Scholar
Leonardo Cotta
Google Scholar ID: 0GI4MyoAAAAJ
Ellison Institute of Technology
Citations & Impact
Citations
313
H-index
8
i10-index
8
Publications
20
Co-authors
0
Publications
Resume (English only)
Academic Achievements
- Recent Pre-prints:
- - Test-Time Fairness and Robustness in Large Language Models, with CJ Maddison
- - Boosting the Predictive Power of Protein Representations with a Corpus of Text Annotations, with H Duan, M Skreta, et al.
- Selected Publications:
- - End-To-End Causal Effect Estimation from Unstructured Natural Language Data
- - Probabilistic Invariant Learning with Randomized Linear Classifiers
- - Causal Lifting and Link Prediction
- - Reconstruction for Powerful Graph Representations
- - Unsupervised Joint k-node Graph Representations with Compositional Energy-Based Models
- - Graph Pattern Mining and Learning through User-defined Relations
- - AoT: Authentication and Access Control for the Entire IoT Device Life-Cycle
Research Experience
- Currently a distinguished postdoc fellow at Vector Institute, hosted by Chris J. Maddison.
Education
- PhD in Computer Science from Purdue University, advised by Bruno Ribeiro; BSc in Computer Science from UFMG, Brazil, where he worked on distributed algorithms (at UFMG) and quantum computing theory (at the University of Calgary).
Background
- Research Interests: Machine learning and causal inference methods, particularly for high-stakes applications of AI such as science and recommendation engines. Long-term goal is to deliver new AI solutions that have real, tangible world impact.
Miscellany
- Personal interests and support activities: Donating to cancer patients in Brazil
Co-authors