I am a Professor and Schmidt Career Advancement Chair in Artificial Intelligence at the faculty of Data and Decisions Science (former Industrial Engineering and Management) of the Technion – Israel Institute of Technology. I am also an affiliated lecturer at The  Centre for Human-Inspired Artificial Intelligence (CHIA) of the  University of Cambridge, UK.

I am a Co-editor in Chief of the Transactions of the Association for Computational Linguistics (TACL, MIT Press) and a fellow of ELLIS. I am active in the hi-tech industry, served as a chief scientist for several companies (e.g.,  suridata.ai, recently purchased by  Fortinet) and provide consultancy services to other companies (among previous companies: Yahoo!, Gong.io, and Meta).

My research has long focused on out-of-distribution generalization, sample-efficient learning, and multilingual natural language processing. While I continue to advance these directions, I am also increasingly interested in multi-modal learning – integrating language with vision, behavioral, and emotional signals. I work at the intersection of causality and machine learning/NLP, particularly in the context of large language model explainability, interpretability, and robustness to distributional shifts. A growing focus of my work lies in developing decision-making agents capable of natural language communication, and in harnessing NLP and multi-modal models for scientific and health-related discovery – especially in neuroscience, mental health, cognition, emotion, and behavior – where language and multimodal models can offer transformative insights and impact.

A more detailed description of my group’s vision and research can be found in Group’s vision and research.

Announcements

Galit Weinstein and I are editing a special issue titled: Artificial Intelligence in Alzheimer’s disease and Related Dementias Research: Unveiling Novel Research Directions” in the Alzheimer’s & Dementia: Diagnosis, Assessment & Disease Monitoring (DADM) journal. The deadline is September 1, 2025 – If you have relevant research that you have not yet submitted to our special issue, we warmly encourage you to do so!

Our paper (Nitay Calderon and Roi Reichart, [PDF: arXiv]): “On Behalf of the Stakeholders: Trends in NLP Model Interpretability in the Era of LLMs.” Received the Senior Area Chair best paper award for Model Analysis and Interpretability in NNACL 2025.

Office: Bloomfield 313 | E-MAIL: roiri at technion dot ac dot il
CV (August 2022)