Research map / Architectures and efficiency / Attention and transformers
Chinese LIWC Lexicon Expansion via Hierarchical Classification of Word Embeddings with Sememe Attention
Xiangkai Zeng, Cheng Yang, Cunchao Tu, Zhiyuan Liu, Maosong Sun
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 sememe attention to hierarchical decoder (ablation) | Hierarchical Decoder · Chinese LIWC expansion, overall | Micro-F1: 0.7023 -> 0.7224 | Improved |
| add sememe attention (ablation) | Hierarchical Decoder · Chinese LIWC expansion, overall | W-M-F1: 0.7000 -> 0.7204 | Improved |
| add sememe attention (ablation) | Hierarchical Decoder · Chinese LIWC expansion, Level 2 | Micro-F1: 0.6658 -> 0.6927 | Improved |
| add sememe attention (ablation) | Hierarchical Decoder · Chinese LIWC expansion, Level 3 | Micro-F1: 0.6113 -> 0.6270 | Improved |
| add sememe attention (ablation) | Hierarchical Decoder · Chinese LIWC expansion, overall | Precision: 0.7216 -> 0.7473 | Improved |
| sememe attention on individual words (case study) (ablation) | Hierarchical Decoder · Chinese LIWC case study (alliance, teardrop) | label correctness | Mixed |
| replace top-down classifiers with seq2seq hierarchical decoder | LIWC expansion · Chinese LIWC expansion, overall | Micro-F1: 0.6283 -> 0.7023 | Improved |
| HDSA vs CSSA recall at level 1 | HDSA · Chinese LIWC expansion, Level 1 | Recall: 0.7973 -> 0.7311 | Hurt |
| raise threshold delta from -2.0 to -1.0 (ablation) | HDSA · Chinese LIWC expansion, overall | Micro-F1: precision 71.8% -> 79.1%, recall 72.9% -> 64.5% | No clear effect |
Closest papers by meaning
- Towards Building a Multilingual Sememe Knowledge Base: Predicting Sememes for BabelNet Synsets (AAAI 2020 · similarity 0.84)
- cw2vec: Learning Chinese Word Embeddings with Stroke n-gram Information (AAAI 2018 · similarity 0.83)
- A Radical-Aware Attention-Based Model for Chinese Text Classification (AAAI 2019 · similarity 0.83)
- WenMind: A Comprehensive Benchmark for Evaluating Large Language Models in Chinese Classical Literature and Language Arts (NeurIPS 2024 · similarity 0.82)
- MMTEB: Massive Multilingual Text Embedding Benchmark (ICLR 2025 · similarity 0.82)
- SeqGPT: An Out-of-the-Box Large Language Model for Open Domain Sequence Understanding (AAAI 2024 · similarity 0.81)
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.
