Research map / Architectures and efficiency / RNNs and normalization
Learning to Embed Sentences Using Attentive Recursive Trees
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| use task-trained attentive word scoring instead of tf-idf scoring (ablation) | AR-Tree · SNLI test, 100D | accuracy (%): 82.3 -> 82.8 | Helped |
| use task-trained attentive word scoring instead of tf-idf scoring (ablation) | AR-Tree · SNLI test, 300D | accuracy (%): 84.5 -> 85.5 | Helped |
| use task-trained attentive word scoring instead of tf-idf scoring (ablation) | AR-Tree · SST-2 | accuracy (%): 88.9 -> 90.4 | Helped |
| use task-trained attentive word scoring instead of tf-idf scoring (ablation) | AR-Tree · SST-5 | accuracy (%): 51.3 -> 52.7 | Helped |
| use task-trained attentive word scoring instead of tf-idf scoring (ablation) | AR-Tree · Age prediction | accuracy (%): 80.20 -> 80.85 | Helped |
| use AR-Tree instead of Gumbel Tree-LSTM | sentence embedding · SST-2 | accuracy (%): 90.1 -> 90.4 | Helped |
| use AR-Tree instead of byte-mLSTM | sentence embedding · SST-2 | accuracy (%): 91.8 -> 90.4 | Hurt |
| use AR-Tree instead of BCN+Char+CoVe | sentence embedding · SST-5 | accuracy (%): 53.7 -> 52.7 | Hurt |
Closest papers by meaning
- A Simple but Tough-to-Beat Baseline for Sentence Embeddings (ICLR 2017 · similarity 0.88)
- Learning to Compose Words into Sentences with Reinforcement Learning (ICLR 2017 · similarity 0.87)
- Scalable Attentive Sentence Pair Modeling via Distilled Sentence Embedding (AAAI 2020 · similarity 0.87)
- Fine-grained Analysis of Sentence Embeddings Using Auxiliary Prediction Tasks (ICLR 2017 · similarity 0.86)
- Augmenting Transformers with Recursively Composed Multi-grained Representations (ICLR 2024 · similarity 0.86)
- Narrowing the Gap between Supervised and Unsupervised Sentence Representation Learning with Large Language Model (AAAI 2024 · similarity 0.85)
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.