Joshua Caleb Peterson

Joshua Peterson

joshuacp "at" princeton "dot" edu

I'm a postdoc in the Computer Science department at Princeton University. My PhD was completed at UC Berkeley, where I focused on repurposing representations learned by deep neural networks as proxies to study humans and bootstrap cognitive models.

Publications

Peterson, J., Battleday, R., Russakovsky, O., & Griffiths, T. (2019). Human uncertainty makes classification more robust.
In Proceedings of the IEEE International Conference on Computer Vision.

*Bourgin, D., *Peterson, J., Reichman, D., Griffiths, T., & Russell, S. (2019). Cognitive model priors for predicting human decisions.
In the Proceedings of the International Conference on Machine Learning (ICML).
  * co-first author

Peterson, J., Abbott, J., & Griffiths, T. (2018). Evaluating (and improving) the correspondence between deep neural networks and human representations.
Cognitive Science, 42(8), 2648-2669.
PDF

Janata, P., Peterson, J., Ngan, C., Keum, B., Whiteside, H., & Ran, S. (forthcoming). Psychological and musical factors underlying engagement with unfamiliar music.
Music Perception.

Peterson, J., Suchow, J., Aghi, K., Ku, A., & Griffiths, T. (2018). Capturing human category representations by sampling in deep feature spaces.
Presented at the 40th Annual Conference of the Cognitive Science Society.
PDF

Peterson, J., Soulos, P., Nematzadeh, A., & Griffiths, T. (2018).
Learning Hierarchical Visual Representations in Deep Neural Networks Using Hierarchical Linguistic Labels.
Presented at the 40th Annual Conference of the Cognitive Science Society.
PDF POSTER

Grant, E., Finn, C., Peterson, J., Abbott, J., Levine, S., Griffiths, T.L. & Darrell, T. (2017).
Concept acquisition via meta-learning: Few-shot learning from positive examples.
In Proceedings of the Workshop on "Cognitively Informed Artificial Intelligence" at NeurIPS 2017.
PDF

Peterson, J., & Griffiths, T. (2017). Evidence for the size principle in semantic and perceptual domains.
Presented at the 39th Annual Conference of the Cognitive Science Society.
PDF

Battleday, R., Peterson, J., & Griffiths, T. (2017). Modeling human categorization of natural images using deep feature representations.
Presented at the 39th Annual Conference of the Cognitive Science Society.

Dawn, C., Peterson, J., & Griffiths, T. (2017). Evaluating vector-space models of analogy.
Presented at the 39th Annual Conference of the Cognitive Science Society.
PDF DATASET

Peterson, J., Abbott, J., & Griffiths, T. (2016). Adapting Deep Network Features to Capture Psychological Representations.
Presented at the 38th Annual Conference of the Cognitive Science Society. (34% acceptance rate)
PDF ABSTRACT
[Winner of the Computational Modeling Prize in Perception/Action]
[Selected for IJCAI-17 Sister Conference Best Paper Track]

*Dubey, R., *Peterson, J., Khosla, A., Yang, M., & Ghanem, B. (2015). What makes an object memorable?
In the Proceedings of the International Conference on Computer Vision. (25% acceptance rate)
PDF WEBSITE ABSTRACT   * co-first author