Publications

Multi-Fidelity Active Learning with GFlowNets. Alex Hernandez-Garcia, Nikita Saxena, Moksh Jain, Cheng-Hao Liu, Yoshua Bengio. Transactions on Machine Learning Research (TMLR). 2023. OpenReview arXiv code

Crystal-GFN: sampling crystals with desirable properties and constraints. Mila AI4Science, Alex Hernandez-Garcia, Alexandre Duval, Alexandra Volokhova, Yoshua Bengio, Divya Sharma, Pierre Luc Carrier, Yasmine Benabed, Michał Koziarski, Victor Schmidt. Neural Information Processing Systems (NeurIPS), Workshop on Accelerated Materials Design (AI4Mat). 2023. arXiv code

Towards equilibrium molecular conformation generation with GFlowNets. Alexandra Volokhova, Michał Koziarski, Alex Hernandez-Garcia, Cheng-Hao Liu, Santiago Miret, Pablo Lemos, Luca Thiede, Zichao Yan, Alán Aspuru-Guzik, Yoshua Bengio. Digital Discovery, Royal Society of Chemistry. 2023. Digital Discovery arXiv code

On the importance of catalyst-adsorbate 3D interactions for relaxed energy predictions. Alvaro Carbonero, Alexandre Duval, Victor Schmidt, Santiago Miret, Alex Hernandez-Garcia, Yoshua Bengio, David Rolnick. Neural Information Processing Systems (NeurIPS), Workshop on Accelerated Materials Design (AI4Mat). 2023. arXiv

GFlowNets for AI-driven scientific discovery. Moksh Jain, Tristan Deleu, Jason Hartford, Cheng-Hao Liu, Alex Hernandez-Garcia, Yoshua Bengio. Digital Discovery, Royal Society of Chemistry. 2023. Digital Discovery arXiv

Multi-variable hard physical constraints for climate model downscaling. Jose González-Abad, Alex Hernandez-Garcia, Paula Harder, David Rolnick, José Manuel Gutiérrez. arXiv:2308.01868. 2023. arXiv

FAENet: Frame Averaging Equivariant GNN for materials modeling. Alexandre Duval, Victor Schmidt, Alex Hernandez-Garcia, Santiago Miret, Fragkiskos D. Malliaros, Yoshua Bengio, David Rolnick. International Conference on Machine Learning (ICML). 2023. ICML arXiv

A theory of continuous generative flow networks. Salem Lahlou, Tristan Deleu, Pablo Lemos, Dinghuai Zhang, Alexandra Volokhova, Alex Hernandez-Garcia, Léna Néhale Ezzine, Yoshua Bengio, Nikolay Malkin. International Conference on Machine Learning (ICML). 2023. ICML arXiv code

Fourier Neural Operators for arbitrary resolution climate data downscaling. Qidong Yang, Alex Hernandez-Garcia, Paula Harder, Venkatesh Ramesh, Prasanna Sattegeri, Daniela Szwarcman, Campbell D. Watson, David Rolnick. arXiv:2305.14452. 2023. arXiv

Counting carbon: A survey of factors influencing the emissions of machine learning. Sasha Luccioni, Alex Hernandez-Garcia. arXiv:2302.08476. 2023. arXiv

PhAST: physics-aware, scalable, and task-specific GNNs for accelerated catalyst design. Alexandre Duval, Victor Schmidt, Santiago Miret, Yoshua Bengio, Alex Hernandez-Garcia, David Rolnick. Journal of Machine Learning Research (JMLR). 2024. JMLR arXiv

Multi-Objective GFlowNets. Moksh Jain, Sharath Chandra Raparthy, Alex Hernandez-Garcia, Jarrid Rector-Brooks, Yoshua Bengio, Santiago Miret, Emmanuel Bengio. International Conference on Machine Learning (ICML). 2023. ICML arXiv

Diversifying design of nucleic acid aptamers using unsupervised machine learning. Siba Moussa, Michael Kilgour, Clara Jans, Alex Hernandez-Garcia, Miroslava Cuperlovic-Culf, Yoshua Bengio, Lena Simine. Journal of Physical Chemistry B. 2022. ACS arXiv

Hard-constrained deep learning for climate downscaling . Paula Harder, Alex Hernandez-Garcia, Venkatesh Ramesh, Qidong Yang, Prasanna Sattigeri, Daniela Szwarcman, Campbell Watson, David Rolnick. Journal of Machine Learning Research (JMLR). 2023. JMLR arXiv

Biological sequence design with GFlowNets. Moksh Jain, Emmanuel Bengio, Alex Hernandez-Garcia, Jarrid Rector-Brooks, Bonaventure F. P. Dossou, Chanakya Ekbote, Jie Fu, Tianyu Zhang, Micheal Kilgour, Dinghuai Zhang, Lena Simine, Payel Das, Yoshua Bengio. International Conference on Machine Learning (ICML). 2023. ICML arXiv

ClimateGAN: Raising climate change awareness by generating images of floods. Victor Schmidt, Alexandra Sasha Luccioni, Mélisande Teng, Tianyu Zhang, Alexia Reynaud, Sunand Raghupathi, Gautier Cosne, Adrien Juraver, Vahe Vardanyan, Alex Hernandez-Garcia, Yoshua Bengio. International Conference on Learning Representations (ICLR). 2022. arXiv OpenReview ICLR

Computational methods for continuous eye-tracking perimetry based on spatio-temporal integration and a deep recurrent neural network. Alessandro Grillini, Alex Hernandez-Garcia, Remco J. Renken, Giorgia Demaria, Frans W. Cornelissen. Frontiers in Neuroscience. 2021. Frontiers in Neuroscience

Benchmark of data processing methods and machine learning models for gut microbiome-based diagnosis of inflammatory bowel disease. Ryszard Kubinski, Jean-Yves Djamen-Kepaou, Timur Zhanabaev, Alex Hernandez-Garcia, Stefan Bauer, Falk Hildebrand, Tamas Korcsmaros, Sani Karam, Prevost Jantchou, Kamran Kafi, Ryan D. Martin. Frontiers in Genetics. 2022. bioRxiv Frontiers in Genetics

Rethinking supervised learning: insights from biological learning and from calling it by its name. Alex Hernandez-Garcia. Neural Information Processing Systems (NeurIPS), Workshop on Shared visual representations between humans and machines (SVRHM). 2020. OpenReview arXiv

A machine learning pipeline to predict vegetation health. Thomas Lees, Gabriel Tseng, Alex Hernandez-Garcia, Clement Atzberger, Simon Dadson, Steven Reece. International Conference on Learning Representations (ICLR), Workshop on Tackling climate change with machine learning. 2020. Spotlight PDF code

Learning representational invariance instead of categorization. Alex Hernandez-Garcia, Peter König. International Conference on Computer Vision (ICCV), Workshop on pre-registration in Computer Vision. 2019. PDF code

Global visual salience of competing stimuli. Alex Hernandez-Garcia, Ricardo Ramos Gameiro, Alessandro Grillini, Peter König. Journal of Vision. 2020. PsyArXiv Journal of Vision code Twitter

Learning robust visual representations using data augmentation invariance. Alex Hernandez-Garcia, Peter König, Tim C. Kietzmann. arXiv:1906.04547. 2019. OpenReview arXiv code BAICS @ ICLR 2020 Video

Data augmentation instead of explicit regularization. Alex Hernandez-Garcia, Peter König. arXiv:1806.03852. 2018. arXiv code

Deep neural networks trained with heavier data augmentation learn features closer to representations in hIT. Alex Hernandez-Garcia, Johannes Mehrer, Nikolaus Kriegeskorte, Peter König, Tim C. Kietzmann. Cognitive Computational Neuroscience (CCN). 2018. PDF

Do deep nets really need weight decay and dropout?. Alex Hernandez-Garcia, Peter König. arXiv:1802.07042. 2018. arXiv code

Further advantages of data augmentation on convolutional neural networks. Alex Hernandez-Garcia, Peter König. International Conference on Artificial Neural Networks (ICANN). 2018. Best Paper Award arXiv code

Emotion and attention: predicting electrodermal activity through video visual descriptors. Alex Hernandez-Garcia, Fernando Fernández-Martínez, Fernando Díaz de María. Int. Workshop on Affective Computing and Emotion Recognition (ACER). 2017. PDF

Exploiting visual saliency for assessing the impact of car commercials upon viewers. Fernando Fernández-Martínez, Alex Hernandez-Garcia, Miguel Ángel Fernández-Torres, Iván González-Díaz, Álvaro García-Faura, Fernando Díaz de María. Multimedia Tools and Applications. 2017. paper

Comparing visual descriptors and automatic rating strategies for video aesthetics prediction. Alex Hernandez-Garcia, Fernando Fernández-Martínez, Fernando Díaz de María. Journal of Signal Processing: Image Communication. 2016. paper

Succeeding metadata based annotation scheme and visual tips for the automatic assessment of video aesthetic quality in car commercials. Fernando Fernández-Martínez, Alex Hernandez-Garcia, Fernando Díaz de María. International Journal Expert Systems with Applications. 2015. paper

Combining audio-visual features for viewers’ perception classification of Youtube car commercials. Fernando Fernández-Martínez, Alex Hernandez-Garcia, Ascensión Gallardo-Antolín, Fernando Díaz de María. Speech, Language and Audio in Multimedia (SLAM). 2014. paper

Jul 2024. AI4Science Seminar (SAIT), Seoul (South Korea). Generative modelling and active learning with GFlowNets for scientific discoveries. slides

May 2024. Computational Energy Materials Design Infrastructure (CEMDI), Montreal, QC (Canada). Crystal-GFN: Sampling crystals with desirable properties and constraints. link video slides

May 2024. Future Horizons, Montreal, QC (Canada). Active learning and generative modelling for scientific discoveries. slides

April 2024. Exploiter l’IA pour accélérer la découverte de nouveaux matériaux et molécules, Montreal, QC (Canada). Crystal-GFN: A generative model to discover crystal structures with desirable properties and constraints. slides

December 2023. AI4Mat at NeurIPS (spotlight), New Orleans, LA (USA). Crystal-GFN: Sampling crystals with desirable properties and constraints. link slides

November 2023. klaviyo, Boston, MA (USA). AI against climate change: Accelerating scientific discovery and raising awareness with visualisations. slides

September 2023. MIF++ seminar, University of Liverpool, Online (The Internet). Multi-fidelity active learning with GFlowNets for drug and materials discovery. link slides

July 2023. NU Circle, Egypt, Online (The Internet). AI against climate change: Accelerating scientific discovery and raising awareness with visualisations. slides

June 2023. Mathematical and Scientific Machine Learning 2023 (invited talk), Providence, RI (USA). GFlowNets to accelerate scientific discovery with machine learning. link slides

June 2023. Artificial Intelligence for Design Challenge program, NRC Canada, Online (The Internet). Multi-fidelity active learning with GFlowNets to accelerate scientific discoveries. slides

April 2023. Cognition in a Changing World, Case Western Reserve University, Cleveland, OH (USA). AI against climate change: Accelerating scientific discovery and raising awareness with visualisations. slides

March 2023. MMICCs Webinar Series, SEA-CROGS, Pacific Northwest National Laboratory, Online (The Internet). GFlowNets: Introduction and Applications to AI-Driven Scientific Discovery. slides

February 2023. Research Society Manipal Institute of Technology, Manipal (India). AI to fight climate change: Visualising extreme climate events and accelerating scientific discovery. slides

October 2022. Unser Dialog, Online (The Internet). ThisClimateDoesNotExist.com: Erhöhung des Klimabewussteins durch KI-Flutprojektionen. link video slides

July 2022. Department of Signal Theory and Communications, University Carlos III of Madrid, Madrid (Spain). Machine learning contra el cambio climático: Simulación de eventos climáticos y descubrimiento de materiales con GFlowNets. slides

March 2022. Deeptails Seminar, MIAI, GrenobleOnline (FranceThe Internet). GFlowNets: a friendly introduction and designing biological sequences (with Moksh Jain). link video slides

January 2022. Intelligent Machines, Emotions, and our Planet, StockholmOnline (SwedenThe Internet). ThisClimateDoesNotExist.com: AI to visualise climate change impacts on street photos. slides

November 2021. GIS Day, McGill University, MontréalOnline (Québec, CanadaThe Internet). ThisClimateDoesNotExist.com: Visualising climate change impacts on street photos. slides

November 2021. Samsung-Mila-NYU Workshop, Online (The Internet). Deep active learning for DNA aptamer design.

August 2021. European Conference on Visual Perception (ECVP), Online (The Internet). Global visual salience of competing stimuli. slides

July 2021. Women in Machine Learning & Data Science, Yaoundé, YaoundéOnline (CameroonThe Internet). ML and DS to fight the climate emergency. slides

June 2021. Kietzmann Lab, Donders Institute for Brain Cognition and Behaviour, NijmegenOnline (NetherlandsThe Internet). Rethinking (un-, semi-, self-, ...)supervised learning: Insights from biological learning and from calling it by its name. slides

December 2020. AFC Lab Talk Series, hosted by Meike Ramon, Online (The Internet). Global visual salience of competing stimuli. link video

November 2020. Brain & Cognitive Society, IIT Kanpur, KanpurOnline (IndiaThe Internet). More than more data: Insights from data augmentation for deep learning and computational neuroscience. link video

November 2020. Machine Learning Research Group, University of Guelph, GuelphOnline (Ontario, CanadaThe Internet). Data augmentation and image understanding.

October 2020. neuromatch 2020 3.0, Online (The Internet). Global visual salience of competing stimuli. link video

June 2020. Berlin Machine Learning Seminar, BerlinOnline (GermanyThe Internet). Learning robust visual representations using data augmentation invariance. link

May 2020. UNIQUE Student Symposium, MontrealOnline (Quebec, CanadaThe Internet). Learning robust visual representations using data augmentation invariance. link video

April 2020. International Conference on Learning Representations (ICLR), Workshop on Bridging AI and Cognitve Science (BAICS), Addis AbabaOnline (EthiopiaThe Internet). Learning robust visual representations using data augmentation invariance. link video

March 2020. neuromatch 2020 1.0, Online (The Internet). Data augmentation invariance for learning robust visual representations. link

November 2019. International Conference on Computer Vision (ICCV), Workshop on pre-registration in Computer Vision, Seoul (Republic of Korea). Learning representational invariance instead of categorization. link video

November 2019. Department of Brain and Cognitive Engineering, Korea University, Seoul (Republic of Korea). Data augmentation for improved regularization and invariance learning.

October 2019. Computational Cognition, Osnabrück (Germany). Learning robust visual representations using data augmentation invariance. link video

July 2019. Neural Information Processing Group, University of Tübingen, Tübingen (Germany). More than more data: undervalued advantages of data augmentation for deep learning and computational neuroscience. link

July 2019. Berlin Machine Learning Meetup, Berlin (Germany). Data augmentation: an alternative to explicit regularization and other undervalued advantages. link

June 2019. Cognitive Neuroscience Center, University Medical Center Groningen, Groningen (Netherlands). Data augmentation: undervalued advantages for deep learning and computational neuroscience.

February 2019. Neural Dynamics of Visual Cognition Lab, Freie Universität Berlin, Berlin (Germany). On the advantages of data augmentation for deep learning and computational neuroscience.

October 2018. Institute of Informatics and Telecommunications, NCSR Demokritos, Athens (Greece). Data augmentation as a biologically plausible alternative to explicit regularization in CNNs. link

May 2018. Group of Multimedia Processing, University Carlos III of Madrid, Madrid (Spain). Data augmentation instead of explicit regularization.

December 2017. Deep Learning Meetup, Berlin (Germany). Data augmentation instead of explicit regularization.

February, 2024. To ‘green’ AI, scientists are making it less resource-hungry, ScienceNewsExplores. link

August, 2023. Wie können KI-Modelle energiesparender werden? Ein Überblick über aktuelle Trends, te.ma. link

March, 2023. The Week in Green Software: How Green is Your Cloud?, Environment Variables (podcast). link

February, 2023. Artificial intelligence training is powered mostly by fossil fuels, New Scientist. link

November, 2021. L’intelligence artificielle connaît une forte croissance à Montréal, Radio Canada. link

October, 2021. Using artificial intelligence to start a conversation about climate change, CBC. link

October, 2021. L'intelligence artificielle pour visualiser des phénomènes climatiques extrêmes, Radio Canada. link

October, 2021. Times Square bajo el agua o el Zócalo en llamas: así afectará el cambio climático a su calle, El País. link