Monday, May 19, 2025, 2:30 – 5:30 pm, Royal Tern

Organizers: William Broderick (Flatiron Institute)

Speakers: William Broderick (Flatiron Institute); Edoardo Balzani (Flatiron Institute); Sarah Jo Venditto (Flatiron Institute); Guillaume Viejo (Flatiron Institute)

This satellite event is a hands-on tutorial for the open source plenoptic python library (https://plenoptic.org/). After a brief introductory presentation, I will walk participants through a jupyter notebook, explaining how to use plenoptic to better understand computational visual models. Participants will be expected to follow along on their laptops, either running the code locally or using a provided binder instance (https://jupyter.org/binder) on the cloud (with the environment installed and GPUs available). The goal is to demonstrate for attendees the sorts of scientific questions that can be addressed with plenoptic and encourage them to use it in their own research.

While the exact content may change, I intend to introduce attendees to model metamers[1] and eigendistortions[2], and show how they can be used to understand and compare a handful of simple visual models (e.g., a linear Gaussian convolutional model, linear center-surround convolutional model, and a simple model of gain control using divisive normalization).

The session will be interactive, with opportunities for attendees to ask and answer questions. I will lead the entire session, and will be accompanied by several teaching assistants, who will circulate throughout the room to address any technical issues attendees encounter. Given this, attendance will be capped at 30, and a smaller room would be beneficial.

[1]: e.g., as used in “A parametric texture model based on joint statistics of complex wavelet coefficients”, Portilla and Simoncelli, 2000

[2]: e.g., as used in “Eigen-distortions of hierarchical representations,” Berardino et al., 2017