Research map / Image generation / GANs
Meta-CoTGAN: A Meta Cooperative Training Paradigm for Improving Adversarial Text Generation
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 Meta-CoTGAN instead of CoT | Text generation on synthetic oracle · Synthetic oracle, sequence length 20 | NLL_oracle: 8.19 -> 7.69 | Helped |
| Use Meta-CoTGAN instead of CoT | Text generation on synthetic oracle · Synthetic oracle, sequence length 20 | NLL_gen: 7.54 -> 6.86 | Helped |
| Use Meta-CoTGAN instead of RelGAN | Text generation at temperature 100 · COCO Image Captions | BLEU-5: 0.331 ± 0.044 -> 0.363 ± 0.009 | Helped |
| Use Meta-CoTGAN instead of RelGAN | Text generation at temperature 100 · COCO Image Captions | NLL_gen: 0.756 ± 0.054 -> 0.578 ± 0.036 | Helped |
| Use Meta-CoTGAN instead of RelGAN | Text generation at temperature 100 · EMNLP2017 WMT News | BLEU-5: 0.319 ± 0.018 -> 0.358 ± 0.015 | Helped |
| Use Meta-CoTGAN instead of RelGAN | Text generation at temperature 100 · EMNLP2017 WMT News | NLL_gen: 2.482 ± 0.031 -> 2.299 ± 0.011 | Helped |
| Add online cooperative updates to the language model (ablation) | Meta-CoTGAN · COCO Image Captions, temperature 100 | BLEU-5: 0.315 ± 0.022 -> 0.363 ± 0.009 | Helped |
| Add online cooperative updates to the language model (ablation) | Meta-CoTGAN · COCO Image Captions, temperature 100 | NLL_gen: 0.580 ± 0.031 -> 0.578 ± 0.036 | No clear effect |
| Add meta optimization after the adversarial update (ablation) | Meta-CoTGAN · COCO Image Captions, temperature 100 | BLEU-5: 0.338 ± 0.014 -> 0.363 ± 0.009 | Helped |
| Add meta optimization after the adversarial update (ablation) | Meta-CoTGAN · COCO Image Captions, temperature 100 | NLL_gen: 0.587 ± 0.019 -> 0.578 ± 0.036 | No clear effect |
Closest papers by meaning
- Generative Cooperative Networks for Natural Language Generation (ICML 2022 · similarity 0.90)
- Branch-GAN: Improving Text Generation with (not so) Large Language Models (ICLR 2024 · similarity 0.90)
- Self-Adversarial Learning with Comparative Discrimination for Text Generation (ICLR 2020 · similarity 0.89)
- Long Text Generation via Adversarial Training with Leaked Information (AAAI 2018 · similarity 0.89)
- TextGAIL: Generative Adversarial Imitation Learning for Text Generation (AAAI 2021 · similarity 0.88)
- MaskGAN: Better Text Generation via Filling in the _______ (ICLR 2018 · similarity 0.88)
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.