Huimin Zeng
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Huimin Zeng has a total of 21 co-authors. Here are the top co-authors by number of shared publications:
- Dong Wang
University of Illinois Urbana-Champaign, Champaign, IL, USA
13 shared publications - Zhenrui Yue
University of Illinois Urbana-Champaign, Champaign, USA
12 shared publications - Lanyu Shang
Department of Computer Science, Loyola Marymount University, Los Angeles, CA, USA
9 shared publications - Yang Zhang
Miami University, Miami, Florida, USA
8 shared publications - Ruohan Zong
University of Illinois Urbana-Champaign, Champaign, IL, USA
5 shared publications - Julian McAuley
University of California San Diego, La Jolla, USA
3 shared publications
Latest Publications
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 419-427
Variable-Length Audio Fingerprinting
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 419-427
Open AccessVariable-Length Audio Fingerprinting
- Open Access
- 4 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Emerging Topics in Computing
- Vol. 14 (2) , 570-587
ColReg: A Generative Subjective Logic-Driven Crowd-AI Collaborative Regression Framework for Vehicle Damage Estimation
The vehicle damage poses a significant concern in urban transportation and public safety due to its potential to disrupt traffic flow, increase infrastructure maintenance costs, and elevate safety risks for motorists and pedestrians in our community. This paper develops a principled human-AI collaborative regression framework that effectively harnesses the collective intelligence of both AI and crowdsourced human workers to quantify the damage incurred by a vehicle following collisions, accidents, or natural events. We refer to such a problem as the collaborative vehicle damage estimation (VDE) problem. Our work is motivated by the complementary strengths of AI and human intelligence. On one hand, manual vehicle damage inspections, while thorough, are known for being time-consuming and prone to human errors. On the other hand, despite AI's efficiency and consistency, it can encounter challenges in providing accurate damage assessments due to its limitations in comprehending the complex context of vehicle damage imagery. To address such challenges, our paper focuses on leveraging crowdsourced human intelligence for contextual understanding and nuanced reasoning to assist AI in addressing the VDE problem. In particular, we develop ColReg, a crowd-AI collaborative regression model that establishes a virtual interaction between the crowd and AI to collectively provide accurate vehicle damage estimation under a principled subjective logic framework. Evaluation results on a real-world VDE application show that ColReg outperforms both AI-regression models and crowd-AI collaboration baselines in accurately estimating vehicle damages across various evaluation scenarios.
- 4 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Emerging Topics in Computing
- Vol. 14 (2) , 570-587
Open AccessColReg: A Generative Subjective Logic-Driven Crowd-AI Collaborative Regression Framework for Vehicle Damage Estimation
The vehicle damage poses a significant concern in urban transportation and public safety due to its potential to disrupt traffic flow, increase infrastructure maintenance costs, and elevate safety risks for motorists and pedestrians in our community. This paper develops a principled human-AI collaborative regression framework that effectively harnesses the collective intelligence of both AI and crowdsourced human workers to quantify the damage incurred by a vehicle following collisions, accidents, or natural events. We refer to such a problem as the collaborative vehicle damage estimation (VDE) problem. Our work is motivated by the complementary strengths of AI and human intelligence. On one hand, manual vehicle damage inspections, while thorough, are known for being time-consuming and prone to human errors. On the other hand, despite AI's efficiency and consistency, it can encounter challenges in providing accurate damage assessments due to its limitations in comprehending the complex context of vehicle damage imagery. To address such challenges, our paper focuses on leveraging crowdsourced human intelligence for contextual understanding and nuanced reasoning to assist AI in addressing the VDE problem. In particular, we develop ColReg, a crowd-AI collaborative regression model that establishes a virtual interaction between the crowd and AI to collectively provide accurate vehicle damage estimation under a principled subjective logic framework. Evaluation results on a real-world VDE application show that ColReg outperforms both AI-regression models and crowd-AI collaboration baselines in accurately estimating vehicle damages across various evaluation scenarios.
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 2728-2739
Red-Teaming Privacy-Protective Perturbations: Blind Face Restoration as an Attack Strategy
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 2728-2739
Open AccessRed-Teaming Privacy-Protective Perturbations: Blind Face Restoration as an Attack Strategy
- Open Access
- 3 August 2025
- Published by Association for Computing Machinery (ACM)
- p. 3692-3703
Preference-Optimized Retrieval and Ranking for Efficient Multimodal Recommendation
- 3 August 2025
- Published by Association for Computing Machinery (ACM)
- p. 3692-3703
Open AccessPreference-Optimized Retrieval and Ranking for Efficient Multimodal Recommendation
- Open Access
- 10 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 395-404
Fair Sequential Recommendation without User Demographics
Much existing literature on fair recommendation (i.e., group fairness) leverages users' demographic attributes (e.g., gender) to develop fair recommendation methods. However, in real-world scenarios, due to privacy concerns and convenience considerations, users may not be willing to share their demographic information with the system, which limits the application of many existing methods. Moreover, sequential recommendation (SR) models achieve state-of-the-art performance compared to traditional collaborative filtering (CF) recommenders, and can represent users solely using user-item interactions (user-free). This leaves a wrong impression that SR models are free from group unfairness by design. In this work, we explore a critical question: how can we build a fair sequential recommendation system without even knowing user demographics? To address this problem, we propose Agnostic FairSeqRec (A-FSR): a model-agnostic and demographic-agnostic debiasing framework for sequential recommendation without requiring users' demographic attributes. Firstly, A-FSR reduces the correlation between the potential stereotypical patterns in the input sequences and final recommendations via Dirichlet neighbor smoothing. Secondly, A-FSR estimates an under-represented group of sequences via a gradient-based heuristic, and implicitly moves training focus towards the under-represented group by minimizing a distributionally robust optimization (DRO) based objective. Results on real-world datasets show that A-FSR achieves significant improvements on group fairness in sequential recommendation, while outperforming other state-of-the-art baselines.
- 10 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 395-404
Open AccessFair Sequential Recommendation without User Demographics
Much existing literature on fair recommendation (i.e., group fairness) leverages users' demographic attributes (e.g., gender) to develop fair recommendation methods. However, in real-world scenarios, due to privacy concerns and convenience considerations, users may not be willing to share their demographic information with the system, which limits the application of many existing methods. Moreover, sequential recommendation (SR) models achieve state-of-the-art performance compared to traditional collaborative filtering (CF) recommenders, and can represent users solely using user-item interactions (user-free). This leaves a wrong impression that SR models are free from group unfairness by design. In this work, we explore a critical question: how can we build a fair sequential recommendation system without even knowing user demographics? To address this problem, we propose Agnostic FairSeqRec (A-FSR): a model-agnostic and demographic-agnostic debiasing framework for sequential recommendation without requiring users' demographic attributes. Firstly, A-FSR reduces the correlation between the potential stereotypical patterns in the input sequences and final recommendations via Dirichlet neighbor smoothing. Secondly, A-FSR estimates an under-represented group of sequences via a gradient-based heuristic, and implicitly moves training focus towards the under-represented group by minimizing a distributionally robust optimization (DRO) based objective. Results on real-world datasets show that A-FSR achieves significant improvements on group fairness in sequential recommendation, while outperforming other state-of-the-art baselines.
- Open Access
- 27 June 2024
- Published by Association for Computing Machinery (ACM)
- p. 63-75
Tripartite Intelligence: Synergizing Deep Neural Network, Large Language Model, and Human Intelligence for Public Health Misinformation Detection (Archival Full Paper)
The threat of rapidly spreading health misinformation through social media during crises like COVID-19 emphasizes the importance of addressing both clear falsehoods and complex misinformation, including conspiracy theories and subtle distortions. This paper designs a novel tripartite collective intelligence approach that integrates deep neural networks (DNNs), large language models (LLMs), and crowdsourced human intelligence (HI) to collaboratively detect complex forms of public health misinformation on social media. Our design is inspired by the collaborative strengths of DNNs, LLMs, and HI, which complement each other. We observe that DNNs efficiently handle large datasets for initial misinformation screening but struggle with complex content and rely on high-quality training data. LLMs enhance misinformation detection with improved language understanding but may sometimes provide eloquent yet factually incorrect explanations, risking misinformation mislabeling. HI provides critical thinking and ethical judgment superior to DNNs and LLMs but is slower and more costly in misinformation detection. In particular, we develop TriIntel, a tripartite collaborative intelligence framework that leverages the collective intelligence of DNNs, LLMs, and HI to tackle the public health information detection problem under a novel few-shot and uncertainty-aware maximum likelihood estimation framework. Evaluation results on a real-world public health misinformation detection application related to COVID-19 show that TriIntel outperforms representative DNNs, LLMs, and human-AI collaboration baselines in accurately detecting public health misinformation under a diverse set of evaluation scenarios.
- 27 June 2024
- Published by Association for Computing Machinery (ACM)
- p. 63-75
Open AccessTripartite Intelligence: Synergizing Deep Neural Network, Large Language Model, and Human Intelligence for Public Health Misinformation Detection (Archival Full Paper)
The threat of rapidly spreading health misinformation through social media during crises like COVID-19 emphasizes the importance of addressing both clear falsehoods and complex misinformation, including conspiracy theories and subtle distortions. This paper designs a novel tripartite collective intelligence approach that integrates deep neural networks (DNNs), large language models (LLMs), and crowdsourced human intelligence (HI) to collaboratively detect complex forms of public health misinformation on social media. Our design is inspired by the collaborative strengths of DNNs, LLMs, and HI, which complement each other. We observe that DNNs efficiently handle large datasets for initial misinformation screening but struggle with complex content and rely on high-quality training data. LLMs enhance misinformation detection with improved language understanding but may sometimes provide eloquent yet factually incorrect explanations, risking misinformation mislabeling. HI provides critical thinking and ethical judgment superior to DNNs and LLMs but is slower and more costly in misinformation detection. In particular, we develop TriIntel, a tripartite collaborative intelligence framework that leverages the collective intelligence of DNNs, LLMs, and HI to tackle the public health information detection problem under a novel few-shot and uncertainty-aware maximum likelihood estimation framework. Evaluation results on a real-world public health misinformation detection application related to COVID-19 show that TriIntel outperforms representative DNNs, LLMs, and human-AI collaboration baselines in accurately detecting public health misinformation under a diverse set of evaluation scenarios.
- Open Access
- 13 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 4653-4663
MMAdapt: A Knowledge-guided Multi-source Multi-class Domain Adaptive Framework for Early Health Misinformation Detection
This paper studies a critical problem of emergent health misinformation detection, aiming to mitigate the spread of misinformation in emergent health domains to support well-informed healthcare decisions towards a Web for good health. Our work is motivated by the lack of timely resources (e.g., medical knowledge, annotated data) during the initial phases of an emergent health event or topic. In this paper, we develop a multi-source domain adaptive framework that jointly exploits medical knowledge and annotated data from different high-resource source domains (e.g., cancer, COVID-19) to detect misleading posts in an emergent target domain (e.g., mpox, polio). Two important challenges exist in developing our solution: 1) how to accurately detect the partially misleading and unverifiable content in an emergent target domain? 2) How to identify the conflicting knowledge facts from different source domains to accurately detect emergent misinformation in the target domain? To address these challenges, we develop MMAdapt, a multi-source multi-class domain adaptive misinformation detection framework that effectively explores diverse knowledge facts from different source domains to accurately detect not only the outright misleading but also the partially misleading or unverifiable posts on the Web. Extensive experimental results on four real-world misinformation datasets demonstrate that MMAdapt substantially outperforms state-of-the-art baselines in accurately detecting misinformation in an emergent health domain.
- 13 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 4653-4663
Open AccessMMAdapt: A Knowledge-guided Multi-source Multi-class Domain Adaptive Framework for Early Health Misinformation Detection
This paper studies a critical problem of emergent health misinformation detection, aiming to mitigate the spread of misinformation in emergent health domains to support well-informed healthcare decisions towards a Web for good health. Our work is motivated by the lack of timely resources (e.g., medical knowledge, annotated data) during the initial phases of an emergent health event or topic. In this paper, we develop a multi-source domain adaptive framework that jointly exploits medical knowledge and annotated data from different high-resource source domains (e.g., cancer, COVID-19) to detect misleading posts in an emergent target domain (e.g., mpox, polio). Two important challenges exist in developing our solution: 1) how to accurately detect the partially misleading and unverifiable content in an emergent target domain? 2) How to identify the conflicting knowledge facts from different source domains to accurately detect emergent misinformation in the target domain? To address these challenges, we develop MMAdapt, a multi-source multi-class domain adaptive misinformation detection framework that effectively explores diverse knowledge facts from different source domains to accurately detect not only the outright misleading but also the partially misleading or unverifiable posts on the Web. Extensive experimental results on four real-world misinformation datasets demonstrate that MMAdapt substantially outperforms state-of-the-art baselines in accurately detecting misinformation in an emergent health domain.
- Open Access
- 13 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 2497-2508
SymLearn: A Symbiotic Crowd-AI Collective Learning Framework to Web-based Healthcare Policy Adherence Assessment
- 13 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 2497-2508
Open AccessSymLearn: A Symbiotic Crowd-AI Collective Learning Framework to Web-based Healthcare Policy Adherence Assessment
- Top Cited PapersOpen Access
- 4 March 2024
- Published by Association for Computing Machinery (ACM)
- p. 930-938
Linear Recurrent Units for Sequential Recommendation
State-of-the-art sequential recommendation relies heavily on self-attention-based recommender models. Yet such models are computationally expensive and often too slow for real-time recommendation. Furthermore, the self-attention operation is performed at a sequence-level, thereby making low-cost incremental inference challenging. Inspired by recent advances in efficient language modeling, we propose linear recurrent units for sequential recommendation (LRURec). Similar to recurrent neural networks, LRURec offers rapid inference and can achieve incremental inference on sequential inputs. By decomposing the linear recurrence operation and designing recursive parallelization in our framework, LRURec provides the additional benefits of reduced model size and parallelizable training. Moreover, we optimize the architecture of LRURec by implementing a series of modifications to address the lack of non-linearity and improve training dynamics. To validate the effectiveness of our proposed LRURec, we conduct extensive experiments on multiple real-world datasets and compare its performance against state-of-the-art sequential recommenders. Experimental results demonstrate the effectiveness of LRURec, which consistently outperforms baselines by a significant margin. Results also highlight the efficiency of LRURec with our parallelized training paradigm and fast inference on long sequences, showing its potential to further enhance user experience in sequential recommendation.
- 4 March 2024
- Published by Association for Computing Machinery (ACM)
- p. 930-938
Top Cited PapersOpen AccessLinear Recurrent Units for Sequential Recommendation
State-of-the-art sequential recommendation relies heavily on self-attention-based recommender models. Yet such models are computationally expensive and often too slow for real-time recommendation. Furthermore, the self-attention operation is performed at a sequence-level, thereby making low-cost incremental inference challenging. Inspired by recent advances in efficient language modeling, we propose linear recurrent units for sequential recommendation (LRURec). Similar to recurrent neural networks, LRURec offers rapid inference and can achieve incremental inference on sequential inputs. By decomposing the linear recurrence operation and designing recursive parallelization in our framework, LRURec provides the additional benefits of reduced model size and parallelizable training. Moreover, we optimize the architecture of LRURec by implementing a series of modifications to address the lack of non-linearity and improve training dynamics. To validate the effectiveness of our proposed LRURec, we conduct extensive experiments on multiple real-world datasets and compare its performance against state-of-the-art sequential recommenders. Experimental results demonstrate the effectiveness of LRURec, which consistently outperforms baselines by a significant margin. Results also highlight the efficiency of LRURec with our parallelized training paradigm and fast inference on long sequences, showing its potential to further enhance user experience in sequential recommendation.
- Open Access
- 26 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Emerging Topics in Computing
- Vol. 12 (4) , 1105-1116
Unsupervised Domain Adaptation via Contrastive Adversarial Domain Mixup: A Case Study on COVID-19
Training large deep learning (DL) models with high performance for natural language downstream tasks usually requires rich-labeled data. However, in a real-world application of COVID-19 information service (e.g., misinformation detection, question answering), a fundamental challenge is the lack of the labeled COVID data to enable supervised end-to-end training of the models for different downstream tasks, especially at the early stage of the pandemic. To address this challenge, we propose an unsupervised domain adaptation framework using contrastive learning and adversarial domain mixup to transfer the knowledge from an existing source data domain to the target COVID-19 data domain. In particular, to bridge the gap between the source domain and the target domain, our method reduces a radial basis function (RBF) based discrepancy between these two domains. Moreover, we leverage the power of domain adversarial examples to establish an intermediate domain mixup, where the latent representations of the input text from both domains could be mixed during the training process. In this paper, we focus on two prevailing downstream tasks in mining COVID-19 text data: COVID-19 misinformation detection and COVID-19 news question answering. Extensive domain adaptation experiments on multiple real-world datasets suggest that our method can effectively adapt misinformation detection and question answering systems to the unseen COVID-19 target domain with significant improvements compared to the state-of-the-art baselines.
- 26 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Emerging Topics in Computing
- Vol. 12 (4) , 1105-1116
Open AccessUnsupervised Domain Adaptation via Contrastive Adversarial Domain Mixup: A Case Study on COVID-19
Training large deep learning (DL) models with high performance for natural language downstream tasks usually requires rich-labeled data. However, in a real-world application of COVID-19 information service (e.g., misinformation detection, question answering), a fundamental challenge is the lack of the labeled COVID data to enable supervised end-to-end training of the models for different downstream tasks, especially at the early stage of the pandemic. To address this challenge, we propose an unsupervised domain adaptation framework using contrastive learning and adversarial domain mixup to transfer the knowledge from an existing source data domain to the target COVID-19 data domain. In particular, to bridge the gap between the source domain and the target domain, our method reduces a radial basis function (RBF) based discrepancy between these two domains. Moreover, we leverage the power of domain adversarial examples to establish an intermediate domain mixup, where the latent representations of the input text from both domains could be mixed during the training process. In this paper, we focus on two prevailing downstream tasks in mining COVID-19 text data: COVID-19 misinformation detection and COVID-19 news question answering. Extensive domain adaptation experiments on multiple real-world datasets suggest that our method can effectively adapt misinformation detection and question answering systems to the unseen COVID-19 target domain with significant improvements compared to the state-of-the-art baselines.