mlsanityData to 14 Oct 2026

Learning Latent Permutations with Gumbel-Sinkhorn Networks

arXiv 1802.08665 ↗First posted 23 Feb 2018stat.ML cs.LGcited by 175

Not on the map. Why?

Abstract (arXiv)

Permutations and matchings are core building blocks in a variety of latent variable models, as they allow us to align, canonicalize, and sort data. Learning in such models is difficult, however, because exact marginalization over these combinatorial objects is intractable. In response, this paper introduces a collection of new methods for end-to-end learning in such models that approximate discrete maximum-weight matching using the continuous Sinkhorn operator. Sinkhorn iteration is attractive because it functions as a simple, easy-to-implement analog of the softmax operator. With this, we can define the Gumbel-Sinkhorn method, an extension of the Gumbel-Softmax method (Jang et al. 2016, Maddison2016 et al. 2016) to distributions over latent matchings. We demonstrate the effectiveness of our method by outperforming competitive baselines on a range of qualitatively different tasks: sorting numbers, solving jigsaw puzzles, and identifying neural signals in worms.

Key facts. Cited by 175 papers in the map (data through 14 Oct 2026). First posted on arXiv 23 Feb 2018.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 28; 2022: 29; 2023: 32; 2024: 20; 2025: 30; 2026: 36.

175 papers.

Sort:

By area:

  1. 2022Git Re-Basin: Merging Models modulo Permutation Symmetriescited by 471
  2. 2025The Diffusion Dualitycited by 135
  3. 2022Learning Vector-Quantized Item Representation for Transferable Sequential Recommenderscited by 112
  4. 2022Equivariance with Learned Canonicalization Functionscited by 109
  5. 2022Tutorial on amortized optimizationcited by 83
  6. 2021A Review of the Gumbel-max Trick and its Extensions for Discrete Stochasticity in Machine Learningcited by 82
  7. 2021Self-supervised learning methods and applications in medical imaging analysis: A surveycited by 77
  8. 2021BCD Nets: Scalable Variational Approaches for Bayesian Causal Discoverycited by 68
  9. 2022On the Symmetries of Deep Learning Models and their Internal Representationscited by 64
  10. 2022Re-basin via implicit Sinkhorn differentiationcited by 62
  11. 2022A Generalist Neural Algorithmic Learnercited by 59
  12. 2022Interventions, Where and How? Experimental Design for Causal Models at Scalecited by 49
38 more
  1. 2022Differentiable DAG Samplingcited by 47
  2. 2024Adaptive Data Optimization: Dynamic Sample Selection with Scaling Lawscited by 47
  3. 2024Discrete Probabilistic Inference as Control in Multi-path Environmentscited by 44
  4. 2021Human-like Controllable Image Captioning with Verb-specific Semantic Rolescited by 41
  5. 2022Unsupervised Deep Multi-Shape Matchingcited by 39
  6. 2023Structure Learning with Continuous Optimization: A Sober Look and Beyondcited by 36
  7. 2022EdiT5: Semi-Autoregressive Text-Editing with T5 Warm-Startcited by 35
  8. 2023Equivariant Deep Weight Space Alignmentcited by 35
  9. 2023BayesDAG: Gradient-Based Posterior Inference for Causal Discoverycited by 33
  10. 2021Differentiable Sorting Networks for Scalable Sorting and Ranking Supervisioncited by 31
  11. 2022A Unified Framework for Implicit Sinkhorn Differentiationcited by 28
  12. 2023Scalable Optimal Transport Methods in Machine Learning: A Contemporary Surveycited by 27
  13. 2021From Alignment to Assignment: Frustratingly Simple Unsupervised Entity Alignmentcited by 26
  14. 2022Towards Compositional Adversarial Robustness: Generalizing Adversarial Training to Composite Semantic Perturbationscited by 26
  15. 2023Learning Probabilistic Symmetrization for Architecture Agnostic Equivariancecited by 26
  16. 2022Monotonic Differentiable Sorting Networkscited by 25
  17. 2023Hierarchical Multi-Marginal Optimal Transport for Network Alignmentcited by 24
  18. 2023Self-Supervised Learning for Multimodal Non-Rigid 3D Shape Matchingcited by 23
  19. 2024A Meta-Learning Approach to Bayesian Causal Discoverycited by 22
  20. 2021Learning with Algorithmic Supervision via Continuous Relaxationscited by 21
  21. 2023Discovering Spatio-Temporal Rationales for Video Question Answeringcited by 21
  22. 2023Unsupervised Entity Alignment for Temporal Knowledge Graphscited by 19
  23. 2021Rethinking Search: Making Domain Experts out of Dilettantescited by 18
  24. 2022Equivariance Discovery by Learned Parameter-Sharingcited by 18
  25. 2021Structured Reordering for Modeling Latent Alignments in Sequence Transductioncited by 17
  26. 2021Discovering Non-monotonic Autoregressive Orderings with Variational Inferencecited by 15
  27. 2021Self-Supervised Learning of Audio Representations from Permutations with Differentiable Rankingcited by 14
  28. 2021Learning Revenue-Maximizing Auctions With Differentiable Matchingcited by 14
  29. 2026Learning to Evict from Key-Value Cachecited by 14
  30. 2021Multilingual Knowledge Graph Completion with Joint Relation and Entity Alignmentcited by 12
  31. 2022Neural Topological Ordering for Computation Graphscited by 12
  32. 2023Non-autoregressive Streaming Transformer for Simultaneous Translationcited by 12
  33. 2023Robot Fleet Learning via Policy Mergingcited by 11
  34. 2024Graph Edit Distance with General Costs Using Neural Set Divergencecited by 11
  35. 2025From Task-Specific Models to Unified Systems: A Review of Model Merging Approachescited by 11
  36. 2026FlashSinkhorn: IO-Aware Entropic Optimal Transport on GPUcited by 11
  37. 2021Stochastic Iterative Graph Matchingcited by 10
  38. 2021Neural Fixed-Point Acceleration for Convex Optimizationcited by 10

50 most cited of 175 All 175 (JSON)