Bofan Chen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Bofan Chen has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
- Shilin Wang
Shanghai Jiao Tong University, Shanghai, China
2 shared publications - Hongyu Zhu
School of Computer Science, Shanghai Jiao Tong University, Shanghai, China
1 shared publication - Junshan Jin
Shanghai Jiao Tong University, Shanghai, China
1 shared publication - Lei Yang
School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai, China
1 shared publication - Mingyuan Zhang
Shanghai Jiao Tong University, Shanghai, China
1 shared publication - Sichu Liang
Southeast University, Nanjing, China
1 shared publication
Latest Publications
- 20 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Fuzzy Systems
- Vol. 34 (4) , 1064-1078
Auditing Partial Dataset Usage in Large Language Models via Fuzzy Membership Aggregation
The remarkable capabilities of Large Language Models (LLMs) are fueled by massive internet-scale corpora. However, scraped data owners often do not consent to its use for training, raising significant legal and ethical concerns over copyright and privacy. Data auditing techniques seek to verify whether a protected dataset was used in training a target LLM, typically framing the task as membership inference: estimating binary sample-level membership and aggregating to a dataset-level decision. In this paper, we identify a fundamental limitation of this crisp binary paradigm: in realistic training pipelines, datasets are rarely used in full. Instead, models are trained on mixtures of partial subsets drawn from multiple sources. Existing auditing techniques, built upon an all-or-none assumption—declaring a dataset either entirely present or absent from training—collapse in partial dataset usage scenarios. Their predictions fluctuate unpredictably with the member ratio, causing unstable performance and high false-negative rates. Inspired by fuzzy set theory, we relax the crisp notion of binary membership to a continuous fuzzy membership in [0,1], quantifying each sample's degree of inclusion in the model's training set. We establish a theoretical bridge between sample-level fuzzy memberships and the dataset-level usage ratio, facilitating inference of the proportion of a protected dataset used during training. A neural network fuzzifier first estimates sample-level fuzzy memberships from binary labels in a reference set, then refines them using dataset-level member ratios as higher-order supervision. Finally, a defuzzification stage aggregates calibrated memberships to determine partial usage. Across LLMs of varying scales and multiple auditing datasets, our Fuzzy Auditor substantially outperforms state-of-the-art crisp binary techniques in detecting partial usage, estimating member proportions, and identifying individual member samples.
- 20 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Fuzzy Systems
- Vol. 34 (4) , 1064-1078
Auditing Partial Dataset Usage in Large Language Models via Fuzzy Membership Aggregation
The remarkable capabilities of Large Language Models (LLMs) are fueled by massive internet-scale corpora. However, scraped data owners often do not consent to its use for training, raising significant legal and ethical concerns over copyright and privacy. Data auditing techniques seek to verify whether a protected dataset was used in training a target LLM, typically framing the task as membership inference: estimating binary sample-level membership and aggregating to a dataset-level decision. In this paper, we identify a fundamental limitation of this crisp binary paradigm: in realistic training pipelines, datasets are rarely used in full. Instead, models are trained on mixtures of partial subsets drawn from multiple sources. Existing auditing techniques, built upon an all-or-none assumption—declaring a dataset either entirely present or absent from training—collapse in partial dataset usage scenarios. Their predictions fluctuate unpredictably with the member ratio, causing unstable performance and high false-negative rates. Inspired by fuzzy set theory, we relax the crisp notion of binary membership to a continuous fuzzy membership in [0,1], quantifying each sample's degree of inclusion in the model's training set. We establish a theoretical bridge between sample-level fuzzy memberships and the dataset-level usage ratio, facilitating inference of the proportion of a protected dataset used during training. A neural network fuzzifier first estimates sample-level fuzzy memberships from binary labels in a reference set, then refines them using dataset-level member ratios as higher-order supervision. Finally, a defuzzification stage aggregates calibrated memberships to determine partial usage. Across LLMs of varying scales and multiple auditing datasets, our Fuzzy Auditor substantially outperforms state-of-the-art crisp binary techniques in detecting partial usage, estimating member proportions, and identifying individual member samples.
- Open Access
- 26 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 40-48
Landmark Guided Visual Feature Extractor for Visual Speech Recognition with Limited Resource
- 26 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 40-48
Open AccessLandmark Guided Visual Feature Extractor for Visual Speech Recognition with Limited Resource