About me - Research - Teaching - Service - CV - Blog - Google Scholar - Github - LinkedIn
About me
I am a researcher at KTH Royal Institute of Technology in Stockholm, Sweden. Before that, I did a PhD and a master’s degree in Machine Learning also at KTH and a bachelor’s degree in Telecommunications engineering at Universitat Politècnica de Catalunya (UPC) in Barcelona, Spain.
Previously, I was an intern at the Machine Learning Group (Summer 2022) and at the Privacy and Machine Learning team at Apple (Summer 2021), the Machine Learning and Computer Vision team at CISCO (Summer 2019), and the Machine Learning and Computer Vision team at Tobii AB (Summer 2018).
I study the intersection between information theory and machine learning, under the supervision of Mikael Skoglund and Ragnar Thobaben. I am particularly interested in:
- Generalization of learning algorithms.
- Privacy and fairness of learning algorithms.
- Uncertainty in learning agorithms.
- Representation learning. More specifically, in no particular order:
- Minimal (and sufficient) representations.
- Disentangled representations.
- Private and fair represenations.
However, I am also interested in talking about other topics in information theory and machine learning. So, please, do not hesitate to contact me for a discussion (borjabrg12 at gmail dot com).
Research
PhD Thesis
Publications (Chronological)
- A note on generalization bounds for losses with finite moments
Borja Rodríguez-Gálvez, Omar Rivasplata, Ragnar Thobaben, and Mikael Skoglund
- More PAC-Bayes bounds: From bounded losses, to losses with general tail behaviors, to anytime-validity
Borja Rodríguez-Gálvez, Ragnar Thobaben, and Mikael Skoglund
- Sections 2 and 3 presented at the PBMIL workshop of ICML 2023. Section 3 as an oral presentation.
- Journal of Machine Learning Research [pdf] [cite]
- The Role of Entropy and Reconstruction for Multi-View Self-Supervised Learning
Borja Rodríguez-Gálvez, Arno Blaas, Pau Rodriguez, Adam Golinski, Xavier Suau, Jason Ramapuram, Dan Busbridge, and Luca Zappella
- Thompson Sampling Regret Bounds for Contextual Bandits with sub-Gaussian rewards
Amaury Gouverneur, Borja Rodríguez-Gálvez, Tobias J. Oechtering, and Mikael Skoglund
- Limitations of information-theoretic generalization bounds for gradient descent methods in stochastic convex optimization
Mahdi Haghifam, Borja Rodríguez-Gálvez, Ragnar Thobaben, Mikael Skoglund, Daniel M. Roy, Gintare Karolina Dziugaite
- An Information-Theoretic Analysis of Bayesian Reinforcement Learning
Amaury Gouverneur, Borja Rodríguez-Gálvez, Tobias J. Oechtering, Mikael Skoglund
- Tighter expected generalization error bounds via Wasserstein distance
Borja Rodríguez-Gálvez, Germán Bassi, Ragnar Thobaben, and Mikael Skoglund
- Presented at ITR3 Workshop of ICML 2021 (Contributed talk) [pdf] [cite]
- NeurIPS 2021
[pdf] [cite]
- Measuring Gender Bias in
Contextualized Embeddings
Styliani Katsarou, Borja Rodrı́guez-Gálvez, and Jesse Shanahan
- Presented at the AIBSD Workshop of AAAI 2022
[pdf] [repo]
- Enforcing fairness in private federated learning via the modified method of differential multipliers
Borja Rodríguez-Gálvez, Filip Granqvist, Rogier van Dalen, and Matt Seigel
- Presented at PriML Workshop of NeurIPS 2021
[pdf] [cite]
- Upper Bounds on the Generalization Error of Private Algorithms
Borja Rodríguez-Gálvez, Germán Bassi, and Mikael Skoglund
- IEEE Transactions on Information Theory
[pdf] [cite]
- A Variational Approach to Privacy and Fairness
Borja Rodríguez-Gálvez, Ragnar Thobaben, and Mikael Skoglund
- On Random Subset Generalization Error Bounds and the Stochastic Gradient Langevin Dynamics Algorithm
Borja Rodríguez-Gálvez, Germán Bassi, Ragnar Thobaben, and Mikael Skoglund
- The Convex Information Bottleneck Lagrangian
Borja Rodríguez-Gálvez, Ragnar Thobaben, and Mikael Skoglund
Talks
-
“A Variational Approach to Privacy and Fairness”. Spotlight presentation at the AAAI PPAI Workshop. Online. February 8th, 2021.
-
“Tighter expected generalization error bounds via Wasserstein distance”. Invited talk at the Information Theory, Machine Learning and Statistics Seminar. Online. April 30th, 2021.
-
“Tighter expected generalization error bounds via Wasserstein distance”. Contributed talk at the ICML ITR3 Workshop. Online. July 24th, 2021.
-
“PAC-Bayes bounds’ parameter optimization via events’ space discretization: new bounds for losses with general tail behaviors”. Contributed talk at the ICML PBMIL Workshop. Honolulu, Hawaii. July 28th, 2023.
-
“A Variational Approach to Privacy and Fairness”. Invited talk at the Swiss Data Science Center. Online. October 25th, 2023.
-
“More PAC-Bayes bounds: From bounded losses, to losses with general tail behaviors, to anytime-validity”. Invited talk at the UCL department of statistics. London, UK. November 15h, 2023.
-
“More PAC-Bayes bounds: From bounded losses, to losses with general tail behaviors, to anytime-validity”. Invited talk at Google Deepmind. London, UK. November 17h, 2023.
Teaching
Service
Reviewing
- Journals: IEEE Trans. IT, IEEE TIFS, JMLR, IEEE TCOM, EURASIP JASP
- Conferences: NeurIPS, ICML, ICLR, AISTATS, IEEE ISIT, UAI
Master’s Theses’ supervising
- Elias Ågeby - Currently exploring his culinary passion
Introducing Sparsity into the Current Landscape of Disentangled Representation Learning - [pdf] [website] [code]
- Johan Sörell - Currently at Academic Work Sweden AB
A General Approach to Inaudible Adversarial Perturbations in a Black-box Setting - [pdf]
- Mihaela Georgieva Stoycheva - Currently at Ocado
Uncertainty Estimation in Deep Neural Object Detectors for Autonomous Driving - [pdf]
- Polixeni Ioannidou - Currently at Ridgeline Discovery
Anomaly Detection in Computer Networks - [pdf]
- Styliani Katsarou - Currently at Peltarion
Improving Multilingual Models for the Swedish Language: Exploring CrossLingual Transferability and Stereotypical Biases - [pdf]
—
Page template forked from evanca