No abstract

No abstract

Sentiment analysis, which addresses the computational treatment of opinion, sentiment, and subjectivity in text, has received considerable attention in recent years. In contrast to the traditional coarse-grained sentiment analysis tasks, such as document-level sentiment classification, we are interested in the fine-grained aspect-based sentiment analysis that aims to identify aspects that users comment on and these aspects' polarities. Aspect-based sentiment analysis relies heavily on syntactic features. However, the reviews that this task focuses on are natural and spontaneous, thus posing a challenge to syntactic parsers. In this paper, we address this problem by proposing a framework of adding a sentiment sentence compression (Sent Comp) step before performing the aspect-based sentiment analysis. Different from the previous sentence compression model for common news sentences, Sent Comp seeks to remove the sentiment-unnecessary information for sentiment analysis, thereby compressing a complicated sentiment sentence into one that is shorter and easier to parse. We apply a discriminative conditional random field model, with certain special features, to automatically compress sentiment sentences. Using the Chinese corpora of four product domains, Sent Comp significantly improves the performance of the aspect-based sentiment analysis. The features proposed for Sent Comp, especially the potential semantic features, are useful for sentiment sentence compression.Index Terms-sentiment analysis, aspect-based sentiment analysis, sentence compression, potential semantic features.1 In this paper, we focus on Chinese sentiment analysis task. The similar method can also be used into other languages. 2 A Chinese natural language processing toolkit, Language Technology Platform (LTP) [10], was used as our dependency parser. More information about the syntactic relations can be found from their paper. The state-of-the-art graph-based dependency parsing model, in the toolkit, was trained on Chinese Dependency Treebank 1.0 (LDC2012T05).

Domain adaptation is an important problem in named entity recognition (NER). NER classifiers usually lose accuracy in the domain transfer due to the different data distribution between the source and the target domains. The major reason for performance degrading is that each entity type often has lots of domainspecific term representations in the different domains. The existing approaches usually need an amount of labeled target domain data for tuning the original model. However, it is a labor-intensive and time-consuming task to build annotated training data set for every target domain. We present a domain adaptation method with latent semantic association (LaSA). This method effectively overcomes the data distribution difference without leveraging any labeled target domain data. LaSA model is constructed to capture latent semantic association among words from the unlabeled corpus. It groups words into a set of concepts according to the related context snippets. In the domain transfer, the original term spaces of both domains are projected to a concept space using LaSA model at first, then the original NER model is tuned based on the semantic association features. Experimental results on English and Chinese corpus show that LaSA-based domain adaptation significantly enhances the performance of NER.

No abstract

Cross-domain sentiment classification has drawn much attention in recent years. Most existing approaches focus on learning domaininvariant representations in both the source and target domains, while few of them pay attention to the domain-specific information. Despite the non-transferability of the domainspecific information, simultaneously learning domain-dependent representations can facilitate the learning of domain-invariant representations. In this paper, we focus on aspectlevel cross-domain sentiment classification, and propose to distill the domain-invariant sentiment features with the help of an orthogonal domain-dependent task, i.e. aspect detection, which is built on the aspects varying widely in different domains. We conduct extensive experiments on three public datasets and the experimental results demonstrate the effectiveness of our method.

Aspect category detection (ACD) in sentiment analysis aims to identify the aspect categories mentioned in a sentence. In this paper, we formulate ACD in the few-shot learning scenario. However, existing few-shot learning approaches mainly focus on single-label predictions. These methods can not work well for the ACD task since a sentence may contain multiple aspect categories. Therefore, we propose a multi-label few-shot learning method based on the prototypical network. To alleviate the noise, we design two effective attention mechanisms. The support-set attention aims to extract better prototypes by removing irrelevant aspects. The query-set attention computes multiple prototype-specific representations for each query instance, which are then used to compute accurate distances with the corresponding prototypes. To achieve multilabel inference, we further learn a dynamic threshold per instance by a policy network. Extensive experimental results on three datasets demonstrate that the proposed method significantly outperforms strong baselines. * Shiwan Zhao is the corresponding author. † The work was (partially) done in IBM. (A) room_cleanliness Support set (B) staff_owner Query set (A) and (C) (B) (B) (C) hotel1) Okay, so it is a cute chain hotel.2) I really don't see how people are giving this hotel such high ratings.1) Hotel is just plain dirty.2) The owners are extremely smart and worldly3) Not a typical customer service response, especially from the owner ! 1) I think the salon has problems starting with the owner.2) The owner is very nice. 1) Cleanliness was great, and the food was really good. 2) People have mentioned, bed bugs on yelp !!

Address standardization is a very challenging task in data cleansing. To provide better customer relationship management and business intelligence for customer-oriented cooperates, millions of free-text addresses need to be converted to a standard format for data integration, de-duplication and householding. Existing commercial tools usually employ lots of hand-craft, domain-specific rules and reference data dictionary of cities, states etc. These rules work better for the region they are designed. However, rule-based methods usually require more human efforts to rewrite these rules for each new domain since address data are very irregular and varied with countries and regions. Supervised learning methods usually are more adaptable than rule-based approaches. However, supervised methods need large-scale labeled training data. It is a labor-intensive and time-consuming task to build a large-scale annotated corpus for each target domain. For minimizing human efforts and the size of labeled training data set, we present a free-text address standardization method with latent semantic association (LaSA). LaSA model is constructed to capture latent semantic association among words from the unlabeled corpus. The original term space of the target domain is projected to a concept space using LaSA model at first, then the address standardization model is active learned from LaSA features and informative samples. The proposed method effectively captures the data distribution of the domain. Experimental results on large-scale English and Chinese corpus show that the proposed method significantly enhances the performance of standardization with less efforts and training data.