About
My name is Nico Lang and I am a Postdoc at the University of Copenhagen associated with the Pioneer centre for AI, where I am co-advised by Serge Belongie and Christian Igel. I received my PhD from ETH Zurich, where I was working in the EcoVision Lab that is part of the Photogrammetry and Remote Sensing (PRS) group under the supervision of Prof. Konrad Schindler and Prof. Jan Dirk Wegner.
[Curriculum Vitae] [Headshot] [Bio]
Research
My research interests are in the area of computer vision, machine learning, and remote sensing and in developing new methods using these techniques to support environmental sciences. My PhD research focused on advancing the way we measure forest structure at global scales using publicly available satellite data. My interests also include uncertainty estimation in deep learning, anomaly detection, and learning from imbalanced data, as these are omnipresent challenges that arise when working on real-world problems. I see great potential in the application of machine learning to address global environmental issues.
Studies
During my Bachelor’s degree in Geomatics and Planning at ETH Zurich, I discovered my fascination for machine learning and computer vision in the context of geospatial data science. Before continuing with my Master in Geomatics at ETH Zurich, I started an internship in computer vision at upicto, a start-up from the Computer Vision Lab at ETH, and continued it at Logitech, as upicto was acquired during my time there.
Collaborations
Following the work started during an early academic visit to Prof. Pietro Perona’s Computational Vision Lab at Caltech, I worked on deep learning approaches to monitor urban trees from street-level images. During my Phd I’ve had the chance to collaborate with the NASA GEDI mission, the Swiss Federal Institute for Forest, Snow and Landscape Research WSL, as well as with companies and institutions outside academia: Barry Callebaut, High carbon stock Approach (HCSA), Hunziker, Zarn und Partner (HZP), and MeteoSwiss.
Leisure
I enjoy spending my free time outdoors hiking and snowboarding, playing volleyball, and traveling where windy takes me kitesurfing. I play the drums and like making music with others.
News
November 2024: I also joined Bluesky here: @nicolang.bsky.social. Interesting starter packs to consider: Deep Learning for Earth Observation 🌎 | AI for Conservation | Computer Vision 1 & Computer Vision 2 by @csprofkgd.bsky.social | Belongie Lab
November 2024: Invited talk at DHI about representation learning for Earth observation data.
November 2024: Joined the research network Climate Change AI Nordics. Invited to give a talk at the 1st workshop on May 13: 2025 Nordic Workshop on AI for Tackling Climate Change.
September 2024: Invited talk at the LifeCLEF 2024 workshop.
August 2024: ECCV workshop paper accepted to CV4E: Multimodal Fusion Strategies for Mapping Biophysical Landscape Features
June 2024: ECCV paper accepted: Labeled Data Selection for Category Discovery. TLDR: Category discovery is improved by selecting labeled data that is neither too similar nor too different to the unlabeled categories.
June 2024: ECCV paper accepted: MMEarth: Exploring Multi-Modal Pretext Tasks For Geospatial Representation Learning. TLDR: We created “MMEarth”, a global dataset with 12 aligned modalities at 1.2M locations to explore multi-modal pretext tasks for learning representations for Sentinel-2 images.
June 2024: We are organizing a summer PhD course on SSL4EO: Self-Supervised Learning for Earth Observation at University of Copenhagen. Looking forward to a week full of interesting talks and discussions. Registration is open, seats are limited.
June 2024: Invited talk at the AI2 Environmental seminar series, Seattle.
May 2024: Working with Lucia Gordon from Harvard University during a summer visit to the Pioneer Centre for AI.
May 2024: New preprint: MMEarth: Exploring Multi-Modal Pretext Tasks For Geospatial Representation Learning. TLDR: We created “MMEarth”, a global dataset with 12 aligned modalities at 1.2M locations to explore multi-modal pretext tasks for learning representations for Sentinel-2 images.
May 2024: Invited as a panelist to the ML4RS workshop at ICLR 2024 in Vienna.
April 2024: I am visiting Oisin Mac Aodha’s group at University of Edinburgh for a two weeks research visit as part of a strategic partnership grant between UoE and KU.
April 2024: I co-organized the Visipedia workshop 2024 at the Pioneer Centre for AI. More info about the Visipedia project can be found here.
March 2024: Paper accepted to CVPR 2024: From Coarse to Fine-Grained Open-Set Recognition. TLDR: We investigate the role of label granularity, semantic similarity, and hierarchical representations in open-set recognition with a new benchmark based on iNat2021.
January 2024: I am co-organizing the 11th Workshop of the Fine-Grained Visual Categorization accepted to CVPR 2024 in Seattle.
September 2023: Journal paper accepted in Nature Ecology and Evolution: A high-resolution canopy height model of the Earth. Our project website links all related resources (data, code).
September 2023: Workshop paper accepted on “Familiarity-Based Open-Set Recognition Under Adversarial Attacks”. It will be presented at The 2nd Workshop and Challenges for Out-of-Distribution Generalization in Computer Vision during ICCV 2023.
September 2023: Invited talk at the RISE Research Institutes of Sweden Learning Machines Seminars. The recording of the talk and the lively discussion is online. [video]
September 2023: New journal publication in Science Advances led by Siyu Liu: The overlooked contribution of trees outside forests to tree cover and woody biomass across Europe.
July 2023: I am very happy to receive the Culmann Prize that recognizes outstanding doctoral theses.
June 2023: I joined the CVPR23 “house band” as a drummer in Vancouver.
May 2023: New journal publication in Nature Food led by Nikolai Kalischek: Cocoa plantations are associated with deforestation in Côte d’Ivoire and Ghana.
May 2023: I am honored to have been named one of the outstanding reviewers for CVPR 2023.
May 2023: Invited talk at AI for Good: “Global vegetation monitoring with probabilistic deep learning”. [video] [slides]
December 2022: I am excited to co-organize the “FGVC10: The 10th Fine-Grained Visual Categorization” workshop which was accepted to CVPR 2023 in Vancouver. CALL FOR PAPERS: Interested in fine-grained learning and its applications? Consider submitting your paper to FGVC10. Deadline: March 20, 2023.
December 2022: German national TV “ARD” presents our research on global canopy height mapping in the quiz show “Wer Weiss Denn Sowas”. Watch the episode here (at 7:45 min).
September 2022: I moved to Copenhagen and started as a Postdoc at the University of Copenhagen, where I am associated with the new Pioneer centre for AI. I am co-advised by Serge Belongie and Christian Igel.
June 2022: My doctoral thesis is now online: Mapping Vegetation Height — Probabilistic Deep Learning for Global Remote Sensing. Here are the slides from my defense.
May 2022: Swiss national TV “SRF” reports on our research in the news program “10 vor 10”: «Living Planet Symposium» mit Schweizer Beteiligung.
May 2022: I defended my PhD! Big thanks to all the great people supporting me on this journey. I posted a picture of my amazing hat on twitter here.
May 2022: Got invited to speak at the at ML for Remote Sensing virtual discussion group organized by Hannah Kerner and Patrick Clifton Gray. See their schedule here.
May 2022: Our research has been covered by several news articles from NVIDIA, NASA, NBC. See the list of links under “Media”.
April 2022: ETH News reported about our research: Neural network can read tree heights from satellite images.
April 2022: New preprint available: A high-resolution canopy height model of the Earth; and a new project website to access the demo and data.
February 2022: I got invited to give a seminar talk to the Alan Turing Institute Environment and Sustainability Special Interest Group.
January 2022: I got invited to give a talk at Google’s Geo for Good Lightning Talks Series #6: Forest & Nature.
October 2021: Our paper together with the NASA GEDI mission is published in the journal Remote Sensing of Environment.
October 2021: Our research project Automated Large-scale High Carbon Stock Estimation from Space was selected to be featured at the AI+X Summit organized by the ETH AI Center.
August 2021: I got invited to give a Radio interview at Swiss Radio SRF1 about our research in a program called “The measurement of forests”. Die Aufzeichnung der Sendung “Treffpunkt” zum Thema: Die Vermessung der Wälder (Swiss German).
July 2021: Our research got featured on the High Carbon Stock Approach (HCSA) website: Publicly available indicative High Carbon Stock Forest maps for Malaysia, Indonesia, and the Philippines
June 2021: Mongabay covered our research in a news article: Chocolate giant funds high resolution carbon map to protect forests
May 2021: News article by Barry Callebaut: Artificial intelligence against deforestation
January 2021: We were interviewed by RESET: “EcoVision Lab Is Mapping Biomass for the Good of the Planet” (German, English translation)
May 2020: News article by ETH Zürich Industry Relations: A global tool against deforestation
August 2019: I wrote a Medium post summarizing my first journal paper as first author: Deep learning reconstructs forests in 3D using only satellite images and initiated the Medium publication of the EcoVision Lab.
March 2019: The U.V. Helava Award for best paper 2018 in ISPRS Journal of Photogrammetry and Remote Sensing (with Jan D. Wegner, Konrad Schindler, and Steve Branson, David Hall, Pietro Perona from the Caltech Computational Vision Group)
October 2018: I am an associate PhD fellow of the Max Planck ETH Center for Learning Systems (CLS)
August 2016: I am visiting Prof. Pietro Perona’s Computational Vision Lab at Caltech for my interdisciplinary project. I am supported by a travel grant from the Erich-Degen Stiftung (ETH Zürich).