Learn to Compress

This workshop is inspired by the belief that data compression, foundations of which are rooted in information theory, is on the brink of a significant transformation. The emergence of deep generative models, like variational autoencoders, generative adversarial networks (GANs), normalizing flows, and diffusion models, has opened up a fresh path for data compression, one that fully taps into the power of machine learning. These methods demonstrate impressive capabilities, particularly with image and video data, yet challenges remain for practical applications.

This workshop will act as a dynamic platform for fostering interdisciplinary collaborations, featuring distinguished experts in machine learning and computer science, who will contribute valuable practical insights to the real-world applications of data compression. For ISIT regulars, it will acquaint them with recent advances in data compression, enabling them to explore the interplay between classical and contemporary methods. For those new to the field and experts alike, it will provide a chance to learn from experienced researchers from the industry and academia, and connect with peers who share similar research interests. The workshop aims to leave attendees with a more comprehensive intellectual toolkit in the era of machine learning.

Accepted papers: OpenReview

Spotlight papers (in alphabetical order):

Estimation of Rate-Distortion Function for Computing with Decoder Side Information.
Heasung Kim, Hyeji Kim, Gustavo De Veciana.
Rate-Distortion-Perception Tradeoff for Vector Gaussian Sources.
Jingjing Qian, Sadaf Salehkalaibar, Jun Chen, Ashish Khisti, Wei Yu, Wuxian Shi, Yiqun Ge, Wen Tong.
Some Notes on the Sample Complexity of Approximate Channel Simulation.
Gergely Flamich, Lennie Wells.
Staggered Quantizers for Perfect Perceptual Quality: A Connection between Quantizers with Common Randomness and Without.
Ruida Zhou, Chao Tian.

Update [April 2024]: Early registration deadline for ISIT'24 is May 13. Registration details can be found here.

Update [May 2024]: The updated schedule of the workshop (including the order of talks) can be found here.

To stay tuned for more information, follow us on Twitter @learn_to_cmpress.

Keynote Speakers

Dr. Johannes Ballé

Google Research

Prof. José Miguel Hernández-Lobato

University of Cambridge

Prof. Shirin Jalali

Rutgers University

Dr. Lucas Theis

Google DeepMind

Organizers

Prof. Elza Erkip

New York University

Ezgi Ozyilkan

New York University

Prof. Aaron B. Wagner

Cornell University

Questions

Contact us at learn.to.compress.workshop@gmail.com.
Follow us on Twitter @learn_to_cmpress.

Special thanks to Ugur Y. Yavuz for technical help with the website.