Research map / Label-efficient and robust learning / Recognition and long-tail
Style Transfer from Non-Parallel Text by Cross-Alignment
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 |
|---|---|---|---|
| add cross-alignment of transferred sentence populations rather than aligning only latent representations (ablation) | Aligned auto-encoder · Word substitution decipherment, 100% substitution | BLEU: 7.2 -> 57.4 | Helped |
| add cross-alignment of transferred sentence populations rather than aligning only latent representations (ablation) | Aligned auto-encoder · Word substitution decipherment, 20%, 40%, 60%, and 80% substitution | BLEU: 81.0 -> 83.8; 68.9 -> 79.1; 50.7 -> 74.7; 45.6 -> 66.1 | Helped |
| add cross-alignment of transferred sentence populations rather than aligning only latent representations (ablation) | Aligned auto-encoder · Word order recovery | BLEU: 5.2 -> 26.1 | Helped |
| replace cross-alignment with a variational auto-encoder (ablation) | Style transfer model · Yelp sentiment modification | sentiment accuracy: 23.2 -> 78.4 | Hurt |
| replace cross-alignment with a variational auto-encoder (ablation) | Style transfer model · Word substitution decipherment, 100% substitution | BLEU: 0.9 -> 57.4 | Hurt |
| replace cross-alignment with a variational auto-encoder (ablation) | Style transfer model · Word order recovery | BLEU: 5.3 -> 26.1 | Hurt |
| use cross-aligned auto-encoder instead of the control-gen baseline | Yelp sentiment modification · Yelp sentiment modification | overall transfer: 41.0 -> 41.5 | Helped |
| use cross-aligned auto-encoder instead of the control-gen baseline | Yelp sentiment modification · Yelp sentiment modification | sentiment accuracy: 83.5 -> 78.4 | Hurt |
| use cross-aligned auto-encoder instead of the control-gen baseline | Yelp sentiment modification · Yelp sentiment modification | fluency: 3.2 -> 2.8 | Hurt |
Closest papers by meaning
- Non-Parallel Text Style Transfer with Self-Parallel Supervision (ICLR 2022 · similarity 0.90)
- ParaGuide: Guided Diffusion Paraphrasers for Plug-and-Play Textual Style Transfer (AAAI 2024 · similarity 0.85)
- StyleStudio: Text-Driven Style Transfer with Selective Control of Style Elements (CVPR 2025 · similarity 0.84)
- Style Aligned Image Generation via Shared Attention (CVPR 2024 · similarity 0.84)
- Adaptive Prompt Routing for Arbitrary Text Style Transfer with Pre-trained Language Models (AAAI 2024 · similarity 0.83)
- CSGO: Content-Style Composition in Text-to-Image Generation (NeurIPS 2025 · similarity 0.83)
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.