Semantic Novelty Detection via Relational Reasoning
This work addresses the challenge of recognizing unknown categories in test data for safety-critical domains, offering a solution that is efficient and avoids the need for fine-tuning under constraints like privacy or computational limits.
The paper tackles the problem of semantic novelty detection, which is crucial for safety-critical applications like autonomous driving and healthcare, by proposing a novel representation learning paradigm based on relational reasoning that learns to measure semantic similarity, resulting in a method that can convert closed-set recognition models into reliable open-set ones without requiring fine-tuning on known categories.
Semantic novelty detection aims at discovering unknown categories in the test data. This task is particularly relevant in safety-critical applications, such as autonomous driving or healthcare, where it is crucial to recognize unknown objects at deployment time and issue a warning to the user accordingly. Despite the impressive advancements of deep learning research, existing models still need a finetuning stage on the known categories in order to recognize the unknown ones. This could be prohibitive when privacy rules limit data access, or in case of strict memory and computational constraints (e.g. edge computing). We claim that a tailored representation learning strategy may be the right solution for effective and efficient semantic novelty detection. Besides extensively testing state-of-the-art approaches for this task, we propose a novel representation learning paradigm based on relational reasoning. It focuses on learning how to measure semantic similarity rather than recognizing known categories. Our experiments show that this knowledge is directly transferable to a wide range of scenarios, and it can be exploited as a plug-and-play module to convert closed-set recognition models into reliable open-set ones.