Research map / Image generation / Editing and style transfer
Controllable Unsupervised Text Attribute Transfer via Editing Entangled Latent Representation
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 entangled latent editing with FGIM rather than the strongest listed baseline | Text attribute transfer · Yelp | Acc: 95.2% -> 95.4% | Helped |
| Use entangled latent editing with FGIM rather than the strongest listed baseline | Text attribute transfer · Yelp | BLEU: 22.8 -> 24.6 | Helped |
| Use entangled latent editing with FGIM rather than the strongest listed baseline | Text attribute transfer · Amazon | PPL: 48.3 -> 47.4 | Helped |
| Use entangled latent editing with FGIM rather than the strongest listed baseline | Text attribute transfer · Captions | Acc: 95.5% -> 92.3% | Hurt |
| Use entangled latent editing with FGIM rather than the strongest listed baseline | Text attribute transfer · Captions | PPL: 31.2 -> 23.7 | Helped |
| Increase the FGIM modification weight | FGIM text transfer · Yelp | attribute accuracy | Helped |
| Increase the FGIM modification weight | FGIM text transfer · Yelp | BLEU | Mixed |
| Increase the FGIM modification weight | FGIM text transfer · Yelp | PPL | No clear effect |
| Use entangled latent editing with a multi-aspect classifier | Beer review attribute transfer · Beer-Advocate | Acc: Appearance 90.2%; Aroma 89.3%; Palate 91.2%; Taste 88.2%; Overall 87.3% | Helped |
Closest papers by meaning
- Multiple-Attribute Text Rewriting (ICLR 2019 · similarity 0.90)
- Revision in Continuous Space: Unsupervised Text Style Transfer without Adversarial Learning (AAAI 2020 · similarity 0.86)
- Graph-based Unsupervised Disentangled Representation Learning via Multimodal Large Language Models (NeurIPS 2024 · similarity 0.83)
- Continuous, Subject-Specific Attribute Control in T2I Models by Identifying Semantic Directions (CVPR 2025 · similarity 0.83)
- Harnessing the Universal Geometry of Embeddings (NeurIPS 2025 · similarity 0.82)
- Retrieval-based Disentangled Representation Learning with Natural Language Supervision (ICLR 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.