Zero-Shot Learning - A Comprehensive Evaluation of the Good, the Bad and the Ugly

نویسندگان

  • Yongqin Xian
  • Christoph H. Lampert
  • Bernt Schiele
  • Zeynep Akata
چکیده

Due to the importance of zero-shot learning, i.e. classifying images where there is a lack of labeled training data, the number of proposed approaches has recently increased steadily. We argue that it is time to take a step back and to analyze the status quo of the area. The purpose of this paper is three-fold. First, given the fact that there is no agreed upon zero-shot learning benchmark, we first define a new benchmark by unifying both the evaluation protocols and data splits of publicly available datasets used for this task. This is an important contribution as published results are often not comparable and sometimes even flawed due to, e.g. pre-training on zero-shot test classes. Moreover, we propose a new zero-shot learning dataset, the Animals with Attributes 2 (AWA2) dataset which we make publicly available both in terms of image features and the images themselves. Second, we compare and analyze a significant number of the state-of-the-art methods in depth, both in the classic zero-shot setting but also in the more realistic generalized zero-shot setting. Finally, we discuss in detail the limitations of the current status of the area which can be taken as a basis for advancing it.

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Zero-Shot Learning - The Good, the Bad and the Ugly

Due to the importance of zero-shot learning, the number of proposed approaches has increased steadily recently. We argue that it is time to take a step back and to analyze the status quo of the area. The purpose of this paper is threefold. First, given the fact that there is no agreed upon zeroshot learning benchmark, we first define a new benchmark by unifying both the evaluation protocols and...

متن کامل

Competition in Healthcare: Good, Bad or Ugly?

The role of competition in healthcare is much debated. Despite a wealth of international experience in relation to competition, evidence is mixed and contested and the debate about the potential role for competition is often polarised. This paper considers briefly some of the reasons for this, focusing on what is meant by “competition in healthcare” and why it is more valuable to think about th...

متن کامل

Optimisation of Healthcare Contracts: Tensions Between Standardisation and Innovation; Comment on “Competition in Healthcare: Good, Bad or Ugly?”

An important determinant of health system performance is contracting. Providers often respond to financial incentives, despite the ethical underpinnings of medicine, and payers can craft contracts to influence performance. Yet contracting is highly imperfect in both single-payer and multi-payer health systems. Arguably, in a competitive, multi-payer environment, contractual innovation may occur...

متن کامل

Good, bad and ugly: Exploring the Machiavellian power dynamics of leadership in medical education

Introduction: Medical education requires participation of variousstakeholders and this contributes to power dynamics operating atmultiple levels. Personality traits of an individual can affect thesmooth execution of the educational programmes and eventuallythe professionalism of the environment. With the increased focuson leadership traits in medical education and collaboration inhealth care se...

متن کامل

"The Good, The Bad And The Ugly": Evaluation of Wi-Fi Steganography

In this paper we propose a new method for the evaluation of network steganography algorithms based on the new concept of “the moving observer”. We considered three levels of undetectability named: “good”, “bad”, and “ugly”. To illustrate this method we chose Wi-Fi steganography as a solid family of information hiding protocols. We present the state of the art in this area covering well-known hi...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:
  • CoRR

دوره abs/1707.00600  شماره 

صفحات  -

تاریخ انتشار 2017