betaFeedback

WorksComputer Science

Learning from Context: Leave-One-Out Prediction in Jump Diffusions

Abstract

Continuous diffusion models generate discrete data through stochastic dynamics in an embedding space. We consider jump diffusions that additionally reset coordinates to embeddings of categorical values. A coordinate's own corrupted state can make prediction easy, potentially reducing the pressure to learn dependencies between coordinates. We study leave-one-out (LOO) prediction, in which each coordinate is predicted from the corrupted states of the other coordinates. We show that, under stated assumptions, jump diffusions with complete coordinate resets and state-independent clocks preserve the prescribed marginals when reset destinations use exact LOO probabilities and the drift uses the full endpoint posterior. We further show that this drift can be constructed from LOO probabilities by analytic likelihood completion without additional network evaluations. We evaluate unrestricted, LOO, and completed predictions on Text8 and LM1B text generation and ZINC250k molecular graph generation. LOO-based variants improve key generation metrics over matched unrestricted controls on all three datasets. Controlled interventions assess reliance on local inputs and context, and we analyze the effect of analytic likelihood completion.

Introduction

Generating text and molecular graphs requires learning dependencies between discrete coordinates. Continuous (jump) diffusion models approach this task by corrupting token embeddings and learning to predict their endpoints [23, 10, 16, 20]. Yet a coordinate’s own noisy embedding can reveal its target without requiring the model to learn the surrounding dependencies. A masked position in masked diffusion or flow matching provides no such local token evidence [34, 13]. For continuous models, increasing noise to remove this shortcut can also destroy useful context [31]. We ask whether excluding a coordinate from its own prediction, so called leave-one-out (LOO) prediction, encourages useful contextual learning, and how to do so while preserving the intended probability path.

We study LOO prediction in embedding-space jump diffusions, where continuous motion is combined with jumps to token embeddings. Let Y=(Y1,…,YL)Y = (Y_1,\ldots,Y_L) be a token sequence, Xt=(X1,t,…,XL,t)X_t = (X_{1,t},\ldots,X_{L,t}) its corrupted embeddings in continuous space at time tt, and denote by X−i,tX_{-i,t} all coordinates except ii. For a token aa, the full and LOO posteriors are

pi,t(a∣x)=P(Yi=a∣Xt=x),p_{i,t}(a \mid x) = \mathbb{P}(Y_i = a \mid X_t = x),
qi,t(a∣x−i)=P(Yi=a∣X−i,t=x−i).q_{i,t}(a \mid x_{-i}) = \mathbb{P}(Y_i = a \mid X_{-i,t} = x_{-i}).

Predicting qi,tq_{i,t} requires estimating YiY_i without access to Xi,tX_{i,t}. Our four-token parity experiment illustrates why this can matter: low full-state prediction loss can precede learning the dependence needed for joint generation (Table 1).

Excluding local inputs is established in blind-spot denoising [1, 22] and discrete diffusion [40, 2] (Gouvritch et al. 2026). For a broader overview of related work, see Appendix § 7. The gap we address is how to use these qi,t(a∣x−i)q_{i,t}(a \mid x_{-i}) predictions in processes with both a continuous drift and coordinate resets, i.e. jumps to the token embeddings. In this case the input of the network provides noisy contextual information before discrete decisions. Continuous motion resumes after each reset, and later resets can replace an earlier, possibly wrong, hypothesis. LOO makes the last point particularly meaningful: the learned reset distribution qθ,i,t(a∣x−i)q_{\theta,i,t}(a \mid x_{-i}) depends only on the surrounding state x−ix_{-i}, so an incorrect local state xix_i at time tt cannot directly influence this reset distribution.

The distinction between pi,tp_{i,t} and qi,tq_{i,t} is essential. Posterior averaging of conditional generators uses the full posterior pi,t(a∣x)p_{i,t}(a \mid x) [17], whereas LOO training learns qi,t(a∣x−i)q_{i,t}(a \mid x_{-i}) without using xix_i. We show that substituting the latter into the drift can change the prescribed marginal path (Example C.2). For jump diffusion, however, exact LOO jump probabilities qi,t(a∣x−i)q_{i,t}(a \mid x_{-i}) preserve the correct probability curve when the drift retains the full posterior, under the assumptions of Proposition 3.2. Analytic likelihood completion combines LOO predictions with the local corruption likelihood to recover the full posterior. Thus one network can supply both prediction roles without additional network evaluations.

We compare unrestricted, raw LOO, and completed predictors on Text8, LM1B, and ZINC250k. LOO improves key generation metrics over matched unrestricted controls, while completion’s effect depends on the path and dataset. Local-state and subject–verb interventions test contextual information use, see § 4.

Our contributions are:

  • A marginal-preservation result for complete coordinate resets using LOO destinations and full-posterior drift, with a counterexample for substituting LOO probabilities into the drift.

  • A repeated-reset Brownian bridge whose marginals converge to the data law, using one LOO network and analytic likelihood completion for the two prediction roles.

  • Matched-budget experiments on Text8, LM1B, and ZINC250k showing when LOO improves over unrestricted prediction and when raw LOO drift outperforms completion despite lacking a marginal-preservation guarantee.

  • We demonstrate empirically that LOO drift without analytic completion can help recover from wrong predictions at earlier time steps, offering a possible explanation for its better performance in some experiments despite lacking a marginal-preservation guarantee. Controlled LM1B interventions show that a token’s own noisy input can dominate unmasked predictions and final token choices, even when it conflicts with context. On LM1B, LOO improves subject tracking and the usefulness of older word order in generated text.

Conditional paths and leave-one-out prediction

Let V\mathcal{V} be a finite vocabulary and let pdata∈P(VL)p_{\mathrm{data}} \in\mathcal{P}(\mathcal{V}^{L}) be a data distribution on sequences of length LL. Choose a random variable Y=(Y1,…,YL)∼pdataY = (Y_1,\ldots,Y_L) \sim p_{\mathrm{data}}, and consider an embedding e:V→Rde:\mathcal{V} \to\mathbb{R}^{d}. For recovery of token sequences, we assume that ee is injective. Write E(Y)=(e(Y1),…,e(YL))E(Y) = (e(Y_1),\ldots,e(Y_L)) and for x=(x1,…,xL)∈RdLx = (x_1,\ldots,x_L) \in\mathbb{R}^{dL} let xi←zx^{i \gets z} denote xx with position ii replaced by zz, i.e. xi←z=(x1,…,xi−1,z,…,xL)x^{i \gets z} = (x_1,\ldots,x_{i-1},z,\ldots,x_L). We write x−i=(x1,…,xi−1,xi+1,…,xL)x_{-i} = (x_1,\ldots,x_{i-1},x_{i+1},\ldots,x_L) for the sequence with position ii omitted, and use X−i,tX_{-i,t} analogously. Time runs from the source at t=0t=0 toward data at t=1t=1. For each process below, we write pt=L(Xt)p_t = \mathcal{L}(X_t) for its state law and pi,t(a∣x)=P(Yi=a∣Xt=x)p_{i,t}(a \mid x) = \mathbb{P}(Y_i = a \mid X_t = x) for the endpoint posterior at position ii.

Discrete flow matching

Discrete flow matching (DFM) uses a probability path on token sequences and learns a jump rate kernel that generates its marginals [13]. Appendix B.1 summarizes the terminology. For the masked path, let V\mathcal{V} be a mask and we consider the following discrete process Xty∈(V∪{m})LX_t^y \in(\mathcal{V} \cup\{m\})^L starting from X0y=(m,…,m)X_0^y = (m,\ldots,m). Let κ:[0,1]→[0,1]\kappa:[0,1] \to[0,1] be continuous and nondecreasing, continuously differentiable on [0,1)[0,1), with κ(0)=0\kappa(0)=0, κ(1)=1\kappa(1)=1, and κ(t)<1\kappa(t)<1 for t<1t<1. For a fixed endpoint yy, take independent reveal times T1,…,TLT_1,\ldots,T_L with P(Ti≤t)=κ(t)\mathbb{P}(T_i \le t)=\kappa(t) and define the conditional process Xty=(X1,ty,…,XL,ty)X_t^y=(X_{1,t}^y,\ldots,X_{L,t}^y) by setting Xi,ty=mX_{i,t}^y=m for t<Tit<T_i and Xi,ty=yiX_{i,t}^y=y_i for t≥Tit\ge T_i. Each position is therefore revealed independently with probability κ(t)\kappa(t) at time tt, so

P(Xty=x)=∏i=1L[(1−κ(t))1{xi=m}+κ(t)1{xi=yi}].(1)\mathbb{P}(X_t^y=x)=\prod_{i=1}^{L}\left[(1-\kappa(t))\mathbf{1}\{x_i=m\}+\kappa(t)\mathbf{1}\{x_i=y_i\}\right]. \tag*{(1)}

Since κ(1)=1\kappa(1)=1, X1y=yX^{y}_{1}=y almost surely. Hence the mixture pt=L(XtY)p_t=\mathcal{L}(X^{Y}_{t}) satisfies p1=pdatap_1=p_{\mathrm{data}}. This conditional process reveals yiy_i at rate κ′(t)/(1−κ(t))\kappa'(t)/(1-\kappa(t)) while position ii is masked and then leaves it fixed. Gat et al. [13] obtain the jump rate kernel for the marginal process XtX_t with law ptp_t through the per-position rates

ri,t(xi→a∣x)=1{xi=m}κ′(t)1−κ(t)pi,t(a∣x),a∈V.(2)r_{i,t}(x_i\to a\mid x)=\mathbf{1}\{x_i=\mathrm{m}\}\frac{\kappa'(t)}{1-\kappa(t)}p_{i,t}(a\mid x),\qquad a\in\mathcal{V}. \tag*{(2)}

Thus generation requires pi,t(a∣x)p_{i,t}(a\mid x), a categorical endpoint posterior at each position, conditioned on the entire current sequence. These posteriors can be learned by cross-entropy on samples of (1) (Gat et al., [13]; Sahoo et al., [34]). At a masked position, the local mask provides no evidence about its target token.

Continuous embeddings and endpoint posteriors

References

  1. [1]Joshua Batson and Loic Royer. Noise2Self: Blind denoising by self-supervision. In Proceedings of the 36th International Conference on Machine Learning, volume 97 of Proceedings of Machine Learning Research, pp. 524–533, 2019. URL https://proceedings.mlr.press/v97/batson19a.html.arxiv.org/abs/1901.11365
  2. [2]Rodrigo Casado Noguerales, Bernhard Schölkopf, Thomas Hofmann, and Aran Raoufi. What does a discrete diffusion model learn? arXiv preprint arXiv:2607.05381, 2026. URL https://arxiv.org/abs/2607.05381.
  3. [3]Huiwen Chang, Han Zhang, Lu Jiang, Ce Liu, and William T. Freeman. MaskGIT: Masked generative image transformer. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 11315–11325, 2022. URL https://openaccess.thecvf.com/content/CVPR2022/html/Chang_MaskGIT_Masked_Generative_Image_Transformer_CVPR_2022_paper.html.arxiv.org/abs/2202.04200
  4. [4]Ciprian Chelba, Tomas Mikolov, Mike Schuster, Qi Ge, Thorsten Brants, Philipp Koehn, and Tony Robinson. One billion word benchmark for measuring progress in statistical language modeling. In Interspeech 2014, pp. 2635–2639, 2014. doi: 10.21437/Interspeech.2014-564. URL https://www.isca-archive.org/interspeech_2014/chelba14_interspeech.html.DOI
  5. [5]Jannis Chemseddine, Gregor Kornhardt, and Gabriele Steidl. Spherical flows for sampling categorical data. arXiv preprint arXiv:2605.05629, 2026. URL https://arxiv.org/abs/2605.05629.
  6. [6]Yuxin Chen, Chumeng Liang, Hangke Sui, Ruihan Guo, Chaoran Cheng, Jiaxuan You, and Ge Liu. LangFlow: Continuous diffusion rivals discrete in language modeling. arXiv preprint arXiv:2604.11748, 2026. URL https://arxiv.org/abs/2604.11748.
  7. [7]Giovanni Conforti and Christian Léonard. Time reversal of Markov processes with jumps under a finite entropy condition. Stochastic Processes and their Applications, 144:85–124, 2022. doi: 10.1016/j.spa.2021.10.002. URL https://doi.org/10.1016/j.spa.2021.10.002.
  8. [8]Fabrizio Costa and Kurt De Grave. Fast neighborhood subgraph pairwise distance kernel. In Proceedings of the 27th International Conference on Machine Learning, 2010. URL https://icml.cc/Conferences/2010/papers/347.pdf.
  9. [9]Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. BERT: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pp. 4171–4186. Association for Computational Linguistics, 2019. doi: 10.18653/v1/N19-1423. URL https://aclanthology.org/N19-1423/.DOI
  10. [10]Sander Dieleman, Laurent Sartran, Arman Roshanan, Nikolay Savinov, Yaroslav Ganin, Pierre H. Richemon d, Arnaud Doucet, Robin Strudel, Chris Dyer, Conor Durkan, Curtis Hawthorne, Rémi Leblond, Will Grathwohl, and Jonas Adler. Continuous diffusion for categorical data. arXiv preprint arXiv:2211.15089, 2022. URL https://arxiv.org/abs/2211.15089.
  11. [11]Daniil Dmitriev, Zhihan Huang, and Yuting Wei. Provably adaptive sampling with uniform and remasking discrete diffusion models. arXiv preprint arXiv:2608.23554, 2026. URL https://arxiv.org/abs/2608.23554.
  12. [12]Martin R. Evans, Satya N. Majumdar, and Grégory Schehr. Stochastic resetting and applications. Journal of Physics A: Mathematical and Theoretical, 53(19):193001, 2020. doi: 10.1088/1751-8121/ab7cfe. URL https://arxiv.org/abs/1910.07993.
  13. [13]Itai Gat, Tal Remez, Neta Shaul, Felix Kreuk, Ricky T. Q. Chen, Gabriel Synnaeve, Yossi Adi, and Yaron Lipman. Discrete flow matching. In Advances in Neural Information Processing Systems, volume 37, 2024. doi: 10.52202/079017-4239. URL https://proceedings.neurips.cc/paper_files/paper/2024/hash/f0d629a734b56a642701bba7bc8bb3ed-Abstract-Conference.html.
  14. [14]Marjan Ghazvininejad, Omer Levy, Yinhan Liu, and Luke Zettlemoyer. Mask-Predict: Parallel decoding of conditional masked language models. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pp. 6112–6121. Association for Computational Linguistics, 2019. doi: 10.18653/v1/D19-1633. URL https://aclanthology.org/D19-1633/.DOI
  15. [15]Samson Gourevitch, Yazid Janati, Dario Shariatian, Umut Simsekli, Eric Moulines, Eric P. Xing, and Alain Durmus. Uniform diffusion models revisited: Leave-one-out denoiser and absorbing state reformulation. arXiv preprint arXiv:2605.22765, 2026. URL https://arxiv.org/abs/2605.22765.
  16. [16]Ishaan Gulrajani and Tatsunori B. Hashimoto. Likelihood-based diffusion language models. In Advances in Neural Information Processing Systems, volume 36, 2023. URL https://papers.nips.cc/paper_files/paper/2023/file/35b5c175e139bff5f22a5361270fce87-Paper-Conference.pdf.arxiv.org/abs/2305.18619
  17. [17]Peter Holderrieth, Marton Havasi, Jason Yim, Neta Shaul, Itai Gat, Tommi Jaakkola, Brian Karrer, Ricky T. Q. Chen, and Yaron Lipman. Generator matching: Generative modeling with arbitrary Markov processes. In International Conference on Learning Representations, 2025. URL https://arxiv.org/abs/2410.20587.
  18. [18]Ari Holtzman, Jan Buys, Li Du, Maxwell Forbes, and Yejin Choi. The curious case of neural text degeneration. In International Conference on Learning Representations, 2020. URL https://arxiv.org/abs/1904.09751.
  19. [19]Jaehyeong Jo, Seoul Lee, and Sung Ju Hwang. Score-based generative modeling of graphs via the system of stochastic differential equations. In Proceedings of the 39th International Conference on Machine Learning, volume 162 of Proceedings of Machine Learning Research, pp. 10362–10383, 2022. URL https://proceedings.mlr.press/v162/jo22a.html.arxiv.org/abs/2202.02514
  20. [20]Pascal Jutras-Dubé, Patrick Pynadath, Jeremy Lu, Yuan Gao, and Ruqi Zhang. Sticky jump diffusions: A unifying view of masked, continuous, and hybrid diffusion. arXiv preprint arXiv:2607.10951, 2026. URL https://arxiv.org/abs/2607.10951.
  21. [21]Hiroshi Kunita. Stochastic Flows and Jump-Diffusions, volume 92 of Probability Theory and Stochastic Modelling. Springer, Singapore, 2019. ISBN 978-981-13-3800-7. doi: 10.1007/978-981-13-3801-4. URL https://doi.org/10.1007/978-981-13-3801-4.
  22. [22]Samuli Laine, Tero Karras, Jaakko Lehtinen, and Timo Aila. High-quality self-supervised deep image denoising. In Advances in Neural Information Processing Systems, volume 32, 2019. URL https://papers.neurips.cc/paper/8920-high-quality-self-supervised-deep-image-denoising.pdf.arxiv.org/abs/1901.10277
  23. [23]Xiang Lisa Li, John Thickstun, Ishaan Gulrajani, Percy Liang, and Tatsunori B. Hashimoto. Diffusion-LM improves controllable text generation. In Advances in Neural Information Processing Systems, volume 35, pp. 4328–4343, 2022. URL https://arxiv.org/abs/2205.14217.
  24. [24]Xingchao Liu, Lemeng Wu, Mao Ye, and Qiang Liu. Let us build bridges: Understanding and extending diffusion generative models. arXiv preprint arXiv:2208.14699, 2022. URL https://arxiv.org/abs/2208.14699.
  25. [25]Aaron Lou, Chenlin Meng, and Stefano Ermon. Discrete diffusion modeling by estimating the ratios of the data distribution. In International Conference on Machine Learning, volume 235 of Proceedings of Machine Learning Research, pp. 32819–32848, 2024. URL https://proceedings.mlr.press/v235/lou24a.html.arxiv.org/abs/2310.16834
  26. [26]Matt Mahoney. About the test data, 2006. URL https://mattmahoney.net/dc/textdata.html. Text8 dataset; section “Relationship of Wikipedia Text to Clean Text”.
  27. [27]Mikhail Papkov, Pavel Chizhov, and Leopold Parts. SwinIA: Self-supervised blind-spot image denoising without convolutions. In Proceedings of the Winter Conference on Applications of Computer Vision, pp. 7071–7080, 2025. URL https://openaccess.thecvf.com/content/WACV2025/html/Papkov_SwinIA_Self-Supervised_Blind-Spot_Image_Denoising_without_Convolutions_WACV_2025_paper.html.arxiv.org/abs/2305.05651
  28. [28]William Peebles and Saining Xie. Scalable diffusion models with transformers. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 4195–4205, 2023. URL https://arxiv.org/abs/2212.09748.
  29. [29]Krishna Pillutla, Swabha Swayamdipta, Rowan Zellers, John Thickstun, Sean Welleck, Yejin Choi, and Zaid Harchaoui. MAUVE: Measuring the gap between neural text and human text using divergence frontiers. In Advances in Neural Information Processing Systems, volume 34, 2021. URL https://papers.nips.cc/paper_files/paper/2021/hash/260c2432a0eecc28ce03c10dadc078a4-Abstract.html.arxiv.org/abs/2102.01454
  30. [30]Kristina Preuer, Philipp Renz, Thomas Unterthiner, Sepp Hochreiter, and Günter Klambauer. Fréchet ChemNet distance: A metric for generative models for molecules in drug discovery. Journal of Chemical Information and Modeling, 58(9):1736–1741, 2018. doi: 10.1021/acs.jcim.8b00234. URL https://doi.org/10.1021/acs.jcim.8b00234.
  31. [31]Patrick Pynadath, Jiaxin Shi, and Ruqi Zhang. CANDI: Hybrid discrete-continuous diffusion models. In International Conference on Machine Learning, 2026a. URL https://arxiv.org/abs/2510.22510.
  32. [32]Patrick Pynadath, Jiaxin Shi, and Ruqi Zhang. Generative frontiers: Why evaluation matters for diffusion language models. arXiv preprint arXiv:2604.02718, 2026b. URL https://arxiv.org/abs/2604.02718.
  33. [33]Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. Language models are unsupervised multitask learners. Technical report, OpenAI, 2019. URL https://cdn.openai.com/better-language-models/language_models_are_unsupervised_multitask_learners.pdf.
  34. [34]Subham Sekhar Sahoo, Marianne Arriola, Yair Schiff, Aaron Gokaslan, Edgar Marroquin, Justin T. Chiu, Alexander Rush, and Volodymyr Kuleshov. Simple and effective masked diffusion language models. In Advances in Neural Information Processing Systems, volume 37, 2024. doi: 10.52202/079017-4135. URL https://proceedings.neurips.cc/paper_files/paper/2024/hash/eb0b13cc515724ab8015bc978fdde0ad-Abstract-Conference.html.
  35. [35]Claude E. Shannon. A mathematical theory of communication. The Bell System Technical Journal, 27(3):379–423, 1948. doi: 10.1002/j.1538-7305.1948.tb01338.x.DOI
  36. [36]Noam Shazeer. GLU variants improve transformer. arXiv preprint arXiv:2002.05202, 2020. URL https://arxiv.org/abs/2002.05202.
  37. [37]Jiaxin Shi, Kehang Han, Zhe Wang, Arnaud Doucet, and Michalis K. Titsias. Simplified and generalized masked diffusion for discrete data. In Advances in Neural Information Processing Systems, volume 37, 2024. URL https://papers.nips.cc/paper_files/paper/2024/hash/bad233b9849f01aead5e5c60c6ef70f-Abstract-Conference.html.arxiv.org/abs/2406.04329
  38. [38]Joongbo Shin, Yoonhyung Lee, Seunghyun Yoon, and Kyomin Jung. Fast and accurate deep bidirectional language representations for unsupervised learning. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pp. 823–835. Association for Computational Linguistics, 2020. doi: 10.18653/v1/2020.acl-main.76. URL https://aclanthology.org/2020.acl-main.76/.DOI
  39. [39]Teague Sterling and John J. Irwin. ZINC 15—ligand discovery for everyone. Journal of Chemical Information and Modeling, 55(11):2324–2337, 2015. doi: 10.1021/acs.jcim.5b00559. URL https://doi.org/10.1021/acs.jcim.5b00559.
  40. [40]Haoran Sun, Lijun Yu, Bo Dai, Dale Schuurmans, and Hanjun Dai. Score-based continuous-time discrete diffusion models. In International Conference on Learning Representations, 2023. URL https://arxiv.org/abs/2211.16750.
  41. [41]Yonghui Wu, Mike Schuster, Zhifeng Chen, Quoc V. Le, Mohammad Norouzi, Wolfgang Macherey, Maxim Krikun, Yuan Cao, Qin Gao, Klaus Macherey, Jeff Klingner, Apurva Shah, Melvin Johnson, Xiaobing Liu, Lukasz Kaiser, Stephan Gouws, Yoshikiyo Kato, Taku Kudo, Hideto Kazawa, Keith Stevens, George Kurian, Nishant Patil, Wei Wang, Cliff Young, Jason Smith, Jason Riesa, Alex Rudnick, Oriol Vinyals, Greg Corrado, Macduff Hughes, and Jeffrey Dean. Google’s neural machine translation system: Bridging the gap between human and machine translation. arXiv preprint arXiv:1609.08144, 2016. URL https://arxiv.org/abs/1609.08144.
  42. [42]Zhihan Yang, Wei Guo, Shuibai Zhang, Subham Sekhar Sahoo, Yongxin Chen, Arash Vahdat, Morteza Mardani, and John Thickstun. Continuous diffusion scales competitively with discrete diffusion for language. arXiv preprint arXiv:2605.18530, 2026. URL https://arxiv.org/abs/2605.18530.
  43. [43]Yixiu Zhao, Jiaxin Shi, Feng Chen, Shaul Druckmann, Lester Mackey, and Scott Linderman. Informed correctors for discrete diffusion models. In Advances in Neural Information Processing Systems, volume 38, 2025. doi: 10.52202/085713-4183. URL https://papers.nips.cc/paper_files/paper/2025/hash/b67d9d07faf071532c420ed5b140756b-Abstract-Conference.html.DOI
  44. [44]Huangjie Zheng, Shansan Gong, Ruixiang Zhang, Tianrong Chen, Jiatao Gu, Mingyuan Zhou, Navdeep Jaitly, and Yizhe Zhang. Continuously augmented discrete diffusion model for categorical generative modeling. In International Conference on Learning Representations, 2026. URL https://arxiv.org/abs/2510.01329.
  45. [45]Cai Zhou, Chenxiao Yang, Yi Hu, Chenyu Wang, Chubin Zhang, Muhan Zhang, Lester Mackey, Tommi Jaakkola, Stephen Bates, and Dinghuai Zhang. Coevolutionary continuous discrete diffusion: Make your diffusion language model a latent reasoner. In International Conference on Machine Learning, 2026. URL https://arxiv.org/abs/2510.03206.

Paper details

vlnc:W_pg39ez45
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC0-1.0
Usage

views · 7

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Computer Science

Not recorded for this paper

View statement
Looking for related papers…
Contents