seminar Department of Industrial Engineering
A Counterfactual Framework for Learning and Evaluating Explanations for Recommender Systems
Liya Gurevitch Advisor: Dr. Noam Koenigstein
Abstract:
In the field of recommender systems, explainability remains a pivotal yet challenging aspect. To address this, we introduce the Learning to eXplain Recommendations (LXR) framework, a posthoc, model-agnostic approach designed for providing counterfactual explanations. LXR is compatible with any differentiable recommender algorithm and scores the relevance of user data in relation to recommended items. A distinctive feature of LXR is its use of novel selfsupervised counterfactual loss terms, which effectively highlight the most influential user data responsible for a specific recommended item. Additionally, we propose several innovative counterfactual evaluation metrics specifically tailored for assessing the quality of explanations in recommender systems.
Bio:
Liya Gurevitch is a M.Sc. student at the department of Industrial Engineering in Tel Aviv University, specializing in Data Science. Liya holds a B.SC. Degree in Industrial Engineering from Tel Aviv University. Her research, supervised by Dr. Noam Koenigstein, focuses on explainability of recommendation systems (XAI) in order to create a generic framework which can be applied for explaining different models in a self-supervised manner.
UTILIZING SKILLS FOR ADVANCED JOB RECOMMENDER SYSTEMS – AN ITERATIVE SCOPING REVIEW
Alon Atzil Advisors: Dr. Hila Chalutz Ben-Gal & Prof. Irad Ben-Gal
Abstract:
In recent years, the widespread adoption of machine learning-based job recommender systems became a major component in predictive human resources analytics. However, recent workforce disruptions, such as freelancing, gig, and flexible work, have resulted in the emerging need to adopt new models and tools to support and implement state-of-the-art job recommender systems, with comprehensive skill utilization methodologies. For this purpose, we first adopt a scoping review method to systematically analyze job recommender systems literature. Subsequently, we introduce a novel conceptual framework for skill extraction. Finally, we provide a decision support tool tailored for scholars and practitioners to assess and implement skillbased job recommender systems, for a variety of workforce challenges. This study contributes to the predictive human resources analytics domain by systematically and iteratively developing a skill-based job recommender systems approach that supports and optimizes human capital management in the new world of work.
Bio:
Alon Atzil is an M.Sc. student in the Department of Industrial Engineering at Tel Aviv University. He holds a BSc degree in Industrial Engineering from Ben-Gurion university. Currently, he serves as a Senior Data Scientist in a fintech company. Prior to this role he has held positions as head of analytics and data science team leader.
