Aspect-based Sentiment Analysis

1. Aspect based Sentiment Analysis:Aspect-based sentiment analysis is a text analysis technique that breaks down text into aspects (attributes or components of a product or service), and then allocates each one a sentiment level (positive, negative or neutral).
2. Example:Great food but the service was dreadful. So delicious was the noodles but terrible vegetables.
3. syntax information, task-related syntatic structures is the key.

Aspect-Oriented Dependency Tree

1. For input sentences, apply a dependency parser to obtain its dependency tree, rij is the relation from node i to node j.
2. place the target aspect at the root
3. set the nodes with direct connections to the aspect as the children.
4. other relations are retained, instead introduce a virtual relation n:con from the aspect to each corresponding nodes, where n represents the distance between two nodes. If one sentence include multi-aspects, build a unique tree for each aspect

Relational Graph Attention Network

1. encode the dependency trees with its labeled edges
2. update each node representation by aggregating neighborhood node representations.
3. extend the original GAT with additional relational heads. We use these heads as gates to control influence fluence from neighborhood nodes. First map the dependency relations to vector.

Model Training

1. BiLSTM encode tree nodes as hi, another BiLSTM encode aspect words as , its average hidden state as the initial representation ha0 of this root.

Baseline Methods

1. Syntax-aware models
2. Attention-based models
3. other recent methods

Richard Huo

Time will tell, 热爱生活的当代码农，狼人杀新手玩家。