References

This page collects the primary literature behind the methods implemented in MedLatents. For where each method lives in the codebase, see Implemented Methods.

Generative model families

  • D3PM – Austin, J., Johnson, D. D., Ho, J., Tarlow, D., & van den Berg, R. (2021). Structured Denoising Diffusion Models in Discrete State-Spaces. NeurIPS.

  • MaskGIT – Chang, H., Zhang, H., Jiang, L., Liu, C., & Freeman, W. T. (2022). MaskGIT: Masked Generative Image Transformer. CVPR.

  • SEDD – Lou, A., Meng, C., & Ermon, S. (2024). Discrete Diffusion Modeling by Estimating the Ratios of the Data Distribution. ICML.

  • MDLM – Sahoo, S. S., et al. (2024). Simple and Effective Masked Diffusion Language Models. NeurIPS.

  • Flow Matching – Lipman, Y., Chen, R. T. Q., Ben-Hamu, H., Nickel, M., & Le, M. (2023). Flow Matching for Generative Modeling. ICLR.

  • Discrete Flow Matching – Gat, I., et al. (2024). Discrete Flow Matching. NeurIPS.

  • Rectified Flow – Liu, X., Gong, C., & Liu, Q. (2023). Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow. ICLR.

  • Bayesian Flow Networks – Graves, A., Srivastava, R. K., Atkinson, T., & Gomez, F. (2023). Bayesian Flow Networks. arXiv:2308.07037.

  • Diffusion Transformer (DiT) – Peebles, W., & Xie, S. (2023). Scalable Diffusion Models with Transformers. ICCV.

Sampling and inference

  • Zero-Terminal-SNR – Lin, S., Liu, B., Li, J., & Yang, X. (2024). Common Diffusion Noise Schedules and Sample Steps are Flawed. WACV.

  • DPM-Solver – Lu, C., Zhou, Y., Bao, F., Chen, J., Li, C., & Zhu, J. (2022). DPM-Solver: A Fast ODE Solver for Diffusion Probabilistic Model Sampling in Around 10 Steps. NeurIPS.

  • DDIM – Song, J., Meng, C., & Ermon, S. (2021). Denoising Diffusion Implicit Models. ICLR.

  • Classifier-Free Guidance – Ho, J., & Salimans, T. (2022). Classifier-Free Diffusion Guidance. arXiv:2207.12598.

  • CFG-Zero* – Fan, W., et al. (2025). CFG-Zero: Improved Classifier-Free Guidance for Flow-Matching Models.* arXiv:2503.18886.

  • ReinMax – Liu, L., et al. (2023). Bridging Discrete and Backpropagation: Straight-Through and Beyond. NeurIPS.

  • RePaint – Lugmayr, A., et al. (2022). RePaint: Inpainting using Denoising Diffusion Probabilistic Models. CVPR.

  • Speculative Decoding – Leviathan, Y., Kalman, M., & Matias, Y. (2023). Fast Inference from Transformers via Speculative Decoding. ICML.

  • Medusa – Cai, T., et al. (2024). Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads. ICML.

Training and representation alignment

  • REPA – Yu, S., et al. (2025). Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Think. ICLR.

  • REPA-E – Leng, X., et al. (2025). REPA-E: Unlocking VAE for End-to-End Tuning with Latent Diffusion Transformers. arXiv:2504.10483.

Preference optimization

  • DPO – Rafailov, R., et al. (2023). Direct Preference Optimization: Your Language Model is Secretly a Reward Model. NeurIPS.

  • IPO – Azar, M. G., et al. (2023). A General Theoretical Paradigm to Understand Learning from Human Preferences. arXiv:2310.12036.

  • KTO – Ethayarajh, K., et al. (2024). KTO: Model Alignment as Prospect Theoretic Optimization. ICML.

  • Diffusion-DPO – Wallace, B., et al. (2023). Diffusion Model Alignment Using Direct Preference Optimization. arXiv:2311.12908.

  • SPO – Liang, Z., et al. (2024). Step-by-Step Preference Optimization. (CVPR 2025).

Reinforcement learning

  • DDPO – Black, K., Janner, M., Du, Y., Kostrikov, I., & Levine, S. (2023). Training Diffusion Models with Reinforcement Learning. arXiv:2305.13301.

  • GRPO – Shao, Z., et al. (2024). DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models. arXiv:2402.03300.

Distillation

  • Consistency Models – Song, Y., Dhariwal, P., Chen, M., & Sutskever, I. (2023). Consistency Models. ICML.

  • Reflow – Liu, X., Gong, C., & Liu, Q. (2022). Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow. arXiv:2209.03003.

Self-play

  • SPIN – Chen, Z., et al. (2024). Self-Play Fine-Tuning Converts Weak Language Models to Strong Language Models. ICML.

  • RFT – Yuan, Z., et al. (2023). Scaling Relationship on Learning Mathematical Reasoning with Large Language Models. arXiv:2308.01825.

Citing MedLatents

If you use MedLatents in your research, please cite the project. See the Getting Started page and the CITATION.cff file at the repository root for the current citation.