Institution profile

University of California, Riverside

Academic institutionnorthamerica · us
Official website
Research library477linked papers
Opportunities0open roles
Selected work

Representative Papers

The Unpaid Toll: Quantifying the Public Health Impact of AI

Dec 09, 2024arXiv.org

The environmental health impacts and associated environmental injustice of AI’s full lifecycle—from semiconductor manufacturing to datacenter operations—remain poorly quantified. Method: We develop the first integrated AI health impact quantification framework, combining multi-source emission inventories, life cycle assessment, atmospheric transport modeling, health risk assessment, and spatial environmental justice statistics. Contribution/Results: Training Llama3.1 generates PM₂.₅ emissions equivalent to over 10,000 round-trip automobile journeys between Los Angeles and New York. Health burdens exhibit pronounced spatial heterogeneity—up to a 200-fold disparity across U.S. census tracts—disproportionately affecting marginalized communities. By 2030, AI-related datacenter operations in the U.S. are projected to incur annual health costs exceeding $20 billion. We propose mandatory disclosure standards for AI health externalities and a health-centered governance framework to advance equitable, sustainable AI development.

13 citations1 influentialRead paper

MMDeepResearch-Bench: A Benchmark for Multimodal Deep Research Agents

Jan 18, 2026

Existing benchmarks struggle to evaluate models’ ability to utilize evidence in end-to-end multimodal deep research. To address this gap, this work introduces a novel evaluation benchmark comprising 140 expert-designed tasks, each presenting an image–text pair and requiring the model to generate a research report grounded in explicit evidence and consistent across modalities. We propose the first comprehensive evaluation framework tailored for report-style multimodal deep research, featuring three fine-grained assessment modules: FLAE (report quality), TRACE (citation alignment), and MOSAIC (multimodal completeness), enabling diagnostic analysis. Experiments across 25 state-of-the-art models reveal systematic trade-offs among generation quality, citation fidelity, and multimodal grounding, highlighting multimodal completeness as a critical bottleneck.

1 citationsRead paper

Parallel Dynamic Spatial Indexes

Jan 08, 2026ACM SIGPLAN Symposium on Principles & Practice of Parallel Programming

This work addresses the inefficiency of parallel batch updates in highly dynamic spatial data by systematically introducing two novel parallel spatial index structures: the P-Orth tree and the SPaC-tree, built upon the Orth-tree and R-tree/BVH frameworks, respectively. By designing efficient parallel batch update algorithms and index maintenance strategies, the proposed methods significantly outperform existing parallel kd-trees and Orth-trees across diverse dynamic workloads. They achieve substantially lower batch update latency while maintaining comparable or even superior query performance. This study establishes a high-performance, scalable indexing foundation for managing dynamic spatial data.

1 citationsRead paper

Testing Neural Network Verifiers: A Soundness Benchmark with Hidden Counterexamples

Dec 04, 2024arXiv.org

Existing neural network verifiers lack “source-grounded ground truth”—i.e., labels for hard instances that are truly unverifiable and resistant to counterexample discovery—making it difficult to validate claims of solving challenging cases. Method: We introduce the first verifier benchmark designed for source-grounded evaluation, proposing implicit counterexample injection training to embed hidden, semantically valid counterexamples imperceptible to standard adversarial attacks. Our framework enables controllable generation of hard-to-verify instances across architectures, activation functions, input dimensions, and perturbation radii. By integrating adversarial robust training, gradient masking mitigation, and multi-dimensional parametric synthesis, we systematically construct instances exposing fundamental soundness flaws. Contribution/Results: The benchmark uncovers both real and synthetic soundness bugs in multiple state-of-the-art verifiers. All code and datasets are publicly released, establishing a standardized foundation for rigorous, reproducible verifier reliability assessment.

1 citationsRead paper

On Approximability of Satisfiable k-CSPs: V

Aug 27, 2024Electron. Colloquium Comput. Complex.

This work addresses the approximability of Max-CSP, establishing for the first time an algorithmic hardness tightness framework applicable to **all satisfiable k-CSP instances**, thereby overcoming Raghavendra’s restriction to nearly satisfiable instances. We introduce the **mixed invariance principle**, which systematically links third-order correlations in discrete domains to expectations over hybrid Gaussian/Abelian group spaces—a novel connection. Our method combines Gaussian elimination with semidefinite programming to yield a hybrid approximation algorithm and constructs a perfectly complete “dictator vs. pseudorandom” test. For a broad class of predicates, we achieve optimal approximation ratios: the algorithm’s performance exactly matches the Unique Games Conjecture (UGC)-based hardness lower bounds, yielding tightness. This resolves the approximability threshold for these CSPs under UGC, unifying algorithm design and hardness analysis across the full spectrum of satisfiable instances.

1 citationsRead paper
Recent publications

Latest Papers