NVIDIA pioneer to commence 4th annual modern AI seminar series


Jan Kautz, Vice President of Learning and Perception at NVIDIA, will kick off NYU Tandon’s fourth annual Modern Artificial Intelligence (AI) series by offering a glimpse into recent progress in generative models, particularly a form of machine learning adept at image categorization: generative adversarial networks (GANs).

Kautz’s presentation, ”Generative Models for Image Synthesis,” will focus on recent progress in the field, advances enabling the translation of images between domains in an unsupervised fashion, the synthesis of completely new images, and the detection of defects in the models’ own training data. 

Sponsored by the Department of Electrical and Computer Engineering (ECE) and organized by Professor Anna Choromanska, the series has featured experts from the likes of Google, Facebook, Microsoft, DeepMind, J.P. Morgan Chase, Columbia University, Carnegie Mellon University, and the California Institute of Technology (CalTech). 

“Once again we have built an impressive schedule of luminaries in the field of artificial intelligence who help us take in the vast scope of current work in AI and speculate on what’s coming next,” says Choromanska. 

Following Kautz are Gabor Lugosi, Research Professor at the Pompeu Fabra University Department of Economics and the Barcelona Graduate School of Economics; Nicolò Cesa-Bianchi, Professor of Computer Science at the University of Milan; and Robert Schapire, Professor of Computer Science at Princeton University and Partner Researcher at Microsoft Research.

  • Lugosi will explore the “archeology” of data structures called “random trees,” presenting results demonstrating that even in gigantic networks a lot of information is preserved from the very early days.
  • Cesa-Biachi will give a presentation on machine learning with partial feedback, particularly the extent to which different forms of partial feedback can affect the learning ability of online algorithms
  • Schapire’s presentation will examine a phenomenon called the “contextual bandits” where the learner must repeatedly decide which action to take in response to an observed context, and is then permitted to observe the received reward, but only for the chosen action.

Past series presenters have included Nobel Laureate Eric Kanel, Turing Award recipient and Facebook’s Vice President and Chief AI Scientist Yann LeCun, another Turing Award recipient and Founder and Scientific Director of the Montreal Institute for Learning Algorithm (MILA) Yoshua Bengio, and famed venture capitalist and director of Microsoft Research Asia, Kai-Fu Lee.

The live-streamed lecture series, held in Brooklyn, New York at 370 Jay Street and NYU Tandon’s MakerSpace at 6 MetroTech Center, has garnered a far-reaching audience numbering in the thousands. So far, 13 seminars have been archived on the YouTube playlist.