[ scholar ] [ datasets ] [ github ] [ blog ] [ short cv ]
I am a computer scientist. My primary interests are machine learning and computer vision, with an emphasis on creating machine learning systems that enable scientific discoveries.
Currently, I am a Research Scientist at GenBio AI. I am working on large AI models for biology, particularly at the cell and tissue level.
Previously, I was a Research Scientist at Altos Labs under Peter Kharchenko. At Altos, I developed machine learning systems for discovering biologically significant cell and tissue states from experimental data (e.g. scRNA-seq, spatial transcriptomics, histology, IF, IHC). I earned my Ph.D. in Pietro Perona’s computer vision group at Caltech, supported by an NSF Graduate Research Fellowship. I spent two years of my Ph.D. working with Serge Belongie at Google Research and completed summer internships at Microsoft Research and the U.S. Air Force Research Laboratory. Before that, I received a B.S.E. from Duke University with a double major in electrical engineering and mathematics. I am grateful to my passionate public school teachers in Utah (Grantsville, Sandy, Salt Lake City), Nebraska (Omaha), and Texas (San Antonio).
Combining Observational Data and Language for Species Range Estimation
Max Hamilton, Christian Lange, Elijah Cole, Alexander Shepard, Samuel Heinrich, Oisin Mac Aodha, Grant Van Horn, Subhransu Maji
NeurIPS 2024
TLDR: Using natural language to enable zero-shot estimation of the spatial distribution of a species.
[paper]
From Coarse to Fine-Grained Open-Set Recognition
Nico Lang, Vésteinn Snæbjarnarson, Elijah Cole, Oisin Mac Aodha, Christian Igel, Serge Belongie
CVPR 2024
TLDR: Revealing the factors that make open-set recognition easy or difficult, with new benchmarks.
[paper]
Active Learning-Based Species Range Estimation
Christian Lange, Elijah Cole, Grant Van Horn, Oisin Mac Aodha
NeurIPS 2023
TLDR: An active learning approach to rapidly estimate the spatial distribution of a novel species.
[paper]
Spatial Implicit Neural Representations for Global-Scale Species Mapping
Elijah Cole, Grant Van Horn, Christian Lange, Alexander Shepard, Patrick Leary, Pietro Perona, Scott Loarie, Oisin Mac Aodha
ICML 2023
TLDR: Implicit neural representations for jointly estimating the spatial range of thousands of species from sparse data, with new benchmarks.
[paper] [website] [demo] [code+data] [blog]
On Label Granularity and Object Localization
Elijah Cole, Kimberly Wilber, Grant Van Horn, Xuan Yang, Marco Fornoni, Pietro Perona, Serge Belongie, Andrew Howard, Oisin Mac Aodha
ECCV 2022
TLDR: Increasing the accuracy and data efficiency of weakly supervised object localization by using coarse-grained labels.
[paper] [website]
When Does Contrastive Visual Representation Learning Work?
Elijah Cole, Xuan Yang, Kimberly Wilber, Oisin Mac Aodha, Serge Belongie
CVPR 2022
TLDR: An empirical analysis of self-supervised learning through the lens of real-world problems.
[paper] [website]
Multi-Label Learning from Single Positive Labels
Elijah Cole, Oisin Mac Aodha, Titouan Lorieul, Pietro Perona, Dan Morris, Nebojsa Jojic
CVPR 2021
TLDR: It’s possible to train effective multi-label image classifiers using only one positive label per image.
[paper] [code] [video]
Benchmarking Representation Learning for Natural World Image Collections
Grant Van Horn, Elijah Cole, Sara Beery, Kimberly Wilber, Serge Belongie, Oisin Mac Aodha
CVPR 2021 (Oral)
TLDR: New benchmarks for self-supervised learning grounded in realistic fine-grained ecology tasks.
[paper] [code+data] [video]
Species Distribution Modeling for Machine Learning Practitioners: A Review
Sara Beery, Elijah Cole, Joseph Parker, Pietro Perona, Kevin Winner
ACM COMPASS 2021
TLDR: A review paper to help ML researchers learn about one of the most important problems in ecology and conservation.
[paper]
Presence-Only Geographical Priors for Fine-Grained Image Classification
Oisin Mac Aodha, Elijah Cole, Pietro Perona
ICCV 2019
TLDR: Efficient algorithms for using spatio-temporal information to improve fine-grained image classification.
[paper] [code] [video]
Feel free to reach out via [email] or [linkedin].