RECORD
Graph Adaptive Semantic Transfer for Cross-domain Sentiment Classification
Computation and Language · Submitted 18 May 2022
abstract compiled · 2026-09-23 · no language-model review
The authors' main claim Extensive experiments on four public datasets indicate that GAST achieves comparable effectiveness to a range of state-of-the-art models.
The compiled record
The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
Open the compiled record
The abstract, sentence by sentence
Cross-domain sentiment classification (CDSC) aims to use the transferable semantics learned from the source domain to predict the sentiment of reviews in the unlabeled target domain. Existing studies in this task attach more attention to the sequence modeling of sentences while largely ignoring the rich domain-invariant semantics embedded in graph structures (i.e., the part-of-speech tags and dependency relations). As an important aspect of exploring characteristics of language comprehension, adaptive graph representations have played an essential role in recent years. To this end, in the paper, we aim to explore the possibility of learning invariant semantic features from graph-like structures in CDSC. Specifically, we present Graph Adaptive Semantic Transfer (GAST) model, an adaptive syntactic graph embedding method that is able to learn domain-invariant semantics from both word sequences and syntactic graphs. More specifically, we first raise a POS-Transformer module to extract sequential semantic features from the word sequences as well as the part-of-speech tags. Then, we design a Hybrid Graph Attention (HGAT) module to generate syntax-based semantic features by considering the transferable dependency relations. Finally, we devise an Integrated aDaptive Strategy (IDS) to guide the joint learning process of both modules. Extensive experiments on four public datasets indicate that GAST achieves comparable effectiveness to a range of state-of-the-art models.
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | Cross-domain sentiment classification (CDSC) aims to use the transferable semantics learned from the source domain to predict the sentiment of reviews in the unlabeled target domain. | background 1.00 | 0.04 | 0.03 | asserted 0.99 |
| s1 | Existing studies in this task attach more attention to the sequence modeling of sentences while largely ignoring the rich domain-invariant semantics embedded in graph structures (i.e., the part-of-speech tags and dependency relations). | gap 0.99 | 0.07 | 0.03 | asserted 1.00 |
| s2 | As an important aspect of exploring characteristics of language comprehension, adaptive graph representations have played an essential role in recent years. | background 1.00 | 0.04 | 0.02 | asserted 1.00 |
| s3 | To this end, in the paper, we aim to explore the possibility of learning invariant semantic features from graph-like structures in CDSC. | contribution 0.91 | 0.30 | 0.05 | hedged possibility 0.90 |
| s4 | Specifically, we present Graph Adaptive Semantic Transfer (GAST) model, an adaptive syntactic graph embedding method that is able to learn domain-invariant semantics from both word sequences and syntactic graphs. | contribution 0.71 | 0.93 | 0.12 | asserted 0.99 |
| s5 | More specifically, we first raise a POS-Transformer module to extract sequential semantic features from the word sequences as well as the part-of-speech tags. | method 0.98 | 0.93 | 0.03 | asserted 0.99 |
| s6 | Then, we design a Hybrid Graph Attention (HGAT) module to generate syntax-based semantic features by considering the transferable dependency relations. | method 0.98 | 0.93 | 0.03 | asserted 0.99 |
| s7 | Finally, we devise an Integrated aDaptive Strategy (IDS) to guide the joint learning process of both modules. | method 0.96 | 0.94 | 0.03 | asserted 0.98 |
| s8 | Extensive experiments on four public datasets indicate that GAST achieves comparable effectiveness to a range of state-of-the-art models. | result 1.00 | 0.96 | 0.80 | hedged possibility 0.61 |
What was examined, and what was not
- arXiv listing abstract v1
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
The paper
Original abstract
Cross-domain sentiment classification (CDSC) aims to use the transferable semantics learned from the source domain to predict the sentiment of reviews in the unlabeled target domain. Existing studies in this task attach more attention to the sequence modeling of sentences while largely ignoring the rich domain-invariant semantics embedded in graph structures (i.e., the part-of-speech tags and dependency relations). As an important aspect of exploring characteristics of language comprehension, adaptive graph representations have played an essential role in recent years. To this end, in the paper, we aim to explore the possibility of learning invariant semantic features from graph-like structures in CDSC. Specifically, we present Graph Adaptive Semantic Transfer (GAST) model, an adaptive syntactic graph embedding method that is able to learn domain-invariant semantics from both word sequences and syntactic graphs. More specifically, we first raise a POS-Transformer module to extract sequential semantic features from the word sequences as well as the part-of-speech tags. Then, we design a Hybrid Graph Attention (HGAT) module to generate syntax-based semantic features by considering the transferable dependency relations. Finally, we devise an Integrated aDaptive Strategy (IDS) to guide the joint learning process of both modules. Extensive experiments on four public datasets indicate that GAST achieves comparable effectiveness to a range of state-of-the-art models.
Human review
How the literature uses it
What this paper does with the arXiv papers it cites
What it cites · 1 sentence without a confident label
Uses the method or tool of 2 papers
-
Inspired by [ref] , we minimize the entropy penalty to disambiguate the positive and negative samples over the unlabeled data from both domains.
Cross-Domain Sentiment Classification with In-Domain Contrastive Learning · Integrated Adaptive Strategy · 0.90
-
To evaluate the impact of different syntactic features, inspired by [ref] , we conduct a comparative experiment using two well-known dependency parsers, i.e., Stanford Parser [ref] and Biaffine Parser[https://github.com/allenai/allennlp.] [ref] , to construct our syntactic graph.
Deep Biaffine Attention for Neural Dependency Parsing · Influence of Adaptive Syntactic Graph · 0.84
Cites as background 3 papers
-
Recent years, many researchers studied the problem [ref] through the deep neural networks.
Learning to Share by Masking the Non-shared for Multi-domain Sentiment Classification · Cross-domain Sentiment Classification · 0.96
-
Most of these studies [ref] focus on extracting features from the word sequences and embed the sequential features into deep semantic representations through various methods, e.g., memory network [ref] , recurrent neural network [ref] , attention mechanism [ref] , and the large-scale pre-trained models [ref] .
Unsupervised Domain Adaptation by Backpropagation · Introduction · 0.90
-
Ghosal et al. [ref] enriched the semantics of a document by exploring the role of external commonsense knowledge.
KinGDOM: Knowledge-Guided DOMain adaptation for sentiment analysis · Cross-domain Sentiment Classification · 0.84
Discussion
Know something about this paper? Leave a comment, or highlight any passage above to comment on it.