Compressing everything with neural networks and the path to production

Talk-Titel: Compressing everything with neural networks and the path to production
Abstract: After a surge of interest in using compressive autoencoders for image compression, the neural compression field has diversified. Now, we also have neural networks that compress videos, audio, 3D data, and even neural networks themselves. In contrast to other ML fields, compression still has a large community that researches non-neural network-based approaches. These remain state-of-the-art for some modalities and compute budgets. I will talk about the unique promises of the neural methods and about the challenges in deploying the algorithms to production

Speaker: Fabian Mentzer (Google Zürich)
Bio: With a Ph.D. from ETH Zürich and a thesis focused on the cutting-edge field of Neural Image Compression, Fabian is a leading authority on both lossy and lossless algorithms. As a Senior Research Scientist at Google Zürich, he applies his expertise to develop groundbreaking solutions for the tech giant.
But Fabian's talents don't stop there. He's also a passionate coffee connoisseur, always seeking out the best brews and brewing methods. And when he's not in the lab, you can find him creating intricate soundscapes with his modular synthesizer.

We are with IMBIT this time and meet in the Nexus Lab. Averbis provide beers and bezels after Fabians' talk and we invite you to stay for a chat.
Thanks for the support from Averbis and IMBIT!