No abstract

No abstract

User intent detection plays a critical role in question-answering and dialog systems. Most previous works treat intent detection as a classification problem where utterances are labeled with predefined intents. However, it is labor-intensive and time-consuming to label users' utterances as intents are diversely expressed and novel intents will continually be involved. Instead, we study the zero-shot intent detection problem, which aims to detect emerging user intents where no labeled utterances are currently available. We propose two capsule-based architectures: INTENT-CAPSNET that extracts semantic features from utterances and aggregates them to discriminate existing intents, and INTENTCAPSNET-ZSL which gives INTENTCAPSNET the zero-shot learning ability to discriminate emerging intents via knowledge transfer from existing intents. Experiments on two real-world datasets show that our model not only can better discriminate diversely expressed existing intents, but is also able to discriminate emerging intents when no labeled utterances are available. * Indicates Equal Contribution. Previously avilable on http://doi.

No abstract

Deep learning has attracted intense interest in Prognostics and Health Management (PHM), because of its enormous representing power, automated feature learning capability and best-in-class performance in solving complex problems. This paper surveys recent advancements in PHM methodologies using deep learning with the aim of identifying research gaps and suggesting further improvements. After a brief introduction to several deep learning models, we review and analyze applications of fault detection, diagnosis and prognosis using deep learning. The survey validates the universal applicability of deep learning to various types of input in PHM, including vibration, imagery, time-series and structured data. It also reveals that deep learning provides a one-fits-all framework for the primary PHM subfields: fault detection uses either reconstruction error or stacks a binary classifier on top of the network to detect anomalies; fault diagnosis typically adds a soft-max layer to perform multi-class classification; prognosis adds a continuous regression layer to predict remaining useful life. The general framework suggests the possibility of transfer learning across PHM applications. The survey reveals some common properties and identifies the research gaps in each PHM subfield. It concludes by summarizing some major challenges and potential opportunities in the domain.

Nowadays, short texts are very prevalent in various web applications, such as microblogs, instant messages. The severe sparsity of short texts hinders existing topic models to learn reliable topics. In this paper, we propose a novel way to tackle this problem. The key idea is to learn topics by exploring term correlation data, rather than the high-dimensional and sparse term occurrence information in documents. Such term correlation data is less sparse and more stable with the increase of the collection size, and can well capture the necessary information for topic learning. To obtain reliable topics from term correlation data, we first introduce a novel way to compute term correlation in short texts by representing each term with its co-occurred terms. Then we formulated the topic learning problem as symmetric non-negative matrix factorization on the term correlation matrix. After learning the topics, we can easily infer the topics of documents. Experimental results on three data sets show that our method provides substantially better performance than the baseline methods.

Meaning Representation (AMR) parsing aims at abstracting away from the syntactic realization of a sentence, and denoting only its meaning in a canonical form. As such, it is ideal for paraphrase detection, a problem in which one is required to specify whether two sentences have the same meaning. We show that naïve use of AMR in paraphrase detection is not necessarily useful, and turn to describe a technique based on latent semantic analysis in combination with AMR parsing that significantly advances state-of-the-art results in paraphrase detection for the Microsoft Research Paraphrase Corpus. Our best results in the transductive setting are 86.6% for accuracy and 90.0% for F 1 measure.

No abstract