Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

ALI-CFM learns multi-marginal flow-matching interpolants by adversarial distribution matching, yielding smooth trajectories that improve trajectory inference on spatial transcriptomics and cell-tracking data.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · deepseek-v4-flash

2026-08-04 12:56 UTC pith:2BVYGPTY

load-bearing objection Novel and useful method; the uniqueness theory is weaker than claimed and has a proof typo, but the empirical work stands on its own. the 3 major comments →

arxiv 2510.01159 v2 pith:2BVYGPTY submitted 2025-10-01 cs.LG

Multi-Marginal Flow Matching with Adversarially Learnt Interpolants

Oskar Kviman , Kirill Tamogashev , Nicola Branchini , V\'ictor Elvira , Jens Lagergren , Nikolay Malkin This is my paper
classification cs.LG
keywords flowmatchingdynamicsinterpolantsmethodmulti-marginalpointstime
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Given snapshots of a moving system at several times, we want to infer the continuous dynamics. Existing flow-matching methods often connect paired points from the first and last snapshot with straight lines or splines; when the shape of the distribution changes over time, these paths kink or leave the data. ALI-CFM instead trains a 'correction' network that bends each path so that the set of all paths at each intermediate time matches the observed distribution in a statistical sense. A discriminator network tries to distinguish real samples at time t from the interpolated samples, while the interpolant tries to fool it—a GAN. To avoid wild, non-unique curves, regularizers pull the paths toward linear or piecewise-linear references, or penalize curvature. Once the interpolant is learned, a standard conditional-flow-matching step trains a neural vector field along these paths, producing a dynamical system. On a synthetic 'knot' distribution, a cell-tracking video with 115 frames, and breast-cancer spatial transcriptomics sections, ALI-CFM produces smooth, accurate trajectories, while on single-cell RNA-seq it matches state-of-the-art methods.

Core claim

The central claim, stated in the abstract, is that ALI-CFM 'uses a GAN-inspired adversarial loss to fit neurally parametrised interpolant curves between source and target points such that the marginal distributions at intermediate time points are close to the observed distributions... These interpolants are subsequently marginalised by a flow matching algorithm, yielding a trained vector field for the underlying dynamics.' If correct, ALI-CFM solves multi-marginal trajectory inference with smooth interpolants that match intermediate distributions, improving accuracy on spatial transcriptomics and cell tracking.

Load-bearing premise

The method assumes that a single time-continuous family of interpolant maps G_phi(·,·,t) exists that pushes the fixed endpoint coupling π to each observed marginal q_t, and that the GAN-plus-regularizer optimization finds it. The uniqueness theorems (Thm 2.1, 2.2) guarantee uniqueness for a fixed t under an absolute-continuity condition, but do not establish the existence of a smooth family across t, nor that the finite-capacity network can represent it. The piecewise-linear regularizer (Eq. 11) uses a Markov-chained OT coupling that depends on the intermediate marginal, so different i give different reference maps; it is not shown that the shared network can satisfy all simultaneously.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper proposes ALI-CFM, a flow-matching method for multi-marginal trajectory inference. Instead of forcing interpolants to pass through observed intermediate samples, the method trains a neural interpolant G_phi(x0,x1,t) with a GAN-style adversarial loss so that its intermediate-time marginals match the observed distributions, together with a regularizer intended to make the interpolant unique. The trained interpolant is then used as the conditional probability path for a standard conditional flow matching objective. The authors provide uniqueness theorems for two regularizers (linear reference and piecewise-linear reference), and evaluate the method on a synthetic knot dataset, cell tracking microscopy data, scRNA-seq trajectory inference, and spatial transcriptomics tumor-coordinate inference, reporting gains on spatial transcriptomics and on par or better results on scRNA-seq compared to several baselines. Code is released.

Significance. The conceptual contribution is appealing: distributional matching of intermediate marginals, rather than pointwise matching, is a natural fit for noisy snapshot data and for settings with many time points. The paper also ships code and provides held-out evaluations (leave-one-out sections and held-out scRNA-seq marginals) that are genuine predictions, not in-sample fits. If the uniqueness theory held as stated, it would be a valuable guarantee. However, the theoretical claims are not established for the actual algorithm: the theorems address a constrained OT problem rather than the unconstrained penalized objective, and the proof of Theorem 2.2 contains a substantive mathematical gap. The empirical results are supportive but uneven, with cell tracking evaluated only visually. The method remains a useful empirical contribution, but the advertised theoretical grounding needs substantial repair.

major comments (3)
  1. [§2.3, Eq. (8) and Theorems 2.1, 2.2] Theorems 2.1 and 2.2 prove uniqueness for the constrained problem 'minimize the regularizer subject to G#π = q_t'. The ALI objective (Eq. 8) is instead the unconstrained sum L_GAN + λ L_reg. For any finite λ, a minimizer of Eq. (8) need not satisfy the marginal constraint, so the theorems do not characterize the optimizer of the actual objective. In the λ→∞ limit, one would need to show that the GAN loss can be driven to zero and that a global min-max optimum is attained by the parametric generator; no such result is supplied. Therefore the abstract's claim that the regularizers 'guarantee unique interpolants' is unsupported for the algorithm actually trained.
  2. [Appendix A, proof of Theorem 2.2] The proof of Theorem 2.2 is invalid as written. It states that the minimizer is G(x0,x1,t) = T∘ℓ(x_t | x0,x1,x_ti,t), but the right-hand side depends on x_ti whereas G is defined as a function of (x0,x1) only. The OT argument establishes uniqueness among deterministic maps x'_t = T(x_t) from the reference marginal to q_t; it does not establish existence or uniqueness among functions of (x0,x1). Unless the Markov-chain coupling is deterministic (i.e., π(x_ti|x0) and π(x1|x_ti) are Monge maps), which is not assumed and is false for minibatch OT plans, the conclusion does not follow. This matters because the spatial transcriptomics experiment uses the piecewise-linear regularizer (Eq. 11).
  3. [§2.3, Eq. (11) and Algorithm 1] Even if the fixed-t uniqueness theorems were correct, they apply to a single t and a single reference coupling. The training procedure optimizes one shared network G_phi over all intermediate times t_i and, for Eq. (11), over t∈[0,1] with a different Markov-chained coupling π_{t_i} for each i. No joint existence or uniqueness result is shown for this simultaneous optimization. The reference maps for different i are different, and the shared network may settle at a compromise that satisfies none of the marginal constraints exactly. This gap is load-bearing for the advertised 'unique interpolants' property.
minor comments (4)
  1. [Appendix A, proof of Theorem 2.1] The line 'G = T∘ℓ − ℓ' should read 'G = T∘ℓ'. This appears to be a typographical error, but it occurs in the proof of the uniqueness result and should be corrected.
  2. [§4.2, cell tracking experiment] The cell tracking evaluation is visual only. Adding quantitative EMD or a similar metric computed on the centroid trajectories, with error bars across subsamples, would substantially strengthen the claim that OT-ALI-CFM outperforms OT-CFM and OT-MFM on this task.
  3. [§2.3, Eq. (11)] The notation L_reg(φ; t_i) is confusing because the expectation is over t∼U[0,1] and the role of t_i is not explicit in the right-hand side. Please clarify the dependence on t_i and the distinction between the training-time index and the continuous time variable.
  4. [Theorem 2.2 statement] The theorem says 'a Markov–chained OT coupling π_{t_i} between q0 and q1', but π_{t_i} is defined as a distribution on triples (x0, x_{t_i}, x1). Rephrase to avoid ambiguity.

Circularity Check

0 steps flagged

No significant circularity: held-out predictions are genuine, uniqueness theorems are self-contained OT proofs, and the only self-citation (Tong et al. 2024) is non-load-bearing.

full rationale

The paper's central derivation is self-contained and its predictive claims are genuine. The ALI objective (Eq. 8) matches the observed intermediate marginals by optimizing the GAN loss (Eq. 7) plus a regularizer; that is the task definition (multi-marginal distribution matching), not a hidden prediction. The actual predictions in the paper are out-of-sample: in §4.4 the ST experiment 'omit[s] either section two (t=1/3) or three (t=2/3) from the training data' and computes EMD against the held-out section; in §4.3 the paper trains on 'the remaining data' and evaluates on 'held-out intermediate-time data'; the cell-tracking experiment trains on 10 subsamples per frame and then pushes all x0 samples through the learned vector field. None of these held-out marginals enters the training objective, so no fitted input is renamed as a prediction. The uniqueness theorems (Thm 2.1, 2.2) are proved in the paper's own appendix by a standard OT/Brenier argument (a.c. first marginal ⇒ unique deterministic Kantorovich plan ⇒ unique interpolant); they do not assume the target result, so there is no self-definitional circularity. Two caveats are correctness concerns, not circularity: (i) the theorems cover the constrained problems (Eqs. 10 and 13), while the actual objective (Eq. 8) is unconstrained with finite λ, so the 'guarantee unique interpolants' claim (end of §2.3 and in §5) is broader than what is proven; (ii) the appendix proof of Thm 2.1 contains an apparent typo, 'G = T∘ℓ − ℓ,' which should read 'T∘ℓ' as in Thm 2.2's proof. The only overlapping-author citation is Tong et al. (2024) (co-author N. Malkin), which supplies the Markov-chained OT coupling used in Eq. (11) and the OT-CFM baseline; that work is published, code-reproducible, and the coupling formula is restated in the paper itself, so the citation is minor and non-load-bearing. Score 2 reflects that single minor self-citation; I find no actual circular step.

Axiom & Free-Parameter Ledger

4 free parameters · 6 axioms · 0 invented entities

Everything the central claim rests on that the reader didn't pay for upstream: the method inherits flow matching and GAN theory; its own premises are the expressivity of the neural interpolant family, the adequacy of the fixed endpoint coupling, and the choice of regularizers. No new physical or mathematical entities are introduced.

free parameters (4)
  • Regularization weight λ = λ=1 (knot, cell), λ=10 (ST), adjusted in scRNA
    Hand-chosen in Eq (8); affects balance between GAN loss and uniqueness regularization.
  • Tumor purity threshold = 0.8
    Chosen based on visual agreement with H&E images (§4.4); determines which coordinates are labeled tumor and thus the target distributions.
  • Time-input Gaussian noise std = 0.001–0.01
    Added to generator time input to smooth interpolants (§D.2, D.3); selected per experiment.
  • Minibatch OT coupling batch size = 128
    Used to construct π; affects coupling quality (D.1-D.4).
axioms (6)
  • standard math The conditional flow matching objective (4) has gradients matching the intractable flow matching objective (2) (Lipman et al. 2023).
    Well-established background result; used without proof in §2.
  • standard math Under absolute continuity of the reference pushforward, the quadratic optimal transport map is unique (Brenier's theorem).
    Used in the proofs of Theorems 2.1 and 2.2 (§A).
  • domain assumption The GAN min-max objective (7) minimizes the Jensen-Shannon divergence between the target and the interpolant's pushforward.
    Standard GAN theory, but relies on discriminator capacity and convergence to the global optimum.
  • domain assumption For each intermediate time, the neural interpolant family is expressive enough to approximate the optimal transport map from the fixed endpoint coupling π to q_t, uniformly in t.
    Unstated but necessary for the method to succeed; not proven.
  • domain assumption The Markov-chained OT coupling π_ti for the piecewise-linear regularizer provides a valid reference for all t∈[0,1].
    Used in Eq. (11); depends on the choice of intermediate marginal and is not shown to be consistent across i.
  • domain assumption The minibatch OT plan between q0 and q1 is a good surrogate for the true coupling of the underlying dynamics.
    Used throughout; standard in flow matching literature but not verified for these datasets.

pith-pipeline@v1.3.0-alltime-deepseek · 17072 in / 13575 out tokens · 101232 ms · 2026-08-04T12:56:04.812179+00:00 · methodology

0 comments
read the original abstract

Learning the dynamics of a process given sampled observations at several time points is an important but difficult task in many scientific applications. When no ground-truth trajectories are available, but one has only snapshots of data taken at discrete time steps, the problem of modelling the dynamics, and thus inferring the underlying trajectories, can be solved by multi-marginal generalisations of flow matching algorithms. This paper proposes a novel flow matching method that overcomes the limitations of existing multi-marginal trajectory inference algorithms. Our proposed method, ALI-CFM, uses a GAN-inspired adversarial loss to fit neurally parametrised interpolant curves between source and target points such that the marginal distributions at intermediate time points are close to the observed distributions. The resulting interpolants are smooth trajectories that, as we show, are unique under mild assumptions. These interpolants are subsequently marginalised by a flow matching algorithm, yielding a trained vector field for the underlying dynamics. We showcase the versatility and scalability of our method by outperforming the existing baselines on spatial transcriptomics and cell tracking datasets, while performing on par with them on single-cell trajectory prediction. Code: https://github.com/mmacosha/adversarially-learned-interpolants.

Figures

Figures reproduced from arXiv: 2510.01159 by Jens Lagergren, Kirill Tamogashev, Nicola Branchini, Nikolay Malkin, Oskar Kviman, V\'ictor Elvira.

Figure 1
Figure 1. Figure 1: Our adversarially learnt interpolants (red curves) follow push-forward distributions (red densities) that approximate the intermediate-time marginal distributions 𝑞𝑡𝑗 and, by construction, have the correct end-marginals 𝑞0 and 𝑞1. Formally, a (deterministic) system in R 𝑛 can be described by an ordinary differential equa￾tion (ODE) 𝑑𝑥𝑡 = 𝑣𝑡(𝑥𝑡) 𝑑𝑡, where 𝑣𝑡 is a time-dependent vector field. In a bimarginal… view at source ↗
Figure 2
Figure 2. Figure 2: Comparison of CFM (Tong et al., 2024), MFM (Kapu [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: (a) and (b) show ten ALIs and piecewise linear interpolants, respectively, based on the subsampled cell tracking data, while (c) and (d) depict the resulting OT-ALI-CFM and OT-CFM vector fields. In the 2D figures we plot the centroids (one per frame) of (e) all of the segmentation data, (f) the subset training data, (g) the OT-ALI-CFM trajectories and (h) the OT-CFM trajectories. Note that the OT-CFM traje… view at source ↗
Figure 4
Figure 4. Figure 4: From left to right we visualize frame 1, 25, 50, 75, 100 and 115 in the sequence of [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: The aligned breast cancer H&E-stained images with overlaid scatter plots of tumor annotated [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: The raw H&E-stained images of the four tissue sections provided by Mo et al. (2024). [PITH_FULL_IMAGE:figures/full_fig_p015_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Different 3D views of all MFM interpolants are shown in [PITH_FULL_IMAGE:figures/full_fig_p016_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: H&E-stained histology image of the third breast cancer tissue section overlayed with [PITH_FULL_IMAGE:figures/full_fig_p017_8.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. ABC: Any-Subset Autoregression via Non-Markovian Diffusion Bridges in Continuous Time and Space

    cs.LG 2026-04 unverdicted novelty 7.0

    ABC enables any-subset autoregressive generation of continuous stochastic processes via non-Markovian diffusion bridges that track physical time and allow path-dependent conditioning.

Reference graph

Works this paper leans on

38 extracted references · 1 linked inside Pith · cited by 1 Pith paper

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

  2. [3]

    Wasserstein generative adversarial networks

    Martin Arjovsky, Soumith Chintala, and Léon Bottou. Wasserstein generative adversarial networks. International Conference on Machine Learning (ICML), 2017

    2017

  3. [4]

    Robust registration of calcium images by learned contrast synthesis

    John A Bogovic, Philipp Hanslovsky, Allan Wong, and Stephan Saalfeld. Robust registration of calcium images by learned contrast synthesis. In 2016 IEEE 13th International Symposium on Biomedical Imaging (ISBI), pages 1123--1126. IEEE, 2016

    2016

  4. [5]

    Learning single-cell perturbation responses using neural optimal transport

    Charlotte Bunne, Stefan G Stark, Gabriele Gut, Jacobo Sarabia Del Castillo, Mitch Levesque, Kjong-Van Lehmann, Lucas Pelkmans, Andreas Krause, and Gunnar R \"a tsch. Learning single-cell perturbation responses using neural optimal transport. Nature methods, 20 0 (11): 0 1759--1768, 2023

    2023

  5. [6]

    Multidimensional scaling

    Michael AA Cox and Trevor F Cox. Multidimensional scaling. In Handbook of data visualization, pages 315--347. Springer, 2008

    2008

  6. [7]

    Generative adversarial nets

    Ian J Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial nets. Advances in Neural Information Processing Systems (NeurIPS), 2014

    2014

  7. [8]

    Clonal evolution in cancer

    Mel Greaves and Carlo C Maley. Clonal evolution in cancer. Nature, 481 0 (7381): 0 306--313, 2012

    2012

  8. [9]

    Estimating epidemiologic dynamics from cross-sectional viral load distributions

    James A Hay, Lee Kennedy-Shaffer, Sanjat Kanjilal, Niall J Lennon, Stacey B Gabriel, Marc Lipsitch, and Michael J Mina. Estimating epidemiologic dynamics from cross-sectional viral load distributions. Science, 373 0 (6552): 0 eabh0635, 2021

    2021

  9. [10]

    The GAN is dead; long live the GAN ! a modern GAN baseline

    Nick Huang, Aaron Gokaslan, Volodymyr Kuleshov, and James Tompkin. The GAN is dead; long live the GAN ! a modern GAN baseline. Advances in Neural Information Processing Systems (NeurIPS), 2024

    2024

  10. [11]

    The relativistic discriminator: A key element missing from standard GAN

    Alexia Jolicoeur-Martineau. The relativistic discriminator: A key element missing from standard GAN . International Conference on Learning Representations (ICLR), 2019

    2019

  11. [12]

    Metric flow matching for smooth interpolations on the data manifold

    Kacper Kapuśniak, Peter Potaptchik, Teodora Reu, Leo Zhang, Alexander Tong, Michael Bronstein, Avishek Joey Bose, and Francesco Di Giovanni. Metric flow matching for smooth interpolations on the data manifold. Advances in Neural Information Processing Systems (NeurIPS), 2024

    2024

  12. [13]

    Kingma and Jimmy Ba

    Diederik P. Kingma and Jimmy Ba. Adam : A method for stochastic optimization. International Conference on Learning Representations (ICLR), 2015

    2015

  13. [14]

    Droplet barcoding for single-cell transcriptomics applied to embryonic stem cells

    Allon M Klein, Linas Mazutis, Ilke Akartuna, Naren Tallapragada, Adrian Veres, Victor Li, Leonid Peshkin, David A Weitz, and Marc W Kirschner. Droplet barcoding for single-cell transcriptomics applied to embryonic stem cells. Cell, 161 0 (5): 0 1187--1201, 2015

    2015

  14. [15]

    Multimodal single cell data integration challenge: results and lessons learned

    Christopher Lance, Malte D Luecken, Daniel B Burkhardt, Robrecht Cannoodt, Pia Rautenstrauch, Anna Laddach, Aidyn Ubingazhibov, Zhi-Jie Cao, Kaiwen Deng, Sumeer Khan, et al. Multimodal single cell data integration challenge: results and lessons learned. Advances in Neural Information Processing Systems (NeurIPS) Competitions and Demonstrations, 2021

    2021

  15. [16]

    Multi-marginal stochastic flow matching for high-dimensional snapshot data at irregular time points

    Justin Lee, Behnaz Moradijamei, and Heman Shakeri. Multi-marginal stochastic flow matching for high-dimensional snapshot data at irregular time points. International Conference on Machine Learning (ICML), 2025

    2025

  16. [17]

    Flow matching for generative modeling

    Yaron Lipman, Ricky TQ Chen, Heli Ben-Hamu, Maximilian Nickel, and Matt Le. Flow matching for generative modeling. International Conference on Learning Representations (ICLR), 2023

    2023

  17. [18]

    Flow straight and fast: Learning to generate and transfer data with rectified flow

    Xingchao Liu, Chengyue Gong, and Qiang Liu. Flow straight and fast: Learning to generate and transfer data with rectified flow. International Conference on Learning Representations (ICLR), 2023

    2023

  18. [19]

    Highly parallel genome-wide expression profiling of individual cells using nanoliter droplets

    Evan Z Macosko, Anindita Basu, Rahul Satija, James Nemesh, Karthik Shekhar, Melissa Goldman, Itay Tirosh, Allison R Bialas, Nolan Kamitaki, Emily M Martersteck, et al. Highly parallel genome-wide expression profiling of individual cells using nanoliter droplets. Cell, 161 0 (5): 0 1202--1214, 2015

    2015

  19. [20]

    The cell tracking challenge: 10 years of objective benchmarking

    Martin Maška, Vladimír Ulman, Pablo Delgado-Rodriguez, Estibaliz Gómez-de Mariscal, Tereza Nečasová, Fidel A Guerrero Peña, Tsang Ing Ren, Elliot M Meyerowitz, Tim Scherr, Katharina Löffler, et al. The cell tracking challenge: 10 years of objective benchmarking. Nature Methods, 20 0 (7): 0 1010--1020, 2023

    2023

  20. [21]

    Tumour evolution and microenvironment interactions in 2D and 3D space

    Chia-Kuei Mo, Jingxian Liu, Siqi Chen, Erik Storrs, Andre Luiz N Targino da Costa, Andrew Houston, Michael C Wendl, Reyka G Jayasinghe, Michael D Iglesia, Cong Ma, et al. Tumour evolution and microenvironment interactions in 2D and 3D space. Nature, 634 0 (8036): 0 1178--1186, 2024

    2024

  21. [22]

    Visualizing structure and transitions in high-dimensional biological data

    Kevin R Moon, David Van Dijk, Zheng Wang, Scott Gigante, Daniel B Burkhardt, William S Chen, Kristina Yim, Antonia van den Elzen, Matthew J Hirn, Ronald R Coifman, et al. Visualizing structure and transitions in high-dimensional biological data. Nature biotechnology, 37 0 (12): 0 1482--1492, 2019

    2019

  22. [23]

    A computational framework for solving Wasserstein lagrangian flows

    Kirill Neklyudov, Rob Brekelmans, Alexander Tong, Lazar Atanackovic, Qiang Liu, and Alireza Makhzani. A computational framework for solving Wasserstein lagrangian flows. International Conference on Machine Learning (ICML), 2024

    2024

  23. [24]

    Broken limits to life expectancy, 2002

    Jim Oeppen and James W Vaupel. Broken limits to life expectancy, 2002

    2002

  24. [25]

    Multisample flow matching: Straightening flows with minibatch couplings

    Aram-Alexandre Pooladian, Heli Ben-Hamu, Carles Domingo-Enrich, Brandon Amos, Yaron Lipman, and Ricky TQ Chen. Multisample flow matching: Straightening flows with minibatch couplings. International Conference on Machine Learning (ICML), 2023

    2023

  25. [26]

    Modeling complex system dynamics with flow matching across time and conditions

    Martin Rohbeck, Edward De Brouwer, Charlotte Bunne, Jan-Christian Huetter, Anne Biton, Kelvin Y Chen, Aviv Regev, and Romain Lopez. Modeling complex system dynamics with flow matching across time and conditions. International Conference on Learning Representations (ICLR), 2025

    2025

  26. [27]

    Optimal-transport analysis of single-cell gene expression identifies developmental trajectories in reprogramming

    Geoffrey Schiebinger, Jian Shu, Marcin Tabaka, Brian Cleary, Vidya Subramanian, Aryeh Solomon, Joshua Gould, Siyan Liu, Stacie Lin, Peter Berube, et al. Optimal-transport analysis of single-cell gene expression identifies developmental trajectories in reprogramming. Cell, 176 0 (4): 0 928--943, 2019

    2019

  27. [28]

    Fiji: an open-source platform for biological-image analysis

    Johannes Schindelin, Ignacio Arganda-Carreras, Erwin Frise, Verena Kaynig, Mark Longair, Tobias Pietzsch, Stephan Preibisch, Curtis Rueden, Stephan Saalfeld, Benjamin Schmid, et al. Fiji: an open-source platform for biological-image analysis. Nature methods, 9 0 (7): 0 676--682, 2012

    2012

  28. [29]

    Integrative spatial and genomic analysis of tumor heterogeneity with tumoroscope

    Shadi Shafighi, Agnieszka Geras, Barbara Jurzysta, Alireza Sahaf Naeini, Igor Filipiuk, Alicja Raczkowska, Hosein Toosi, ukasz Koperski, Kim Thrane, Camilla Engblom, et al. Integrative spatial and genomic analysis of tumor heterogeneity with tumoroscope. Nature Communications, 15 0 (1): 0 9343, 2024

    2024

  29. [30]

    Visualization and analysis of gene expression in tissue sections by spatial transcriptomics

    Patrik L Ståhl, Fredrik Salmén, Sanja Vickovic, Anna Lundmark, José Fernández Navarro, Jens Magnusson, Stefania Giacomello, Michaela Asp, Jakub O Westholm, Mikael Huss, et al. Visualization and analysis of gene expression in tissue sections by spatial transcriptomics. Science, 353 0 (6294): 0 78--82, 2016

    2016

  30. [31]

    Towards a better global loss landscape of GANs

    Ruoyu Sun, Tiantian Fang, and Alexander Schwing. Towards a better global loss landscape of GANs . Advances in Neural Information Processing Systems (NeurIPS), 2020

    2020

  31. [32]

    TrajectoryNet : A dynamic optimal transport network for modeling cellular dynamics

    Alexander Tong, Jessie Huang, Guy Wolf, David Van Dijk, and Smita Krishnaswamy. TrajectoryNet : A dynamic optimal transport network for modeling cellular dynamics. International Conference on Machine Learning (ICML), 2020

    2020

  32. [33]

    Improving and generalizing flow-based generative models with minibatch optimal transport

    Alexander Tong, Kilian Fatras, Nikolay Malkin, Guillaume Huguet, Yanlei Zhang, Jarrid Rector-Brooks , Guy Wolf, and Yoshua Bengio. Improving and generalizing flow-based generative models with minibatch optimal transport. Transactions on Machine Learning Research, 2024

    2024

  33. [34]

    The epigenotype

    Conrad H Waddington. The epigenotype. Endeavour, 1: 0 18--20, 1942

    1942

  34. [35]

    Distance metric learning with application to clustering with side-information

    Eric Xing, Michael Jordan, Stuart J Russell, and Andrew Ng. Distance metric learning with application to clustering with side-information. Advances in Neural Information Processing Systems (NeurIPS), 2002

    2002

  35. [36]

    Alignment and integration of spatial transcriptomics data

    Ron Zeira, Max Land, Alexander Strzalkowski, and Benjamin J Raphael. Alignment and integration of spatial transcriptomics data. Nature Methods, 19 0 (5): 0 567--575, 2022

    2022

  36. [37]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should not add it explicitly Type <Return> for now, but then later remove the command n...

  37. [38]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@first@sw \@firstoftwo \@ifundefined NAT@b*@#2 \@firstoftwo @num @NAT@ctr \@secondoft...

  38. [39]

    K PF!hg L_2ta: N , *,[ D @Ϸ@ p@; a ^zM@ LOԮc 2

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibsetup #1 @NAT@ctr @ @openbib .11em \@plus.33em \@minus.07em 4000 4000 `\.\@m @bibit...

    Pith/arXiv arXiv 2021