The purpose of XAI-Healthcare 2023 event is to provide a place for intensive discussion on all aspects of eXplainable Artificial Intelligence (XAI) in the medical and healthcare field. This should result in cross-fertilization among research on Machine Learning, Decision Support Systems, Natural Language, Human-Computer Interaction, and Healthcare sciences. This meeting will also provide attendees with an opportunity to learn more on the progress of XAI in healthcare and to share their own perspectives. The panel discussion will provide participants with the insights on current developments and challenges from the researchers working in this fast-developing field.
Explainable AI (XAI) aims to address the problem of understanding how decisions are made by AI systems by designing formal methods and frameworks for easing their interpretation. The impact of AI in clinical settings and the trust placed in such systems by clinicians have been a growing concern related to the risk of introducing AI into the healthcare environment. XAI in healthcare is a multidisciplinary area addressing this challenge by combining AI technologies, cognitive modeling, healthcare science, ethical and legal issues.
Previous editions
- XAI-Healthcare 2022 during IEEE ICHI @ Rochester, USA. [proceedings] [video-proceedings]
- XAI-Healthcare 2021 during AIME @ Porto, Portugal. [proceedings] [video-proceedings]
PROCEEDINGS
Papers selected and presented during XAI-Healthcare 2023 were published in a CCIS Springer book series as part of the Proceedings of the International Workshops on Explainable Artificial Inteligence and Process Mining for Healthcare 2023.
Preface
In this preface we described the motivation, papers presented, and discussions during XAI-Healthcare 2023 event. [PREFACE]
Papers published in the proceedings
- Mozhgan Salimiparsa, Kamran Sedig and Daniel Lizotte Unlocking the Power of Explainability in Ranking Systems: A Visual Analytics Approach with XAI Techniques [PAPER]
- Laurent Cervoni, Rita Sleiman, Damien Jacob and Mehdi Roudesli Explainable artificial intelligence in response to the failures of musculoskeletal disorder rehabilitation [PAPER]
- Syed Hamail Hussain Zaidi, Bilal Hashmat and Muddassar Farooq An Explainable AI Framework for Treatment Failure Model for Oncology Patients [PAPER]
- Olga Kamińska, Tomasz Klonecki and Katarzyna Kaczmarek Majer Feature selection in bipolar disorder episode classification using cost-constrained methods [PAPER]
- Enrique Valero-Leal, Pedro Larrañaga and Concha Bielza ProbExplainer: A library for unified explainability of probabilistic models and an application in interneuron classification[PAPER]
- Carlos Hernández-Pérez, Cristian Pachón-García, Pedro Delicado and Verónica Vilaplana Interpreting Machine Learning models for Survival Analysis: A study of Cutaneous Melanoma using the SEER Database [PAPER]
- William Van Woensel, Floriano Scioscia, Giuseppe Loseto, Oshani Seneviratne, Evan Patton and Samina Abidi Explanations of Symbolic Reasoning to Effect Patient Persuasion and Education[PAPER]
Invited speaker
Keynote: Prof. Mihaela van der Schaar
Professor of Machine Learning, Artificial Intelligence and Medicine at the University of Cambridge.
Fellow at The Alan Turing Institute in London.
Title of the talk: New frontiers in machine learning interpretability (online)
About M. van der Schaar and her lab.
Activities & Program
XAI-Healthcare encourage face-to-face interaction between researchers of the XAI and healthcare field. In this third edition the activities included are: XAI hackathon, the Keynote of Prof. Mihaela van der Schaar and the presentation of selected high-quality research paper submissions.
XAI Hackathon 2023
This activity was focused on researchers interested in knowing the most practical aspects of the use of XAI techniques in healthcare, through the Phython library AraucanaXAI [AraucanaXAI paper].
Hands-on session, the attendees had a chance to:
- Familiarize with our XAI approach and its Python implementation
- Apply it to your own dataset/problem of interest
- Extend it to add new cool features like working with unstructured data (text, image, you name it), temporal data, bio-signals, graphs, etc.
Scope
We expect the contributions received to describe explanation methods, AI techniques and a targeted healthcare problem. Some examples are provided below for guidance, but the list of topics is not limited to these specific methods, techniques and problems.
Explanation Approaches:
- Model agnostic methods
- Feature analysis
- Visualization approaches
- Example and counterfactuals based explanations
- Fairness, accountability and trust
- Evaluating XAI
- Fairness and bias auditing
- Human-AI interaction
- Human-Computer Interaction (HCI) for XAI
- Natural Language Processing (NLP) Explainability
AI techniques:
- Blackbox ML approaches: DL, random forest, etc.
- Interpretable ML models: Rules, Trees, Bayesian networks, etc.
- Statistical models and reasoning
- Case-based reasoning
- Natural language processing and generation
- Abductive Reasoning
Target healthcare problems:
- Infection challenges (COVID, Antibiotic Resistance, etc.)
- Trustworthy AI
- Chronic diseases
- Ageing & home care
- Diagnostic systems
Important Dates
- April 24, 2023 Paper submission
- May 11, 2023 Acceptance
- May 15, 2023 Final mansucript
- Jun 15, 2023 1-day Workshop
Submissions
Papers should be submitted to the XAI-Healthcare Easy Chair Website at SUBMIT LINK .
Papers should be formatted according to Springer Lecture Notes Format , either for LaTeX or for Word. Springer's proceedings LaTeX templates are available in Overleaf (overleaf template).
The workshop features regular papers in two categories: short papers (up to 5 pages) describing work-in-progress and full papers (up to 10 pages) describing original and solid results.
We are aiming at providing selected papers in a post-proceedings volume, gathering all papers presented at the workshop. Furthermore, we are also considering a proposal of a special issue of a journal like Journal of Healthcare Informatics Research (JHIR) or a similar specialized journal.
Workshop Organisers
- Concha Bielza, Dept. of Artificial Intelligence, Universidad Politecnica de Madrid [contact]
- Pedro Larrañaga, Dept. of Artificial Intelligence, Universidad Politecnica de Madrid [contact]
- Primoz Kocbek, Faculty of Health Sciences, University of Maribor [contact]
- Jose M. Juarez, Faculty of Computer Science, University of Murcia [contact]
- Gregor Stiglic, Faculty of Health Sciences, University of Maribor [contact]
- Alfredo Vellido, Universitat Politecnica de Catalunya and UPC BarcelonaTech [contact]
- Alejandro Rodriguez, Universidad Politecnica de Madrid, Spain
- Bernardo Canovas-Segura, University of Murcia
- Carlo Combi, University of Verona, Italy
- Caroline Konig, Universitat Politecnica de Catalunya, Spain
- Huang Zhengxing, Faculty of Biomedical Engineering,China
- Jean-Baptiste Lamy, LIMICS, France
- Jose M. Alonso, Universidad de Santiago de Compostela, Spain
- Lluis Belanche, Universitat Politecnica de Catalunya, Spain
- Milos Hauskrecht, University of Pittsburgh, USA
- Nava Tintarev, Maastricht University, The Netherlands
- Paulo Felix, Universidad de Santiago de Compostela, Spain
- Pedro Cabalar, University of Coruna, Spain
- Przemyslaw Biecek, Warsaw University of Technology, Poland
- Riccardo Bellazzi, University of Pavia, Italy
- Shashikumar Supreeth, University of California San Diego, USA
- Pham Thai-Hoang, The Ohio University, USA
- Zhe He, Florida State University, USA
Program Committee