Research map / Image generation / GANs
CoT: Cooperative Training for Generative Modeling of Discrete Data
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 CoT instead of MLE | Synthetic sequence generation · Synthetic Turing test | NLL_oracle: 9.08 -> 8.19 | Helped |
| Use CoT instead of MLE | Synthetic sequence generation · Synthetic Turing test | NLL_test (final): 8.97 -> 8.03 | Helped |
| Use CoT instead of SeqGAN | Synthetic sequence generation · Synthetic Turing test | time/epoch: 817.64 ± 5.41s -> 53.94 ± 1.01s | Helped |
| Use CoT-strong* instead of LeakGAN* | Text generation · EMNLP 2017 WMT News test | BLEU4: 0.437 -> 0.510 | Helped |
| Use CoT-basic instead of MLE | Text generation · EMNLP 2017 WMT News | eWMD_test: 1.015 -> 0.766 | Helped |
| Drop the REINFORCE-like gradient term (set γ=0.0) (ablation) | CoT generator · Synthetic sequence modeling | gradient variance | Helped |
| Train the mediator on balanced real and generated samples rather than fitting a predictive model only to real data (ablation) | CoT mediator · Sequential discrete data | generator guidance stability | Helped |
Closest papers by meaning
- CoopInit: Initializing Generative Adversarial Networks via Cooperative Learning (AAAI 2023 · similarity 0.90)
- COT-GAN: Generating Sequential Data via Causal Optimal Transport (NeurIPS 2020 · similarity 0.89)
- Generative Cooperative Networks for Natural Language Generation (ICML 2022 · similarity 0.88)
- MCGAN: Enhancing GAN Training with Regression-Based Generator Loss (AAAI 2025 · similarity 0.87)
- Direct Discriminative Optimization: Your Likelihood-Based Visual Generative Model is Secretly a GAN Discriminator (ICML 2025 · similarity 0.87)
- Learning Energy-Based Models by Cooperative Diffusion Recovery Likelihood (ICLR 2024 · similarity 0.87)
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.