[论文] Aspect Based Sentiment Analysis with Gated Convolutional Networks

Aspect Based Sentiment Analysis with Gated Convolutional Networks

 

摘要：

Aspect based sentiment analysis (ABSA)

two subtasks:

aspect-category sentiment analysis (ACSA) and aspect-term sentiment analysis (ATSA)

 

大多数先前的方法采用长期短期记忆和注意机制来预测相关目标的情绪极性，这通常是复杂的并且需要更多的训练时间。我们提出了一种基于卷积神经网络和门控机制的模型，该模型更加准确和高效。

 数据集：SemEval

 

介绍：

ABSA不是预测整体情绪极性。具体而言，对文本中方面类别或目标实体的情感极性感兴趣。

The goal of ACSA is to predict the sentiment polarity with regard to the given aspect, which is one of a few predefined categories. On the other hand, the goal of ATSA is to identify the sentiment polarity concerning the target entities that appear in the text instead, which could be a multi-word phrase or a single word.

 

Attention-based LSTM:

 It first computes the alignment scores between context vectors and target vector; then carry out a weighted sum with the scores and the context vectors.

 LSTM and attention layer are very time-consuming during training.

we propose a fast and effective neural network for ACSA and ATSA based on convolutions and gating mechanisms, which has much less training time than LSTM based networks, but with better accuracy.

 

 two separate convolutional layers on the top of the embedding layer, whose outputs are combined by novel gating units. Convolutional layers with multiple filters can efficiently extract n-gram features at many granularities on each receptive field. The proposed gating units have two nonlinear gates, each of which is connected to one convolutional layer. With the given aspect information, they can selectively extract aspect-specific sentiment information for sentiment prediction.

To the best of our knowledge, no CNN-based model has been proposed for aspect based sentiment analysis so far.

 

相关工作：

2.1： Neural Networks

Recursive Neural Tensor Network

Tree-LSTM

Recurrent Neural Networks (RNNs)

LSTM

GRU

2.2： Aspect based Sentiment Analysis

Aspect-Term Sentiment Analysis

Aspect-Category Sentiment Analysis

 

Gated Convolutional Network with Aspect Embedding：

Each convolutional filter computes n-gram features at different granularities from the embedding vectors at each position individually.

Moreover, our model is equipped with two kinds of effective filtering mechanisms: the gating units on top of the convolutional layers and the max pooling layer, both of which can accurately generate and select aspect-related sentiment features.

The outputs of two gates are element-wisely multiplied for the max pooling layer.

 

va is the embedding vector of the given aspect category in ACSA or computed by another CNN over aspect terms in ATSA.

si and ai are responsible for generating sentiment features and aspect features respectively. The above max-over-time pooling layer

 

Gating Mechanisms：

it can output a similarity score according to the relevance between the given aspect information va and the aspect feature ai at position t.

GTRU is more effective than these two gating units.

 

GCAE on ATSA：

ATSA task is defined to predict the sentiment polarity of the aspect terms in the given sentence. We simply extend GCAE by adding a small convolutional layer on aspect terms, as shown in Figure 2.

In ACSA, the aspect information controlling the flow of sentiment features in GTRU is from one aspect word; while in ATSA, such information is provided by a small CNN on aspect terms [wi, wi+1, . . . , wi+k].