Zero-shot recognition with unreliable attributes
نویسندگان
چکیده
In this document, we provide supplementary material for our NIPS 2014 paper “Zero-Shot Recognition with Unreliable Attributes”. Sec 1 shows how unlearnable attributes are avoided by our method. Sec 2 discusses the details of the signature uncertainty model introduced in Sec 3.2.3 of the paper. Sec 3 gives additional details for our controlled noise experiments (Sec 4.1 of the paper). Sec 4 lists the 10 SUN database test classes chosen at random. Sec 5 shows more few-shot results, as a continuation of Sec 4.2 in the paper. Sec 6 contains pseudocode for our proposed method, and Sec 7 contains schematics illustrating our method and its ablated variants.
منابع مشابه
Zero-Shot Activity Recognition with Verb Attribute Induction
In this paper, we investigate large-scale zero-shot activity recognition by modeling the visual and linguistic attributes of action verbs. For example, the verb “salute” has several properties, such as being a light movement, a social act, and short in duration. We use these attributes as the internal mapping between visual and textual representations to reason about a previously unseen action....
متن کاملAlternative Semantic Representations for Zero-Shot Human Action Recognition
A proper semantic representation for encoding side information is key to the success of zero-shot learning. In this paper, we explore two alternative semantic representations especially for zero-shot human action recognition: textual descriptions of human actions and deep features extracted from still images relevant to human actions. Such side information are accessible on Web with little cost...
متن کاملImproving Semantic Embedding Consistency by Metric Learning for Zero-Shot Classiffication
This paper addresses the task of zero-shot image classification. The key contribution of the proposed approach is to control the semantic embedding of images – one of the main ingredients of zero-shot learning – by formulating it as a metric learning problem. The optimized empirical criterion associates two types of sub-task constraints: metric discriminating capacity and accurate attribute pre...
متن کاملZero-Shot Learning by Generating Pseudo Feature Representations
Zero-shot learning (ZSL) is a challenging task aiming at recognizing novel classes without any training instances. In this paper we present a simple but high-performance ZSL approach by generating pseudo feature representations (GPFR). Given the dataset of seen classes and side information of unseen classes (e.g. attributes), we synthesize feature-level pseudo representations for novel concepts...
متن کاملPreserving Semantic Relations for Zero-Shot Learning
Zero-shot learning has gained popularity due to its potential to scale recognition models without requiring additional training data. This is usually achieved by associating categories with their semantic information like attributes. However, we believe that the potential offered by this paradigm is not yet fully exploited. In this work, we propose to utilize the structure of the space spanned ...
متن کامل