Eugene Yujun Fu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Eugene Yujun Fu has a total of 78 co-authors. Here are the top co-authors by number of shared publications:
- Grace Ngai
The Hong Kong Polytechnic University
16 shared publications - Youqian Zhang
The Hong Kong Polytechnic University, Hong Kong, Hong Kong
11 shared publications - Hong Va Leong
The Hong Kong Polytechnic University, Hong Kong, Hong Kong
8 shared publications - Mk Michael Cheung
Hong Kong Polytechnic University, Hong Kong, China
7 shared publications - Xiapu Luo
The Hong Kong Polytechnic University, Hong Kong, China
6 shared publications - Chi-Wai Do
School of Optometry, The Hong Kong Polytechnic University, Hong Kong, China
4 shared publications
Latest Publications
- Open Access
- 6 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 5304-5315
SAERec: Constructing Fine-grained Interpretable Intent Priors via Sparse Autoencoders for Recommendation
- 6 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 5304-5315
Open AccessSAERec: Constructing Fine-grained Interpretable Intent Priors via Sparse Autoencoders for Recommendation
- Open Access
- 2 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-13
Seeing Beyond the Chart: Behaviour-Aware Self-Directed Visual Acuity Screening with Multimodal Interaction
- 2 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-13
Open AccessSeeing Beyond the Chart: Behaviour-Aware Self-Directed Visual Acuity Screening with Multimodal Interaction
- 29 October 2026
- Published by Springer Nature
- p. 309-338
Lightweight Detection of Electromagnetic Signal Injection Attacks on Image Sensors
- 29 October 2026
- Published by Springer Nature
- p. 309-338
Lightweight Detection of Electromagnetic Signal Injection Attacks on Image Sensors
- 19 October 2026
- Published by Springer Nature
- p. 616-631
The Access Fallacy: Why Naive Comparisons Overstate the Value of Course Chatbots
- 19 October 2026
- Published by Springer Nature
- p. 616-631
The Access Fallacy: Why Naive Comparisons Overstate the Value of Course Chatbots
- 9 October 2026
- Published by Springer Nature
- p. 464-484
A Simulation Framework for Electromagnetic Signal Injection Attacks on Image Sensors
- 9 October 2026
- Published by Springer Nature
- p. 464-484
A Simulation Framework for Electromagnetic Signal Injection Attacks on Image Sensors
- Open Access
Sex differences in pediatric EEG inter-subject correlation during naturalistic movie watching: a large-scale characterization of the Healthy Brain Network EEG dataset
Inter-subject correlation (ISC) of EEG during naturalistic viewing is increasingly used as a candidate pediatric biomarker, yet its behavior at cohort scale remains uncharacterized. We characterize ISC in 1143 children and adolescents (765 male, 378 female, ages 5 to 21) from the Healthy Brain Network EEG dataset viewing four naturalistic film clips. Three results follow. First, ISC separates narrative from abstract stimuli by 1.7 to 2.4 fold and exceeds a resting-state baseline by 42 to 99 fold, and the previously reported developmental decline replicates throughout. Second, ISC is sex-dependent: males exceed females in every movie (t = 9.7 to 13.5; Cohen's d = 0.61 to 0.85), concentrated in delta and theta, peaking at a frontocentral cluster and at ages 11 to 14. The direction matches an earlier report where the effect was marginal, and we resolve structure that sample could not. It survives three demographic and clinical robustness analyses, same-sex and size-matched templates, ocular component removal, and exclusion of the releases overlapping that report, and is attenuated, by approximately 10 percent adjusting for interpolation count and by 23 percent in recordings free of cluster interpolation, but not explained by differential channel interpolation. Evoked response magnitude, previously proposed as an alternative account, does differ by sex here (d = 0.55 to 0.60) but mediates only 8.6 to 18.1 percent of the effect. Third, ISC predicts none of four CBCL bifactor dimensions after FDR correction at any feature resolution: zero of 16 grand-mean, 2064 per-channel and 16 cluster tests, with confidence intervals excluding standardized associations beyond |β| = 0.106.
Open AccessSex differences in pediatric EEG inter-subject correlation during naturalistic movie watching: a large-scale characterization of the Healthy Brain Network EEG dataset
Inter-subject correlation (ISC) of EEG during naturalistic viewing is increasingly used as a candidate pediatric biomarker, yet its behavior at cohort scale remains uncharacterized. We characterize ISC in 1143 children and adolescents (765 male, 378 female, ages 5 to 21) from the Healthy Brain Network EEG dataset viewing four naturalistic film clips. Three results follow. First, ISC separates narrative from abstract stimuli by 1.7 to 2.4 fold and exceeds a resting-state baseline by 42 to 99 fold, and the previously reported developmental decline replicates throughout. Second, ISC is sex-dependent: males exceed females in every movie (t = 9.7 to 13.5; Cohen's d = 0.61 to 0.85), concentrated in delta and theta, peaking at a frontocentral cluster and at ages 11 to 14. The direction matches an earlier report where the effect was marginal, and we resolve structure that sample could not. It survives three demographic and clinical robustness analyses, same-sex and size-matched templates, ocular component removal, and exclusion of the releases overlapping that report, and is attenuated, by approximately 10 percent adjusting for interpolation count and by 23 percent in recordings free of cluster interpolation, but not explained by differential channel interpolation. Evoked response magnitude, previously proposed as an alternative account, does differ by sex here (d = 0.55 to 0.60) but mediates only 8.6 to 18.1 percent of the effect. Third, ISC predicts none of four CBCL bifactor dimensions after FDR correction at any feature resolution: zero of 16 grand-mean, 2064 per-channel and 16 cluster tests, with confidence intervals excluding standardized associations beyond |β| = 0.106.
- Open Access
- 26 July 2026
- Published by Springer Nature in Journal on Image and Video Processing
- Vol. 2026 (1)
Beyond landmarks: modeling backchannel agreement via reconstruction-based holistic facial motion representations
- 26 July 2026
- Published by Springer Nature in Journal on Image and Video Processing
- Vol. 2026 (1)
Open AccessBeyond landmarks: modeling backchannel agreement via reconstruction-based holistic facial motion representations
- 24 February 2026
- Published by Springer Nature
- p. 415-434
Using Topic Detection to Analyze Student Reflections in Service-Learning: A Text Mining Method for Understanding Learning Gains
- 24 February 2026
- Published by Springer Nature
- p. 415-434
Using Topic Detection to Analyze Student Reflections in Service-Learning: A Text Mining Method for Understanding Learning Gains
- 24 February 2026
- Published by Springer Nature
- p. 503-520
Multimodal Learning Analytics for Predicting Learning Gains in Online Service-Learning Programs
- 24 February 2026
- Published by Springer Nature
- p. 503-520
Multimodal Learning Analytics for Predicting Learning Gains in Online Service-Learning Programs