Hao Feng
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Hao Feng has a total of 21 co-authors. Here are the top co-authors by number of shared publications:
- Chaokun Wang
Tsinghua University
3 shared publications - Yi Yang
Department of Electronic Engineering, Tsinghua University, Beijing 100084, China
2 shared publications - Yi Yang
Department of Electronic Engineering, Tsinghua University, Beijing 100084, China
2 shared publications - Yiming Cheng
Department of Computer Science, University of Chicago, Chicago, IL 60637, USA
2 shared publications - Zhu Han
University of Houston
2 shared publications Ziyang Liu
School of Software, Tsinghua University, Beijing, China
2 shared publications
Latest Publications
- Open Access
- 12 August 2026
- Published by MDPI AG in Electronics
- Vol. 15 (16) , 3591
Large Language Model Agents for Recommender Systems: Bridging Behavior and Semantics with Long-Short Term Interest Modeling
A recommender system extracts user preferences from past interactions to suggest items, widely used in platforms like user-generated content, online shopping, and urban services. These systems aim to provide accurate recommendations, reduce user interaction burden, and enhance user experience while improving socio-economic benefits. Current technologies include content-based, behavior-based, and hybrid recommendations. Behavior-based technologies, like collaborative filtering, are mature due to extensive user–item interaction data but focus on behavior over intrinsic features, lacking comprehensive understanding. Advancements in natural language processing (NLP), text vectorization, and large language models (LLMs) have made content-based recommendations based on semantic understanding more prominent. This study presents an exploratory LLM-agent-based hybrid recommender system framework, building upon a multi-armed bandit (MAB) approach, which enables semantic understanding of item content and integrates both short-term and long-term user interaction behaviors for recommendation. Specifically, it optimizes collaborative filtering models to capture long-term and short-term interests, and uses an agentic LLM memory and reasoning component to identify preference shifts and schedule appropriate recommendation modules. It further vectorizes content using NLP and LLMs to generate retrieval terms that enhance candidate recall, and aligns LLM-optimized content recommendations with user behavior to integrate both aspects. This hybrid system achieved a recall rate improvement of up to 26.63% in behavior recommendation and, in ranking-oriented (diversity-focused) recommendation tasks on the MovieLens dataset, raised NDCG@100 from 0.1428 to 0.3541, a relative improvement of approximately 148% (i.e., about 2.48× the baseline value), albeit against a deliberately simple multi-armed bandit baseline and with a corresponding decrease in Recall@k. However, this work represents an initial exploration in this emerging research direction, and more comprehensive comparisons with state-of-the-art methods are needed in future work to fully assess the framework’s effectiveness.
- 12 August 2026
- Published by MDPI AG in Electronics
- Vol. 15 (16) , 3591
Open AccessLarge Language Model Agents for Recommender Systems: Bridging Behavior and Semantics with Long-Short Term Interest Modeling
A recommender system extracts user preferences from past interactions to suggest items, widely used in platforms like user-generated content, online shopping, and urban services. These systems aim to provide accurate recommendations, reduce user interaction burden, and enhance user experience while improving socio-economic benefits. Current technologies include content-based, behavior-based, and hybrid recommendations. Behavior-based technologies, like collaborative filtering, are mature due to extensive user–item interaction data but focus on behavior over intrinsic features, lacking comprehensive understanding. Advancements in natural language processing (NLP), text vectorization, and large language models (LLMs) have made content-based recommendations based on semantic understanding more prominent. This study presents an exploratory LLM-agent-based hybrid recommender system framework, building upon a multi-armed bandit (MAB) approach, which enables semantic understanding of item content and integrates both short-term and long-term user interaction behaviors for recommendation. Specifically, it optimizes collaborative filtering models to capture long-term and short-term interests, and uses an agentic LLM memory and reasoning component to identify preference shifts and schedule appropriate recommendation modules. It further vectorizes content using NLP and LLMs to generate retrieval terms that enhance candidate recall, and aligns LLM-optimized content recommendations with user behavior to integrate both aspects. This hybrid system achieved a recall rate improvement of up to 26.63% in behavior recommendation and, in ranking-oriented (diversity-focused) recommendation tasks on the MovieLens dataset, raised NDCG@100 from 0.1428 to 0.3541, a relative improvement of approximately 148% (i.e., about 2.48× the baseline value), albeit against a deliberately simple multi-armed bandit baseline and with a corresponding decrease in Recall@k. However, this work represents an initial exploration in this emerging research direction, and more comprehensive comparisons with state-of-the-art methods are needed in future work to fully assess the framework’s effectiveness.
- 3 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Knowledge and Data Engineering
- Vol. 38 (5) , 3234-3249
Training-Free and Unbiased Graph Collaborative Filtering for Personalized Recommendations
With the widespread adoption of collaborative filtering techniques for personalized recommendations, exposure bias has become a significant challenge. Exposure bias refers to the tendency of recommendation models to disproportionately favor items with high exposure over those with low exposure. In graph collaborative filtering that uses graph neural networks (GNNs) for recommendations, exposure bias can be exacerbated due to 1) the reliance on positive feedback during graph construction and 2) the effects of the neighbor aggregation step in GNNs. To tackle this challenge, we propose a novel and efficient framework called FUGCF (training-Free and Unbiased Graph Collaborative Filtering) to improve both the accuracy and bias mitigation of graph-based personalized recommendations. FUGCF employs a two-stage calculation strategy: it estimates exposure probabilities in the first stage and then leverages these exposure probabilities to help derive debiased node embeddings in the second stage. Furthermore, we design a training-free estimation method for FUGCF based on closed-form solutions to enhance its computation efficiency. The extensive experiments on a synthetic dataset and three real-world datasets demonstrate the effectiveness of FUGCF in reducing exposure bias, improving recommendation accuracy, and optimizing computation efficiency.
- 3 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Knowledge and Data Engineering
- Vol. 38 (5) , 3234-3249
Training-Free and Unbiased Graph Collaborative Filtering for Personalized Recommendations
With the widespread adoption of collaborative filtering techniques for personalized recommendations, exposure bias has become a significant challenge. Exposure bias refers to the tendency of recommendation models to disproportionately favor items with high exposure over those with low exposure. In graph collaborative filtering that uses graph neural networks (GNNs) for recommendations, exposure bias can be exacerbated due to 1) the reliance on positive feedback during graph construction and 2) the effects of the neighbor aggregation step in GNNs. To tackle this challenge, we propose a novel and efficient framework called FUGCF (training-Free and Unbiased Graph Collaborative Filtering) to improve both the accuracy and bias mitigation of graph-based personalized recommendations. FUGCF employs a two-stage calculation strategy: it estimates exposure probabilities in the first stage and then leverages these exposure probabilities to help derive debiased node embeddings in the second stage. Furthermore, we design a training-free estimation method for FUGCF based on closed-form solutions to enhance its computation efficiency. The extensive experiments on a synthetic dataset and three real-world datasets demonstrate the effectiveness of FUGCF in reducing exposure bias, improving recommendation accuracy, and optimizing computation efficiency.
- 2 January 2026
- Published by Springer Nature
- p. 123-134
Analogy-Based Learning Approach for Teaching AI Fundamentals to Teens
- 2 January 2026
- Published by Springer Nature
- p. 123-134
Analogy-Based Learning Approach for Teaching AI Fundamentals to Teens
- Top Cited PapersOpen Access
- 27 March 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15 (7) , 3676
Minimizing Hallucinations and Communication Costs: Adversarial Debate and Voting Mechanisms in LLM-Based Multi-Agents
The emergence of large language models (LLMs), such as GPT and Claude, has revolutionized AI by enabling general and domain-specific natural language tasks. However, hallucinations, characterized by false or inaccurate responses, pose serious limitations, particularly in critical fields like medicine and law, where any compromise in reliability can lead to severe consequences. This paper addresses the hallucination issue by proposing a multi-agent LLM framework, incorporating adversarial and voting mechanisms. Specifically, the framework employs repetitive inquiries and error logs to mitigate hallucinations within single LLMs, while adversarial debates and voting mechanisms enable cross-verification among multiple agents, thereby determining when external knowledge retrieval is necessary. Additionally, an entropy compression technique is introduced to enhance communication efficiency by reducing token usage and task completion time. Experimental results demonstrate that the framework significantly improves accuracy, showing a steady increase in composite accuracy across 20 evaluation batches while reducing hallucinations and optimizing task completion time. Notably, the dynamic weighting mechanism effectively prioritized high-performing models, leading to a reduction in error rates and improved consistency in the final responses.
- 27 March 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15 (7) , 3676
Top Cited PapersOpen AccessMinimizing Hallucinations and Communication Costs: Adversarial Debate and Voting Mechanisms in LLM-Based Multi-Agents
The emergence of large language models (LLMs), such as GPT and Claude, has revolutionized AI by enabling general and domain-specific natural language tasks. However, hallucinations, characterized by false or inaccurate responses, pose serious limitations, particularly in critical fields like medicine and law, where any compromise in reliability can lead to severe consequences. This paper addresses the hallucination issue by proposing a multi-agent LLM framework, incorporating adversarial and voting mechanisms. Specifically, the framework employs repetitive inquiries and error logs to mitigate hallucinations within single LLMs, while adversarial debates and voting mechanisms enable cross-verification among multiple agents, thereby determining when external knowledge retrieval is necessary. Additionally, an entropy compression technique is introduced to enhance communication efficiency by reducing token usage and task completion time. Experimental results demonstrate that the framework significantly improves accuracy, showing a steady increase in composite accuracy across 20 evaluation batches while reducing hallucinations and optimizing task completion time. Notably, the dynamic weighting mechanism effectively prioritized high-performing models, leading to a reduction in error rates and improved consistency in the final responses.
- 31 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Knowledge and Data Engineering
- Vol. 36 (12) , 8120-8134
Efficient Unsupervised Graph Embedding With Attributed Graph Reduction and Dual-Level Loss
Graph embedding aims to extract low-dimensional representation vectors, commonly referred to as embeddings, from graph data. The generated embeddings simplify subsequent data analysis and machine learning tasks. Recently, researchers have proposed the use of contrastive learning on graphs to extract node embeddings in an unsupervised manner. Although existing graph contrastive learning methods have significantly advanced this field, there is still potential for further exploration, particularly in optimizing training efficiency and enhancing embedding quality . In this paper, we propose an efficient unsupervised graph embedding method named GEARED. First, the method involves an attributed graph reduction module that converts the raw graph into a reduced graph, greatly improving model training efficiency. Second, GEARED employs a dual-level loss with adaptive scaling factors to ensure the acquisition of high-quality embeddings. Finally, we conduct a partial derivative analysis to elucidate the specific mechanisms through which GEARED is capable of generating high-quality embeddings. Extensive experimental evaluations on 14 benchmark datasets show that GEARED consistently outperforms state-of-the-art methods in terms of training efficiency and classification accuracy. For instance, GEARED achieves a training speedup of over 40 times on both the CS and Physics datasets while maintaining superior classification accuracy.
- 31 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Knowledge and Data Engineering
- Vol. 36 (12) , 8120-8134
Efficient Unsupervised Graph Embedding With Attributed Graph Reduction and Dual-Level Loss
Graph embedding aims to extract low-dimensional representation vectors, commonly referred to as embeddings, from graph data. The generated embeddings simplify subsequent data analysis and machine learning tasks. Recently, researchers have proposed the use of contrastive learning on graphs to extract node embeddings in an unsupervised manner. Although existing graph contrastive learning methods have significantly advanced this field, there is still potential for further exploration, particularly in optimizing training efficiency and enhancing embedding quality . In this paper, we propose an efficient unsupervised graph embedding method named GEARED. First, the method involves an attributed graph reduction module that converts the raw graph into a reduced graph, greatly improving model training efficiency. Second, GEARED employs a dual-level loss with adaptive scaling factors to ensure the acquisition of high-quality embeddings. Finally, we conduct a partial derivative analysis to elucidate the specific mechanisms through which GEARED is capable of generating high-quality embeddings. Extensive experimental evaluations on 14 benchmark datasets show that GEARED consistently outperforms state-of-the-art methods in terms of training efficiency and classification accuracy. For instance, GEARED achieves a training speedup of over 40 times on both the CS and Physics datasets while maintaining superior classification accuracy.
- Open Access
Emotion-Aware Scene Adaptation: A Bandwidth-Efficient Approach for Generating Animated Shorts
Semantic communication technology in the 6G wireless system focuses on semantic extraction in communication, that is, only the inherent meaning of the intention in the information. Existing technologies still have challenges in extracting emotional perception in the information, high compression rates, and privacy leakage due to knowledge sharing in communication. Large-scale generative-model technology could rapidly generate multimodal information according to user requirements. This paper proposes an approach that leverages large-scale generative models to create animated short films that are semantically and emotionally similar to real scenes and characters. The visual content of the data source is converted into text expression through semantic understanding technology; emotional clues from the data source media are added to the text form through reinforcement learning technology; and finally, a large-scale generative model is used to generate visual media, which is consistent with the semantics of the data source. This paper develops a semantic communication process with distinct modules and assesses the enhancements garnered from incorporating an emotion enhancement module. This approach facilitates the expedited generation of broad media forms and volumes according to the user’s intention, thereby enabling the creation of generated multimodal media within applications in the metaverse and in intelligent driving systems.
Open AccessEmotion-Aware Scene Adaptation: A Bandwidth-Efficient Approach for Generating Animated Shorts
Semantic communication technology in the 6G wireless system focuses on semantic extraction in communication, that is, only the inherent meaning of the intention in the information. Existing technologies still have challenges in extracting emotional perception in the information, high compression rates, and privacy leakage due to knowledge sharing in communication. Large-scale generative-model technology could rapidly generate multimodal information according to user requirements. This paper proposes an approach that leverages large-scale generative models to create animated short films that are semantically and emotionally similar to real scenes and characters. The visual content of the data source is converted into text expression through semantic understanding technology; emotional clues from the data source media are added to the text form through reinforcement learning technology; and finally, a large-scale generative model is used to generate visual media, which is consistent with the semantics of the data source. This paper develops a semantic communication process with distinct modules and assesses the enhancements garnered from incorporating an emotion enhancement module. This approach facilitates the expedited generation of broad media forms and volumes according to the user’s intention, thereby enabling the creation of generated multimodal media within applications in the metaverse and in intelligent driving systems.
- 6 January 2023
- Published by Springer Nature in The VLDB Journal
- Vol. 32 (4) , 815-843
Time-topology analysis on temporal graphs
- 6 January 2023
- Published by Springer Nature in The VLDB Journal
- Vol. 32 (4) , 815-843
Time-topology analysis on temporal graphs