Who are you, then?

I'm Colorado Reed: a Computer Science Ph.D. student at the Berkeley Artificial Intelligence Research (BAIR) lab. I'm advised by Kurt Keutzer & Trevor Darrell.

Okay. So what do you?

My work focuses on advancing fundamental computer vision research through applying unsupervised learning to real-world problems.

Why work on that?

Aside from being inherently interesting, I work on unsupervised computer vision because of its potential impact in climate/earth/environmental science and remote sensing. Said another way, unsupervised computer vision will be one of the most powerful diagnostic, planning, and evaluation tools in the fight against climate change and the various problems it creates.

How did you wind up here?

Before restarting my Ph.D. in January 2020, I left grad school to start a software company with a few of the most talented folks I've ever met (Sean, Joey, and Avi). And before that, I worked with Maneesh Agrawala and also completed a research-based masters at Cambridge with Zoubin Ghahramani working on statistical machine learning (Bayesian nonparametrics).

What else do you?

I also created Metacademy with Roger Grosse: a web site that uses a dependency graph of concepts to create personalized learning plans for machine learning and related fields.

Yeah, but what else do you?

I also run ultra marathons, bake, and drink oat milk cappuccinos. I love to talk about and do these things, so reach out if you'd like to join me. Oh, and my brother is an amazingly talented and prolific artist.

Your name is really "Colorado"?

Yup. Since birth. Sharing a name with a state is surprisingly common, see Tennessee Williams, Indiana Jones, Virginia Woolf, Dakota Fanning, Georgia O'Keeffe, Tex Ritter, or York Bowen ;-).


  • Region Similarity Representation Learning.
    Tete Xiao*, Colorado J Reed*, Xiaolong Wang, Kurt Keutzer, Trevor Darrell
    Maintain spatial relationships in the convolutional feature maps when performing instance contrastive pretraining -- this is useful for detection-related tasks.
  • Self-Supervised Pretraining Improves Self-Supervised Pretraining
    Colorado J Reed*, Xiangyu Yue*, Ani Nrusimha, Sayna Ebrahimi, Vivek Vijaykumar, Richard Mao, Bo Li, Shanghang Zhang, Devin Guillory, Sean Metzger, Kurt Keutzer, Trevor Darrell
    Initializing self-supervised pretraining with a pretrained model is pretty much always a good idea.
  • SelfAugment: Automatic Augmentation Policies for Self-Supervised Learning.
    Colorado J Reed*, Sean Metzger*, Aravind Srinivas, Trevor Darrell, Kurt Keutzer
    Conference on Computer Vision and Pattern Recognition (CVPR), 2021.
    Determine strong augmentation policies for instance contrastive learning without using supervised evaluations.
  • Multi-source Domain Adaptation in the Deep Learning Era: A Systematic Survey.
    Sicheng Zhao, Bo Li, Colorado J Reed, Pengfei Xu, Kurt Keutzer
    A recent survey on multi-source domain adaptation.
  • Video Digests: A Browsable, Skimmable Format for Informational Lecture Videos.
    Amy Pavel, Colorado J Reed, Bjoern Hartmann, and Maneesh Agrawala
    ACM 27th Symposium on User Interface Software and Technology, 2014.
    Browse an informational video just like a textbook.
  • Scaling the Indian Buffet Process via Submodular Maximization.
    Colorado J Reed, Zoubin Ghahramani
    International Conference on Machine Learning (ICML), 2013.
    Use submodular optimization to shortcut training of latent feature models.
  • Submodular MAP Inference for Scalable Latent Feature Models.
    Colorado J Reed
    Master's Thesis, University of Cambridge, 2013.
    Use submodular optimization to shortcut training, described in 100+ glorious pages.
  • Semi-supervised eigenbasis novelty detection.
    David R. Thompson, Walid A. Majid, Colorado J Reed, and Kiri L. Wagstaff
    Statistical Analysis and Data Mining, 2012
    PCA is useful of novelty detection even with temporal astronomical data (expanded version of below workshop paper).
  • Semi-supervised novelty detection with adaptive eigenbases, and application to radio transients.
    David R. Thompson, Walid A. Majid, Colorado J Reed, and Kiri L. Wagstaff
    Conference on Intelligent Data Understanding, 2011. Best Paper Award
    PCA is useful for novelty detection even with temporal astronomical data.
  • What's trending? Mining topical trends in UGC systems with YouTube as a case study.
    Colorado J Reed, Todd Elvers, and Padmini Srinivasan
    17th ACM SIGKDD, 2011.
    Create clusters of emerging topics on Twitter, YouTube, etc.