CVDec 16, 2020

Towards Recognizing New Semantic Concepts in New Visual Domains

arXiv:2012.09058v1
AI Analysis

This work addresses the fundamental problem of deep learning models' inability to generalize to novel visual domains and semantic concepts, which is a critical limitation for real-world AI applications. It is an incremental step towards more adaptable and robust AI systems.

This thesis explores methods for deep learning models to recognize new semantic concepts in new visual domains, addressing the limitations of models trained on finite datasets. It investigates solutions for generalizing to new visual domains through knowledge transfer and extending pretrained models to new semantic concepts without original training data, including a novel approach for recognizing unseen concepts in unseen domains via domain and semantic mixing.

Deep learning models heavily rely on large scale annotated datasets for training. Unfortunately, datasets cannot capture the infinite variability of the real world, thus neural networks are inherently limited by the restricted visual and semantic information contained in their training set. In this thesis, we argue that it is crucial to design deep architectures that can operate in previously unseen visual domains and recognize novel semantic concepts. In the first part of the thesis, we describe different solutions to enable deep models to generalize to new visual domains, by transferring knowledge from a labeled source domain(s) to a domain (target) where no labeled data are available. We will show how variants of batch-normalization (BN) can be applied to different scenarios, from domain adaptation when source and target are mixtures of multiple latent domains, to domain generalization, continuous domain adaptation, and predictive domain adaptation, where information about the target domain is available only in the form of metadata. In the second part of the thesis, we show how to extend the knowledge of a pretrained deep model to new semantic concepts, without access to the original training set. We address the scenarios of sequential multi-task learning, using transformed task-specific binary masks, open-world recognition, with end-to-end training and enforced clustering, and incremental class learning in semantic segmentation, where we highlight and address the problem of the semantic shift of the background class. In the final part, we tackle a more challenging problem: given images of multiple domains and semantic categories (with their attributes), how to build a model that recognizes images of unseen concepts in unseen domains? We also propose an approach based on domain and semantic mixing of inputs and features, which is a first, promising step towards solving this problem.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes