Tyler R. Scott

Tyler R. Scott

Machine Learning Researcher

I am a senior software engineer at Google with a PhD in computer science from the University of Colorado, Boulder. I was advised by Michael C. Mozer. Previously, I’ve interned with the Brain and Perception teams at Google, interned at Facebook Reality Labs, and worked as a machine learning research scientist at Sensory Inc. Outside of research, I enjoy rock climbing, skiing, and reading.

Research areas: machine learning, deep learning, representation learning, few–shot learning, zero–shot learning, model robustness, clustering.

My research interests are centered on deep representation learning: learning low-dimensional, semantic representations of data that capture relevant factors of variation. Better representations can improve downstream tasks including few–shot learning, zero–shot learning, inductive transfer learning, robustness, verification, clustering, and retrieval.

Conference & Journal Publications

  • Mayo, D., Scott, T. R., Ren, M., Elsayed, G., Hermann, K., Jones, M., and Mozer, M. C. (2023). Multitask Learning via Interleaving: A Neural Network Investigation. Proceedings of the Annual Conference of the Cognitive Science Society (COGSCI 2023).
  • Jones, M., Scott, T. R., Ren, M., Elsayed, G., Hermann, K., Mayo, D., and Mozer, M. C. (2023). Learning in Temporally Structured Environments. Proceedings of the International Conference on Learning Representations (ICLR 2023). Link.
  • Scott, T. R., Liu, T., Mozer, M. C., and Gallagher, A. C. (2022). An Empirical Study on Clustering Pretrained Embeddings: Is Deep Strictly Better? Submitted for publication. Also arXiv:2211.05183 [cs.CV].
  • Ren, M., Scott, T. R., Iuzzolino, M. L., Mozer, M. C., and Zemel, R. (2021). Online Unsupervised Learning of Visual Representations and Categories. Submitted for publication. Also arXiv:2109.05675 [cs.CV].
  • Scott, T. R., Gallagher, A. C., and Mozer, M. C. (2021). von Mises-Fisher Loss: An Exploration of Embedding Geometries for Supervised Learning. Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV 2021). Also arXiv:2103.15718 [cs.LG]. [26% acceptance rate].
  • Scott, T. R., Ridgeway, K., and Mozer, M. C. (2018). Adapted Deep Embeddings: A Synthesis of Methods for k-Shot Inductive Transfer Learning. Advances in Neural Information Processing Systems 31 (NeurIPS 2018) (pp. 76-85). Also arXiv:1805.08402 [cs.LG]. [Selected as spotlight; 4% acceptance rate].

Workshop Publications

  • Kim, D. Y. J., Scott, T. R., Mallick, D., and Mozer, M. C. (2021). Using Semantics of Textbook Highlights to Predict Student Comprehension and Knowledge Retention. AIED Workshop on Intelligent Textbooks. Link.
  • Scott, T. R., Shvartsman, M., and Ridgeway, K. (2020). Unifying Few- and Zero-Shot Egocentric Action Recognition. CVPR Workshop on Egocentric Perception, Interaction, and Computing. Also arXiv:2006.11393 [cs.CV].
  • Scott, T. R., Ridgeway, K., and Mozer, M. C. (2019). Stochastic Prototype Embeddings. ICML Workshop on Uncertainty and Robustness in Deep Learning. Also arXiv:1909.11702 [stat.ML].