Talks
Invited talks, seminars and panels by Sonali Parbhoo.
-
Dharma, Demonstration and the Alignment of Machines · Slides
at Hinduism and AI: A Roundtable, University of Cambridge, Sep 2026 -
Causal Discovery in the Age of Weak Supervision: From Preference Data to Language Model Oracles · Slides
at MCML workshop on causal machine learning, Munich, Aug 2026 -
Interpretability in the Wild: When Explainability Meets Causality and Clinical Reality
at University of Cambridge, Mar 2026 -
From Research to Certification with Data-Driven Medical Decision Support Systems
at Dagstuhl Seminar, 2026 -
Fix-A-Shortcut: Gradient-Based Shortcut Mitigation for Causal Representation Learning in Healthcare · Slides
at Royal Statistical Society, Emerging Applications invited talk, Dec 2025 -
The Pitfalls of Shortcuts in Multimodal Representation Learning for Healthcare
at NeurIPS 2025 workshop on ML for Multimodal Representation Learning in Healthcare, Dec 2025 -
Mind Your Evaluation: On the Challenges with Off-Policy Evaluation for Real World Decision-Making · Slides
at University of Oxford, 2025
at EPFL, 2025
at RLC I Can’t Believe It’s Not Better Workshop, 2024
at CIFAR Workshop on World Models: Causality, Neuroscience and AI Safety, Max Planck Institute, Tübingen, 2024
at Novel Trends in Data Science Workshop, ETH Zurich, 2023
at Open Data Science Conference, London, 2023 -
Insights from the Inverse: Leveraging Inverse Reinforcement Learning for Reconstructing LLM Training Goals
at Google DeepMind, 2025 -
On the Promises of Data-Driven Approaches for Early Glaucoma Detection
at Aravind Eye Institute, Madurai, 2024 -
Taking the Genie Out Of The Bottle(neck): Concept-Based Insights for AI Oversight · Slides
at KDD Workshop on Interpretability, 2024 -
Interpretable Reinforcement Learning for Actionable Decision-Making in Healthcare
at RLC InterpPol Workshop, 2024 -
Lessons from the Past for Smarter Decisions and Future Wins · Slides
at Microsoft Research Cambridge, 2024 -
Insights from the Inverse: Leveraging Inverse Reinforcement Learning for Actionable Decision-Making in Healthcare · Slides
at Apple Healthcare Workshop, Mountain View, 2024 -
Larger Models are Not Always Better: Towards Creating ML People Can Use in Practice
at ISN Tea, Imperial College London, 2024 -
When Information Is Not Enough: Towards Uncertainty-Aware Real Time Decisions for Health
at Critical Care Summit and Datathon, Tarragona, 2022 -
Help, I Need Somebody! Deferring to Humans for Sequential Decision-Making in Uncertainty · Slides
at Boston Consulting Group, 2022 -
On Addressing Bias and Uncertainty in Machine Learning for Equity in Healthcare
at DSxHE, University College London, 2022 -
Machine Learning Methods for Sepsis Management in the ICU
at IMSE, Imperial College London, 2022 -
Towards Risk-Aware Decision-Making and Policy Evaluation for Actionable Healthcare
at Oxford Machine Learning Summer School, 2022 -
Towards Responsible and Effective AI for Health
at Imperial College London, 2022 -
Building Evidence-Based Models for Targeted Healthcare
at Purdue University, 2022
at EECS, Imperial College London, 2021
at University of Edinburgh, 2021 -
Why Uncertainty Matters: Implications for Medical Evaluation and Beyond
at MIT, 2022 -
Counterfactual Representation Learning for Multiple Concurrent Therapies
at IMES and HMS, MIT, 2021 -
Understanding Uncertainty in Decision-Making with Visualisations
at CSAIL, MIT, 2021 -
The Role of Counterfactuals in High-Stake Decision-Making
at Stanford University, 2021 -
Robust Machine Learning Methods for Targeted Healthcare
at Rising Stars in AI for Social Good, Harvard CRCS, 2021 -
The Need For Interpretable and Fair Algorithms in Health Policy
at Womxn@Broad Symposium, Broad Institute of MIT and Harvard, 2021 -
The Role of Ethics in Computer Science
at ABCD Committee, Harvard University, 2021 -
Treatment-Effect Estimation from Well-Curated to Less-Resourced Populations with HIV
at Michigan AI & Health Symposium, University of Michigan, 2020 -
Information Bottleneck Methods for Estimating Treatment Effects
at Deep Learning IndabaX, South Africa, 2020 -
Kernel Methods for HIV Therapy Selection
at Harvard University, 2016 -
Bayesian Neighbourhood Estimation
at University of the Witwatersrand, 2016