João Monteiro
Research Scientist
Low Data Learning
João is a Research Scientist at ServiceNow Research with interests in devising learning algorithms and models classes that improve out-of-distribution generalization, as well as to define learning procedures that can be applied to multiple different tasks. João obtained a PhD from the Institut National de la Recherche Scientifique in Montreal where he focused on applications of generative modeling and metric learning to voice biometrics settings. His goal was to improve the state-of-the-art in tasks such as speaker verification and spoken language identification, while also improving the robustness of such approaches against spoofing attackers.
Interests
- Machine Learning
- Text2Code
- Retrieval
- Uncertainty Estimation
Publications
StarCoder: may the source be with you!.
ArXiv,
2023.
Constraining Representations Yields Models That Know What They Don't Know.
International Conference of Learning Representations (ICLR),
2023.
Multilingual Code Retrieval Without Paired Data: A New Benchmark and Experiments.
Workshop at the International Conference on Learning Representations (ICLR),
2023.
Constraining Low-level Representations to Define Effective Confidence Scores.
Workshop at the Neural Information Processing Systems (NeurIPS),
2022.
Contrastive Self-supervision Defines General-Purpose Similarity Functions.
Workshop at the Neural Information Processing Systems (NeurIPS),
2022.
Leveraging Activation Patterns to Define Classifiers Able to Detect and Reject Anomalies.
Montreal AI Symposium (MAIS),
2022.
Monotonicity Regularization: Improved Penalties and Novel Applications to Disentangled Representation Learning and Robust Classification.
Conference on Uncertainty in Artificial Intelligence (UAI),
2022.