Dekai Zhang
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
Research Interests
Dekai Zhang has a total of 22 co-authors. Here are the top co-authors by number of shared publications:
- Francesca Toni
Imperial College London, London, United Kingdom
4 shared publications - Adam Dejl
Imperial College London, London, United Kingdom
2 shared publications - Matthew Williams
Affiliations Department of Radiotherapy, Charing Cross Hospital, Imperial College Healthcare NHS Trust, London, United Kingdom, Computational Oncology Laboratory, Institute of Global Health Innovation, Imperial College London, London, United Kingdom
2 shared publications - Adam Dejl
Department of Computing Imperial College London
1 shared publication Amir Sanchez
The ¶Department of Pediatrics, Baylor College of Medicine, Houston, Texas 77030
1 shared publication- Carlos Ignacio Villalobos Sánchez
Department of Computing Department of Computing Department of Computing Department of Computing Department of Computing Department of Computing Department of Computing Department of Computing Imperial College London Imperial College London Imperial College London Imperial College London Imperial College London Imperial College London Imperial College London Imperial College London
1 shared publication
Latest Publications
- Open Access
- 23 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 1498-1542
Hidden Conflicts in Neural Networks and their Implications for Explainability
- 23 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 1498-1542
Open AccessHidden Conflicts in Neural Networks and their Implications for Explainability
- Open Access
- 23 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 2892-2905
XAI-Units: Benchmarking Explainability Methods with Unit Tests
- 23 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 2892-2905
Open AccessXAI-Units: Benchmarking Explainability Methods with Unit Tests
- Open Access
XAI-Units: Benchmarking Explainability Methods with Unit Tests
Feature attribution (FA) methods are widely used in explainable AI (XAI) to help users understand how the inputs of a machine learning model contribute to its outputs. However, different FA models often provide disagreeing importance scores for the same model. In the absence of ground truth or in-depth knowledge about the inner workings of the model, it is often difficult to meaningfully determine which of the different FA methods produce more suitable explanations in different contexts. As a step towards addressing this issue, we introduce the open-source XAI-Units benchmark, specifically designed to evaluate FA methods against diverse types of model behaviours, such as feature interactions, cancellations, and discontinuous outputs. Our benchmark provides a set of paired datasets and models with known internal mechanisms, establishing clear expectations for desirable attribution scores. Accompanied by a suite of built-in evaluation metrics, XAI-Units streamlines systematic experimentation and reveals how FA methods perform against distinct, atomic kinds of model reasoning, similar to unit tests in software engineering. Crucially, by using procedurally generated models tied to synthetic datasets, we pave the way towards an objective and reliable comparison of FA methods.
Open AccessXAI-Units: Benchmarking Explainability Methods with Unit Tests
Feature attribution (FA) methods are widely used in explainable AI (XAI) to help users understand how the inputs of a machine learning model contribute to its outputs. However, different FA models often provide disagreeing importance scores for the same model. In the absence of ground truth or in-depth knowledge about the inner workings of the model, it is often difficult to meaningfully determine which of the different FA methods produce more suitable explanations in different contexts. As a step towards addressing this issue, we introduce the open-source XAI-Units benchmark, specifically designed to evaluate FA methods against diverse types of model behaviours, such as feature interactions, cancellations, and discontinuous outputs. Our benchmark provides a set of paired datasets and models with known internal mechanisms, establishing clear expectations for desirable attribution scores. Accompanied by a suite of built-in evaluation metrics, XAI-Units streamlines systematic experimentation and reveals how FA methods perform against distinct, atomic kinds of model reasoning, similar to unit tests in software engineering. Crucially, by using procedurally generated models tied to synthetic datasets, we pave the way towards an objective and reliable comparison of FA methods.
- 29 November 2022
- Published by Springer Nature
- p. 171-185
A Federated Cox Model with Non-proportional Hazards
- 29 November 2022
- Published by Springer Nature
- p. 171-185
A Federated Cox Model with Non-proportional Hazards
- Open Access
- 1 November 2009
- Published by Elsevier in Journal of Biological Chemistry
- Vol. 284 (48) , 33088-33096
Transcriptional Regulation of Tlr11 Gene Expression in Epithelial Cells
- 1 November 2009
- Published by Elsevier in Journal of Biological Chemistry
- Vol. 284 (48) , 33088-33096
Open AccessTranscriptional Regulation of Tlr11 Gene Expression in Epithelial Cells