Authors: Atsuyuki Miyai, Jingkang Yang, Jingyang Zhang, Yifei Ming, Yueqian Lin, Qing Yu, Go Irie, Shafiq Joty, Yixuan Li, Hai Li, Ziwei Liu, Toshihiko Yamasaki, Kiyoharu Aizawa
Category: Computer Vision and Pattern Recognition, Artificial Intelligence, Machine Learning
ArXiv: http://arxiv.org/abs/2407.21794v1
Abstract: Detecting out-of-distribution (OOD) samples is crucial for ensuring the safety of machine learning systems and has shaped the field of OOD detection. Meanwhile, several other problems are closely related to OOD detection, including anomaly detection (AD), novelty detection (ND), open set recognition (OSR), and outlier detection (OD). To unify these problems, a generalized OOD detection framework was proposed, taxonomically categorizing these five problems. However, Vision Language Models (VLMs) such as CLIP have significantly changed the paradigm and blurred the boundaries between these fields, again confusing researchers. In this survey, we first present a generalized OOD detection v2, encapsulating the evolution of AD, ND, OSR, OOD detection, and OD in the VLM era. Our framework reveals that, with some field inactivity and integration, the demanding challenges have become OOD detection and AD. In addition, we also highlight the significant shift in the definition, problem settings, and benchmarks; we thus feature a comprehensive review of the methodology for OOD detection, including the discussion over other related tasks to clarify their relationship to OOD detection. Finally, we explore the advancements in the emerging Large Vision Language Model (LVLM) era, such as GPT-4V. We conclude this survey with open challenges and future directions.
Summary: The document is a survey titled “Generalized Out-of-Distribution Detection and Beyond in Vision Language Model Era: A Survey,” authored by Atsuyuki Miyai and others. The survey discusses the importance of detecting out-of-distribution (OOD) samples for ensuring the safety of machine learning systems. It outlines a generalized OOD detection framework that unifies and categorizes related problems such as anomaly detection (AD), novelty detection (ND), open set recognition (OSR), and outlier detection (OD).
**Problems Identified:**
– Confusion and overlap in the definitions and approaches of OOD detection and related tasks due to subtle differences in definitions.
– The emergence of Vision Language Models (VLMs), such as CLIP, which has changed the recognition paradigm and blurred the distinctions between these problems, creating challenges in directions for research.
**How to Solve the Problems:**
– The authors propose a generalized OOD detection framework to clarify definitions and relationships among OOD detection and related tasks.
– The survey revisits and presents an updated framework, “generalized OOD detection v2,” to encapsulate the evolution of these problems in the VLM era, highlighting changes in definitions, problem settings, and benchmarks.
**Results:**
– It reveals that the demanding challenges in the VLM era are OOD detection and AD.
– The survey provides a comprehensive review of methodologies for OOD detection and discusses related tasks to clarify their relationship to OOD detection.
– It explores advancements in the emerging Large Vision Language Model (LVLM) era, such as GPT-4V. – The survey concludes with open challenges and future directions in the field.
The document itself is structured as a research paper, detailing the introduction, methodology, review of related tasks, and a conclusion with future perspectives. The authors also provide a GitHub resource link for further information related to the survey.
Review：
Q:What is OOD Prompts?
A:Is this a question or are you just testing functionality on the site? It says that you are an author already.
Related video：