Research map / Label-efficient and robust learning / Recognition and long-tail
Word-Level Contextual Sentiment Analysis with Interpretability
Tomoki Ito, Kota Tsubouchi, Hiroki Sakaji, Tatsuo Yamashita, Kiyoshi Izumi
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add lexical initialization of word sentiment from a small dictionary (ablation) | SINN · WOSL, EcoRev I | macro F1: .492 -> .839 | Improved |
| add lexical initialization (ablation) | SINN · LWCL, EcoRev I | macro F1: .350 -> .800 | Improved |
| add lexical initialization (ablation) | SINN · GWCL, EcoRev I | Pearson correlation: .053 -> .588 | Improved |
| add lexical initialization (ablation) | SINN · WOSL, Sentiment 140 | macro F1: .444 -> .737 | Improved |
| reduce dictionary words in LEXIL from 200 to 50 (ablation) | SINN · WOSL, EcoRev I | macro F1: .839 -> .844 | No clear effect |
| reduce dictionary words in LEXIL from 200 to 50 (ablation) | SINN · LWCL, EcoRev I | macro F1: .800 -> .776 (100 words: .815) | Mixed |
| reduce dictionary words in LEXIL from 200 to 50 (ablation) | SINN · Word-level polarity, EcoRev I | macro F1: .719 -> .695 | Hurt |
| use SINN instead of best prior GWCL attention method | GWCL · Yahoo | Pearson correlation: .405 (LBSA) vs .285 (SINN 100) | Hurt |
| use SINN instead of IntGrad+RNN | WCSA · Document-level polarity, EcoRev II | macro F1: .929 -> .944 | Improved |
Closest papers by meaning
- How recurrent networks implement contextual processing in sentiment analysis (ICML 2020 · similarity 0.85)
- Human-Level Interpretable Learning for Aspect-Based Sentiment Analysis (AAAI 2021 · similarity 0.84)
- Beyond Word Importance: Contextual Decomposition to Extract Interactions from LSTMs (ICLR 2018 · similarity 0.83)
- Learning Sentiment-Specific Word Embedding via Global Sentiment Representation (AAAI 2018 · similarity 0.83)
- You Only Read Once: Constituency-Oriented Relational Graph Convolutional Network for Multi-Aspect Multi-Sentiment Classification (AAAI 2024 · similarity 0.82)
- Aspect-Based Sentiment Analysis with Explicit Sentiment Augmentations (AAAI 2024 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
