Our guest speaker is Thomas Kipf from Google Brain and you are all cordially invited to the CVG Seminar on July 22nd at 2:30 p.m. CET on Zoom (passcode is 809285).

Abstract

The world around us — and our understanding of it — is rich in compositional structure: from atoms and their interactions to objects and entities in our environments. How can we learn models of the world that take this structure into account and generalize to new compositions in systematic ways? This talk focuses on an emerging class of slot-based neural architectures that utilize attention mechanisms to perform perceptual grouping of scenes into objects and abstract entities without direct supervision.
I will briefly introduce the Slot Attention mechanism as a core representative for this class of models and cover our recent extension of Slot Attention to multi-object video (SAVi). I will further give an overview of our new work on 1) extending SAVi to real-world video on the Waymo Open dataset (SAVi++), and 2) using Slot Attention in a scene representation transformer architecture to radically speed up 3D-centric object discovery via novel view synthesis (Object Scene Representation Transformer, OSRT).

Bio

Thomas is a Senior Research Scientist at Google Brain. He obtained his Ph.D. at the University of Amsterdam working with Max Welling. For his Ph.D. thesis on Deep Learning with Graph-Structured Representations, he received the ELLIS Ph.D. Award 2021. He is broadly interested in developing and studying machine learning models that can reason about the rich structure of both the physical and digital world and its combinatorial complexity. This includes topics in graph representation learning, object-centric learning, and causal representation learning.

Our guest speaker is Dmitry Dylov from Skoltech and you are all cordially invited to the CVG Seminar on June 3rd at 2:30 p.m. on Zoom (passcode is 859119) or in-person (room 302 at the institute of informatics)

Bio

Our guest speaker is Greg Yang from Microsoft Research and you are all cordially invited to the CVG Seminar on April 8th at 2:30 p.m. on Zoom (passcode is 757974) or in-person (room 302 at the institute of informatics)

Abstract

You can’t train GPT-3 on a single GPU, much less tune its hyperparameters (HPs)…or so it seems. I’m here to tell you this is not true: you can tune its HPs on a single GPU even if you can’t train it that way! In the first half of this talk, I’ll describe how, in the so-called maximal update parametrization (abbreviated µP), narrow and wide neural networks share the same set of optimal HPs. This lets us tune any large model by just tuning a small version of it — we call this µTransfer. In particular, this allowed us to tune the 6.7 billion parameter version of GPT-3 using only 7% of its pretraining compute budget, and, with some asterisks, we get a performance comparable to the original GPT-3 model with twice the parameter count. In the second half of this talk, I’ll discuss the theoretical reason µP has this special property and the connection to the study of infinite-width neural networks and, more generally, the theory of Tensor Programs. The first half will target general practitioners or empirical researchers in machine learning, while the second half targets those who are more theoretically curious. This talk is based on http://arxiv.org/abs/2203.03466.

Bio

Greg Yang is a researcher at Microsoft Research in Redmond, Washington. He joined MSR after he obtained Bachelor's in Mathematics and Master's degrees in Computer Science from Harvard University, respectively advised by ST Yau and Alexander Rush. He won the Hoopes prize at Harvard for best undergraduate thesis as well as Honorable Mention for the AMS-MAA-SIAM Morgan Prize, the highest honor in the world for an undergraduate in mathematics. He gave an invited talk at International Congress of Chinese Mathematicians 2019.

Our guest speaker is Alaa El-Nouby from Meta AI Research and Inria Paris and you are all cordially invited to the CVG Seminar on March 11th at 3:30 p.m. on Zoom (passcode is 913674).

Abstract

Pre-training models on large scale datasets, like ImageNet, is a standard practice in computer vision. This paradigm is especially effective for tasks with small training sets, for which high-capacity models tend to overfit. In this work, we consider a self-supervised pre-training scenario that only leverages the target task data. We consider datasets, like Stanford Cars, Sketch or COCO, which are order(s) of magnitude smaller than Imagenet. Our study shows that denoising autoencoders, such as BEiT or a variant that we introduce in this paper, are more robust to the type and size of the pre-training data than popular self-supervised methods trained by comparing image embeddings. We obtain competitive performance compared to ImageNet pre-training on a variety of classification datasets, from different domains. On COCO, when pre-training solely using COCO images, the detection and instance segmentation performance surpasses the supervised ImageNet pre-training in a comparable setting.

Bio

Alaa El-Nouby is a PhD student at Meta AI Research and Inria Paris advised by Hervé Jégou and Ivan Laptev. His research interests are metric learning, self-supervised learning and transformers for computer vision. Prior to pursuing his PhD, Alaa received his Msc from the University of Guelph and the Vector institute, advised by Graham Taylor, where he conducted research in spatio-temporal representation learning and text-to-image synthesis with generative models.

Our guest speaker is Arash Vahdat from NVIDIA Research and you are all cordially invited to the CVG Seminar on Feb 10th at 5:30 p.m. on Zoom (passcode is 908626).

Abstract

A wide variety of deep generative models has been developed in the past decade. Yet, these models often struggle with simultaneously addressing three key requirements including high sample quality, mode coverage, and fast sampling. We call the challenge imposed by these requirements the generative learning trilemma, as the existing models often trade between them. Particularly, denoising diffusion models (DDMs) have shown impressive sample quality and diversity, but their expensive sampling does not yet allow them to be applied in many real-world applications. 

Project Pages

https://nvlabs.github.io/LSGM/

https://nvlabs.github.io/denoising-diffusion-gan/

Bio