Rate My Professor Sara Magliacane

Her research centers on the intersection of causality and machine learning, with key directions including causal representation learning from high-dimensional data like image sequences, efficient causal discovery, and causality-inspired machine learning and reinforcement learning for robustness to distribution shifts, domain adaptation, and dynamical systems modeling. Notable publications include 'Domain Adaptation by Using Causal Inference to Predict Invariant Conditional Distributions' (NeurIPS 2018, cited over 300 times), 'Ancestral Causal Inference' (NeurIPS 2017), 'Joint Causal Inference from Multiple Contexts' (JMLR 2020), 'CITRIS: Causal Identifiability from Temporal Intervened Sequences' (ICML 2022), 'BISCUIT: Causal Representation Learning from Binary Interactions' (ICLR 2023), and 'SNAP: Sequential Non-Ancestor Pruning for Targeted Causal Effect Estimation' (AISTATS 2025). In November 2025, she received an NWO VIDI grant of €850,000 for her project 'CANES: a CAusal NEuro-Symbolic approach to integrating perception and abstract reasoning,' aimed at learning understandable concepts from few annotations. Magliacane has supervised PhD students who graduated cum laude, served as program chair for UAI 2025 and general chair for UAI 2026, co-organized workshops on causality and logic-AI, and given invited talks at Harvard Data Science Initiative Causal Seminar, ONRL AI Seminar Series, and ML in PL 2025.