and empirical validation using memcached benchmarks. Results demonstrate that tiering substantially outperforms replication under low device failure rates and high update workloads. Key contributions include: (1) a deployable, low-overhead configuration algorithm; (2) quantitative design guidelines for heterogeneous cache deployment; and (3) theoretical foundations for the reliability–performance trade-off in tiered caching systems.

8 citationsRead paper

Conformal prediction after efficiency-oriented model selection

Aug 13, 2024

Model selection and conformal prediction jointly optimized under limited calibration data face a fundamental trade-off: reusing the validation set for both tasks risks invalidating coverage guarantees. This challenge is to simultaneously select the optimal model and construct narrow yet reliable prediction sets while preserving nominal coverage. Method: We propose a novel conformal prediction framework that avoids additional data splitting. It integrates continuity analysis of model selection, bias-correction mechanisms, and statistical generalization bound derivation. Contributions/Results: The method ensures strict nominal coverage under finite samples and achieves asymptotically optimal prediction interval width. It balances theoretical rigor with computational efficiency. Experiments on synthetic and real-world datasets demonstrate significant interval reduction (12–28% on average) while stably maintaining target coverage, validating its robustness and practicality in low-data regimes.

6 citationsRead paper

An Integrated Framework for Contextual Personalized LLM-Based Food Recommendation

Apr 25, 2025

Existing food recommendation systems suffer from fragmented component design, poor generalization under massive imbalanced data, and insufficient domain adaptation of general-purpose large language models (LLMs). To address these limitations, we propose Food-RLP—a novel context-aware food recommendation paradigm integrating multimodal food logging, geospatial modeling, and domain-specific LLM fine-tuning. We construct a multimedia food logging platform and the World Food Atlas to enable fine-grained, geography-aware food representation. Compared to generic RLP approaches, Food-RLP significantly improves recommendation accuracy and interpretability, enabling context-sensitive, personalized, and cross-regional dietary recommendations in real-world settings. Key innovations include: (1) a food-domain-specific architectural design; (2) joint geospatial–nutritional modeling; and (3) food-semantic-enhanced LLM adaptation mechanisms.

4 citationsRead paper
Recent publications

Latest Papers