Adriana Romero Soriano
RESEARCH SCIENTIST | MONTREAL, CANADA
Adriana is currently a research scientist at Facebook AI Research and an adjunct professor at McGill University. The goal of her research is to develop models and algorithms that are able to learn from multi-modal and real world data, understand and reason about conceptual relations, and recognize their uncertainties, while addressing impactful problems. The playground of her research has been defined by problems which require inferring full observations from limited sensory data. Previously, she was a post-doctoral researcher at Mila, advised by Prof. Yoshua Bengio. her postdoctoral research revolved around deep learning techniques to tackle biomedical challenges, such as the ones posed by multi-modal data, high dimensional data and graph structured data. She received her Ph.D. from University of Barcelona with a thesis on assisting the training of deep neural networks, advised by Dr. Carlo Gatta.
Research Areas
Adriana's Work
Adriana's Publications
October 19, 2025
RESEARCH
NLP
Controlling Multimodal LLMs via Reward-guided Decoding
Aishwarya Agrawal, Adriana Romero Soriano, Koustuv Sinha, Michal Drozdzal, Oscar Mañas, Pierluca D'Oro
October 19, 2025
December 12, 2024
COMPUTER VISION
EvalGIM: A Library for Evaluating Generative Image Models
Melissa Hall, Abhishek Charnalia, Adriana Romero Soriano, Candace Ross, Carolina Braga, Jakob Verbeek, Karen Ullrich, Maeve Ryan, Mark Ibrahim, Marton Havasi, Michal Drozdzal, Mike Rabbat, Oscar Mañas, Pietro Astolfi, Reyhane Askari, Tariq Berrada Ifriqi, Yohann Benchetrit
December 12, 2024
June 14, 2024
COMPUTER VISION
Decomposed evaluations of geographic disparities in text-to-image models
Abhishek Sureddy, Dishant Padalia, Nandhinee Periyakaruppa, Oindrila Saha, Polina Kirichenko, Melissa Hall, Adina Williams, Adriana Romero Soriano, Megan Richards
June 14, 2024
April 04, 2024
CORE MACHINE LEARNING
DP-RDM: Adapting Diffusion Models to Private Domains Without Fine-Tuning
Chuan Guo, Adriana Romero Soriano, Jonathan Lebensold, Kamalika Chaudhuri, Maziar Sanjabi, Mike Rabbat, Pietro Astolfi
April 04, 2024
November 03, 2021
CORE MACHINE LEARNING
Parameter Prediction for Unseen Deep Architectures
Boris Knyazev, Graham Taylor, Adriana Romero Soriano, Michal Drozdzal
November 03, 2021
October 26, 2021
Active 3D Shape Reconstruction from Vision and Touch
Edward Smith, Adriana Romero Soriano, Jitendra Malik, Luis Pineda, Michal Drozdzal, Roberto Calandra, David Meger
October 26, 2021
November 04, 2020
REINFORCEMENT LEARNING
COMPUTER VISION
Active MR k-space Sampling with Reinforcement Learning
Roberto Calandra, Sumana Basu, Adriana Romero Soriano, Luis Pineda, Michal Drozdzal
November 04, 2020
October 30, 2020
3D Shape Reconstruction from Vision and Touch
Adriana Romero Soriano, Edward Smith, Georgia Gkioxari, Jitendra Malik, Michal Drozdzal, Roberto Calandra, David Meger
October 30, 2020
June 16, 2019
RESEARCH
COMPUTER VISION
Inverse Cooking: Recipe Generation from Food Images
Michal Drozdzal, Adriana Romero Soriano, Amaia Salvador Aguilera, Xavier Giro-i-Nieto
June 16, 2019
June 03, 2019
CONVERSATIONAL AI
RESEARCH
GEOMetrics: Exploiting Geometric Structure for Graph-Encoded Objects
Adriana Romero Soriano, Dave Meger, Edward Smith, Scott Fujimoto
June 03, 2019
May 10, 2019
RESEARCH
COMPUTER VISION
Reducing Uncertainty in Undersampled MRI Reconstruction with Active Acquisition
Michal Drozdzal, Adriana Romero Soriano, Pascal Vincent, Lin Yang, Matthiew Muckley, Zizhao Zhang
May 10, 2019
June 18, 2018
RESEARCH
COMPUTER VISION
On the iterative refinement of densely connected representation levels for semantic segmentation
Adriana Romero Soriano, Michal Drozdzal, Arantxa Casanova, Guillem Cucurull, Yoshua Bengio
June 18, 2018
April 29, 2018
RESEARCH
COMPUTER VISION
Graph Attention Networks
Adriana Romero Soriano, Arantxa Casanova, Guillem Cucurull, Petar Velickovic, Pietro Lio, Yoshua Bengio
April 29, 2018
Our approach
Latest news
Foundational models
Meta © 2026