Amir H. Karimi

Bio

Dr. Amir-Hossein Karimi is an award-winning educator and researcher, and an Assistant Professor in the Department of Electrical and Computer Engineering and the Cheriton School of Computer Science (cross-appointed) at the University of Waterloo, and a Vector Institute Faculty Affiliate. He leads the Collaborative Human-AI Reasoning Machines (✨CHARM✨) Lab, dedicated to advancing safe and trustworthy human-AI collaborations. Prior to joining Waterloo, Dr. Karimi accumulated significant industry experience at leading tech companies such as BlackBerry (2013), Meta (Facebook) (2014-6), Google Brain (2021), and DeepMind (2022), and provided consulting services for various startups and incubators including for NEXT AI. His contributions have earned him multiple accolades, such as the UofToronto Spirit of Engineering Science Award (2015), the UWaterloo Alumni Gold Medal Award (2018), the NSERC Canada Graduate Scholarship - Doctorate (2018), the Google PhD Fellowship (2021), the ETH Zurich Medal (2024), the NSERC Discovery Grants & Supplements (2024), and the Igor Ivkovic Teaching Excellence Award (2024).

✨CHARM Lab✨

The mandate of the Collaborative Human-AI Reasoning Machines (✨CHARM✨) Lab is to enhance the integration of AI systems into human decision-making, ensuring they are not only powerful but also safe, reliable, and aligned with human values. As AI becomes more embedded in everyday life, e.g., 🏥 healthcare, 🎓 education, 💼 finance, and 🚗 transportation, our dependency on these systems grows, and so does the risk of potential consequential errors. Our mission is to develop AI systems that can detect 🕵️‍♂️ potential issues, correct 🛠️ mistakes, and ultimately perfect 🤝 human-AI partnership where humans and machines work together to achieve better outcomes. 🌍✨

The CHARM Lab focuses on causal inference, explainable AI, and neuro-symbolic approaches in order to build systems that allow users to understand, challenge, and improve AI decisions. We occasionally also borrow insights from, and collaborate with leading experts in, such fields as social sciences, cognitive science, human-computer interaction, multi-agent reinforcement learning, game theory, and behavioral economics.

Prospective Students

The lab is always on the lookout for exceptional and highly motivated students/visitors across all levels (bachelor's, master's, doctoral, postdoctoral). If you are passionate about building the future of trustworthy human-AI symbiosis, and have a strong background in machine learning, computer science, or related fields please fill out this form.

Current Members

Amir-Hossein Karimi

Principal Investigator (PI)

Mina Kebriaee

PhD Student
(w/ Prof. Tahvildari)

Zahra Khotanlou

PhD Student

Maryam Ghorbansabagh

Master's Student
(w/ Prof. Grossmann)

Zachary Wu

Research Assistant
(w/ Prof. Tahvildari)

Lab Alumni

Abubakar Bello

Research Assistant
(next: Microsoft Inc.)

Mohammadreza Alavi

Research Assistant

Past Mentees

Ahmad Ehyaei

Mentee (w/ Prof. Farnadi)
(next: Intl. Max Planck Research Schools)

Miriam Rateike

Mentee (w/ Prof. Valera)
(next: Google PhD Fellow 2023)

Ricardo Dominguez-Olmedo

Mentee (w/ Prof. Schölkopf)
(next: Intl. Max Planck Research Schools)

Kiarash Mohammadi

Mentee (w/ Prof. Valera)
(next: MILA AI Institute)

Alexandra Walter

Mentee (w/ Prof. Valera)
(next: Helmholtz Data Science School of Health)

Publications

Dr. Karimi's scholarly contributions have been showcased almost exclusively at top-tier AI and ML venues including NeurIPS, ICML, AAAI, AISTATS, ACM FAccT, and ACM AIES. He has authored influential publications such as a comprehensive survey paper in the prestigious ACM Computing Surveys, holds a patent, and is a contributing author of a book chapter. Dr. Karimi’s work on algorithmic recourse has notably elevated its prominence in responsible AI research, with its presence growing from almost none to hundreds on Google Scholar in just five years; algorithmic recourse is now a mandatory criterion in key sectors, including Canada’s Treasury Board Directive on Automated Decision-Making. Several of Dr. Karimi's papers have received over 100 citations each. Dedicated to knowledge mobilization and reproducibility, his open-source code has garnered over 100 GitHub stars.

Most recent publications are available on Google Scholar.
^‡ indicates equal contribution.

Paper

📝 Imagining and building wise machines: The centrality of AI metacognition
✍️ Johnson, Karimi, Bengio, Chater, Gerstenberg, Larson, Levine, Mitchell, Rahwan, Schölkopf, Grossmann
✨ TL;DR: AI systems should integrate metacognitive strategies to enhance robustness, explainability, cooperation, and safety in complex real-world scenarios

Paper Slide

📝 On the Relationship Between Explanation and Prediction: A Causal View
✍️ Karimi, Muandet, Kornblith, Schölkopf, Kim
🏛️ ICML 2023 (acceptance rate: 27.9%)
✨ TL;DR: do explanations actually explain the predictions? yes, among other things....

Paper

📝 A Survey of Algorithmic Recourse: Contrastive Explanations & Consequential Recommendations
✍️ Karimi, Barthe, Schölkopf, Valera
🏛️ ACM CSUR 2022 (impact factor: 23.8)
✨ TL;DR: definitions, formulations, solutions, and prospects for recourse research

Paper

📝 Key-Value Memory Networks for Directly Reading Documents
✍️ Miller, Fisch, Dodge, Karimi, Bordes, Weston
🏛️ EMNLP 2016 (acceptance rate: 24.3%)
✨ TL;DR: use key-value pairs to efficiently retrieve and store information for improved document comprehension

Paper

📝 Robustness Implies Fairness in Causal Algorithmic Recourse
✍️ Ehyaei, Karimi, Schölkopf, Maghsudi
🏛️ ACM FAccT 2023 (acceptance rate: 24.6%)
✨ TL;DR: as the title says...

Paper

📝 On the Robustness of Causal Algorithmic Recourse
✍️ Dominguez-Olmedo, Karimi, Schölkopf
🏛️ ICML 2022 (acceptance rate: 21.9%)
✨ TL;DR: robustness of recourse is separate from robustness of prediction

Paper

📝 On the Fairness of Causal Algorithmic Recourse
✍️ von Kügelgen, Karimi, Bhatt, Valera, Weller, Schölkopf
🏛️ AAAI 2022 (acceptance rate: 15.0%)
✨ TL;DR: fairness of recourse is separate from fairness of prediction

Paper

📝 A Survey of Algorithmic Recourse: Contrastive Explanations & Consequential Recommendations
✍️ Karimi, Barthe, Schölkopf, Valera
🏛️ ACM CSUR 2022 (impact factor: 23.8)
✨ TL;DR: definitions, formulations, solutions, and prospects for recourse research

Paper Code Video

📝 Algorithmic Recourse under Imperfect Causal Knowledge: a Probabilistic Approach
✍️ Karimi^‡, von Kügelgen^‡, Schölkopf, Valera
🏛️ NeurIPS 2020 (acceptance rate: 20.1%)
✨ TL;DR: causal assumptions are rarely available; instead generate recourse probabilistically

Paper Code Video

📝 Algorithmic Recourse: from Counterfactual Explanations to Interventions
✍️ Karimi, Schölkopf, Valera
🏛️ ACM FAccT 2021 (acceptance rate: 25.0%)
✨ TL;DR: real-world causal assumptions are needed for counterfactual explanations & recourse

Paper

📝 Scaling Guarantees for Nearest Counterfactual Explanations
✍️ Mohammadi, Karimi, Barthe, Valera
🏛️ ACM AIES 2021 (acceptance rate: 38.0%)
✨ TL;DR: generate optimal counterfactual explanations using mixed-integer linear programs

Paper Code Video

📝 Model-Agnostic Counterfactual Explanations for Consequential Decisions
✍️ Karimi, Barthe, Valera
🏛️ AISTATS 2019 (acceptance rate: 32.4%)
✨ TL;DR: generate optimal counterfactual explanations using satisfiability solvers

Paper

📝 Causal Adversarial Perturbations for Individual Fairness and Robustness in Heterogeneous Data Spaces
✍️ Ehyaei, Mohammadi, Karimi, Samadi, Farnadi
🏛️ AAAI 2023 (acceptance rate: 23.7%)

Paper Slide

📝 On the Relationship Between Explanation and Prediction: A Causal View
✍️ Karimi, Muandet, Kornblith, Schölkopf, Kim
🏛️ ICML 2023 (acceptance rate: 27.9%)
✨ TL;DR: do explanations actually explain the predictions? yes, among other things....

Paper

📝 On Data Manifolds Entailed by Structural Causal Models
✍️ Dominguez-Olmedo, Karimi, Arvanitidis, Schölkopf
🏛️ ICML 2023 (acceptance rate: 27.9%)

Paper

📝 Robustness Implies Fairness in Causal Algorithmic Recourse
✍️ Ehyaei, Karimi, Schölkopf, Maghsudi
🏛️ ACM FAccT 2023 (acceptance rate: 24.6%)
✨ TL;DR: as the title says...

Paper

📝 On the Robustness of Causal Algorithmic Recourse
✍️ Dominguez-Olmedo, Karimi, Schölkopf
🏛️ ICML 2022 (acceptance rate: 21.9%)
✨ TL;DR: robustness of recourse is separate from robustness of prediction

Paper

📝 On the Fairness of Causal Algorithmic Recourse
✍️ von Kügelgen, Karimi, Bhatt, Valera, Weller, Schölkopf
🏛️ AAAI 2022 (acceptance rate: 15.0%)
✨ TL;DR: fairness of recourse is separate from fairness of prediction

Paper

📝 A Survey of Algorithmic Recourse: Contrastive Explanations & Consequential Recommendations
✍️ Karimi, Barthe, Schölkopf, Valera
🏛️ ACM CSUR 2022 (impact factor: 23.8)
✨ TL;DR: definitions, formulations, solutions, and prospects for recourse research

Paper Code Video

📝 Algorithmic Recourse under Imperfect Causal Knowledge: a Probabilistic Approach
✍️ Karimi^‡, von Kügelgen^‡, Schölkopf, Valera
🏛️ NeurIPS 2020 (acceptance rate: 20.1%)
✨ TL;DR: causal assumptions are rarely available; instead generate recourse probabilistically

Paper Code Video

📝 Algorithmic Recourse: from Counterfactual Explanations to Interventions
✍️ Karimi, Schölkopf, Valera
🏛️ ACM FAccT 2021 (acceptance rate: 25.0%)
✨ TL;DR: real-world causal assumptions are needed for counterfactual explanations & recourse

Paper

📝 Imagining and building wise machines: The centrality of AI metacognition
✍️ Johnson, Karimi, Bengio, Chater, Gerstenberg, Larson, Levine, Mitchell, Rahwan, Schölkopf, Grossmann
✨ TL;DR: AI systems should integrate metacognitive strategies to enhance robustness, explainability, cooperation, and safety in complex real-world scenarios

Paper

📝 Prospector Heads: Generalized Feature Attribution for Large Models & Data
✍️ Machiraju, Derry, Desai, Guha, Karimi, Zou, Altman, Ré, Mallick
🏛️ ICML 2024 (acceptance rate: 27.5%)

Paper Slide

📝 On the Relationship Between Explanation and Prediction: A Causal View
✍️ Karimi, Muandet, Kornblith, Schölkopf, Kim
🏛️ ICML 2023 (acceptance rate: 27.9%)
✨ TL;DR: do explanations actually explain the predictions? yes, among other things....

Paper

📝 Scaling Guarantees for Nearest Counterfactual Explanations
✍️ Mohammadi, Karimi, Barthe, Valera
🏛️ ACM AIES 2021 (acceptance rate: 38.0%)
✨ TL;DR: generate optimal counterfactual explanations using mixed-integer linear programs

Paper Code Video

📝 Model-Agnostic Counterfactual Explanations for Consequential Decisions
✍️ Karimi, Barthe, Valera
🏛️ AISTATS 2019 (acceptance rate: 32.4%)
✨ TL;DR: generate optimal counterfactual explanations using satisfiability solvers

Paper

📝 Imagining and building wise machines: The centrality of AI metacognition
✍️ Johnson, Karimi, Bengio, Chater, Gerstenberg, Larson, Levine, Mitchell, Rahwan, Schölkopf, Grossmann
✨ TL;DR: AI systems should integrate metacognitive strategies to enhance robustness, explainability, cooperation, and safety in complex real-world scenarios

Paper

📝 SRP: Efficient class-aware embedding learning for large-scale data via supervised random projections
✍️ Karimi, Wong, Ghodsi
🏛️ arXiv

Paper

📝 Ensembles of Random Projections for Nonlinear Dimensionality Reduction
✍️ Karimi, Shafiee, Ghodsi, Wong
🏛️ CVIS 2017 (acceptance rate: 40.0%)

Paper

📝 Synthesizing Deep Neural Network Architectures using Biological Synaptic Strength Distributions
✍️ Karimi, Shafiee, Ghodsi, Wong
🏛️ CCN 2017 (acceptance rate: 40.0%)

Paper

📝 Distance Correlation Autoencoder
✍️ Wang, Karimi, Ghodsi
🏛️ IJCNN 2018 (acceptance rate: 22.7%)

Paper

📝 Key-Value Memory Networks for Directly Reading Documents
✍️ Miller, Fisch, Dodge, Karimi, Bordes, Weston
🏛️ EMNLP 2016 (acceptance rate: 24.3%)
✨ TL;DR: use key-value pairs to efficiently retrieve and store information for improved document comprehension

Paper

📝 Spatio-temporal Saliency Detection using Abstracted Fully-Connected Graphical Models
✍️ Karimi, Shafiee, Scharfenberger, BenDaya, Haider, Talukdar, Clausi, Wong
🏛️ ICIP 2016 (acceptance rate: 40.0%)

Teaching

Dr. Karimi enjoys teaching diverse audiences, covering broad topics in AI for general audiences, courses on ML for undergrad/grad students, and specialized courses for practitioners. In his first teaching term at the University of Waterloo, he was honored with the 2024 Igor Ivkovic Teaching Excellence Award. Dr. Karimi has been invited to present talks, lectures, and tutorials at institutions such as MIT, Harvard, ETH Zurich, MILA, University College London, Cyber Valley Health, Institute of Mathematical Statistics, NEC Europe Labs, DeepMind, and Google Brain. He gave notable tutorials on Causal Explainable AI at KDD 2023 and Algorithmic Recourse at the Toronto ML Summit 2024. Please reach out to book Dr. Karimi for teaching engagements at your organization or event.

For everyone...

From AI to Life Lessons

Watch on YouTube

Success Studio

Read on Medium

AI for Autism

Listen on BBC Podcasts

For students...

Machine Learning Essentials

👉🏼 Master ML Today 👈🏼

CausEthical Machine Learning

Seminar | Workshop

Dimension Reduction & PCA

Watch on YouTube

For professionals...

Machine Learning Essentials

👉🏼 Master ML Today 👈🏼

Algorithmic Recourse

Watch on YouTube

Causal Explainable AI

Get the slides

Vitæ

Full CV in PDF. Having immigrated five times across three continents for studies and work, from Iran to Canada, and then onward to the USA, Germany, Switzerland, and the UK, Dr. Amir-Hossein Karimi has accumulated over 15 years of technical experience in both research, industry, and startup roles. Now an Assistant Professor at the University of Waterloo and a Vector Institute Faculty Affiliate, he has held prominent research positions at Google DeepMind, Google Brain, alongside software engineering roles at BlackBerry and Meta (Facebook). His academic journey includes a Ph.D. at the Max Planck Institute & ETH Zürich, with a focus on causal inference and explainable AI.

UWaterloo & Vector Institute Sep 2023 - Present

Assistant Professor
Electrical and Computer Engineering Department
Google DeepMind May 2022 - Oct 2022

Research Scientist Intern
Mentors: Dr. Lars Buesing, Dr. Jessica Hamrick, David Amos
Google Brain Dec 2021 - Apr 2022

Research Scientist Intern
Mentors: Dr. Been Kim, Dr. Simon Kornblith
Max Planck Institute & ETH Zürich Oct 2018 - Aug 2023

Ph.D. Student in Computer Science
Advisors: Prof. Bernhard Schölkopf, Prof. Isabel Valera
University of Waterloo Sep 2016 - Aug 2018

M.Math in Computer Science
Advisors: Prof. Alexander Wong, Prof. Ali Ghodsi
Meta (Facebook) Inc. Aug 2015 - Sep 2016

Software Engineer
Mentors: Dr. Antoine Bordes, Dr. Jason Weston, Kaustubh Karkare
Meta (Facebook) Inc. Feb 2014 - Apr 2014

Software Engineer Intern
Mentor: Chris Triolo
BlackBerry Inc. May 2013 - Dec 2013

Software Engineer Intern
Mentor: Chris Joe
Stanford University May 2012 - Aug 2012

Undergraduate Research Assistant
Advisor: Prof. Amin Arbabian
University of Toronto Sep 2010 - May 2015

B.A.Sc. in Engineering Science – Electrical and Computer Stream
Advisors: Prof. Chris Eliasmith, Prof. Richard Zemel

Press, Funding, & Consultations

Dr. Karimi is grateful for the generous funding support from the University of Waterloo, NSERC, Google, and Waterloo.AI, enabling his team to push the boundaries of human-AI research.

In addition to his academic contributions, Dr. Karimi has consulted internationally for several startups, leveraging his expertise for discovery, product-market fit, and scaling of operations, in addition to fundraising and grant writing. He is available to discuss how his research and experience can deliver value to your organization and stakeholders. Book a time via this link.

For press inquiries, feel free to reach out.