Joëlle Hanna

About me

I am a Research Scientist at Google DeepMind in Zürich, where I work on Earth observation as part of the Biosphere team within Frontier AI’s Sustainability efforts. My research focuses on computer vision and remote sensing, particularly on building multimodal models for Earth observation with an emphasis on implicit supervision and modality-aware learning. I completed my PhD at the University of St. Gallen under the supervision of Damian Borth, with co-supervision from Michael Mommert and Ribana Roscher. Prior to joining Google DeepMind, I was a Student Researcher at Google Research in Zürich. I received my M.Sc. and B.Sc. in Electrical and Electronic Engineering from EPFL.

News

2026

  • Joined Google DeepMind in August 2026 as a Research Scientist within the Frontier AI, Sustainability Team.
  • Defended my PhD Thesis on "Rethinking Learning Objectives and Supervision for Geospatial Representation Learning".
  • One paper accepted at CVPR "GeoSANE: Learning Geospatial Representations from Models, Not Data" [arxiv][blog]
  • One paper accepted in the IEEE Transactions on Geoscience and Remote Sensing on "MAPEX: Modality-Aware Pruning of Experts for Remote Sensing Foundation Models" [paper]
  • Reviewer for CVPR, ECCV, EarthVision, NeurIPS

2025

  • One paper accepted at ICCV "Know Your Attention Maps: Class-specific Token Masking for Weakly Supervised Semantic Segmentation" [arxiv]
  • One poster accepted at the Swiss Remote Sensing Days
  • Gave a talk at the ESA-NASA International workshop on AI Foundation Model for Earth Observation on a Modality-Aware Pruning of Experts for Multi-Modal Remote Sensing Foundation Models
  • Reviewer for CVPR, EarthVision, ICCV

2024

  • Co-supervised a master thesis on "SAR-to-RGB Translation with Latent Diffusion for Earth Observation" [arxiv]
  • One poster presented at the ETH AI + Environment summits (more info here) on "Sparse Multimodal Vision Transformer for Weakly Supervised Semantic Segmentation"
  • Gave a tutorial on "Data-Efficient Deep Learning for Earth Observation" at the 2024 IEEE International Geoscience and Remote Sensing Symposium (IGARSS)
  • Teacher Assistant for the Artificial Intelligence course - taught by Damian Borth.
  • Teacher Assistant for the Computer Vision course - taught by Christos Sakaridis.
  • Reviewer for CVPR, EarthVision, ECCV, WACV

2023

  • Did an internship at Google Research (Health AI), on building a Conversational Assistant for Radiotherapy Patients using MedPalm
  • Co-supervised a bachelor thesis on "Deep Clustering for Traffic Camera Vehicle Detection"
  • One paper accepted at the CVPR EarthVision Workshop on "Sparse Multimodal Vision Transformer for Weakly Supervised Semantic Segmentation" [paper]
  • One paper accepted in the IEEE Transactions on Geoscience and Remote Sensing on "Physics-Guided Multitask Learning for Estimating Power Generation and CO2 Emissions from Satellite Imagery" [paper]
  • One paper accepted in the IEEE International Geoscience and Remote Sensing Symposium on "Ben-ge: Extending BigEarthNet with Geographical and Environmental Data" [arxiv]
  • Gave a tutorial on "Data-Efficient Deep Learning for Earth Observation" at the 2023 IEEE International Geoscience and Remote Sensing Symposium (IGARSS)
  • One poster accepted at the Swiss Remote Sensing days.
  • Co-hosted the Swiss Remote Sensing days at the University of St.Gallen.
  • Teacher Assistant for the Computer Vision course - taught by Michael Mommert.
  • Defended my PhD proposal on "multimodal representation learning for Earth Observation".
  • Reviewer for WACV

2022

  • Won the Best Student Paper Award for the paper "Self-supervised Vision Transformers for Land-cover Segmentation and Classification" at CVPR EarthVision Workshop
  • One paper accepted at the CVPR EarthVision Workshop on "Self-supervised Vision Transformers for Land-cover Segmentation and Classification" [paper]
  • One paper accepted in the IEEE International Geoscience and Remote Sensing Symposium on "A Multimodal Approach for Event Detection: Study of UK Lockdowns in the year 2020" [paper]
  • One poster accepted at the Swiss Remote Sensing days.
  • Gave a Talk at the AI4EO Symposium in the Technical University of Munich.
  • Invited Talk at the Visual Intelligence seminar, The Arctic University of Norway on "Self-supervised Vision Transformers for Land-cover Segmentation and Classification".

2021

  • One paper accepted at the Tackling Climate Change with Machine Learning workshop in NeurIPS, on "Multitask Learning for Estimating Power Plant Greenhouse Gas Emissions from Satellite Imagery" [paper]
  • One abstract accepted at the ESA-ECMWF Workshop on "Estimating Industrial Greenhouse Gas Emissions from Satellite Imagery" [paper]
  • Started my PhD in the AIML Lab in the University of St Gallen