pith. sign in

arxiv: 2606.01628 · v1 · pith:DZR5G6VCnew · submitted 2026-06-01 · 🧬 q-bio.BM · cs.AI

Demystifying Multimodal Biomolecular Co-design With Intrinsic Geodesic Coupling

Keyue Qiu , Xintong Wang , Zhilong Zhang , Hao Zhou , Wei-Ying Ma This is my paper

Pith reviewed 2026-06-28 11:56 UTC · model grok-4.3

classification 🧬 q-bio.BM cs.AI
keywords biomolecular co-designgenerative modelsgeodesic couplingtemporal couplingprotein designdrug designmultimodal generationsequence-structure interplay
0
0 comments X

The pith

Optimizing temporal couplings between generative processes via geodesic alignment produces more consistent and valid biomolecular designs.

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

The paper establishes that the temporal coupling between separate generative processes for different biomolecular modalities, such as sequence and three-dimensional structure, is a key but overlooked design choice that affects consistency during joint generation. Fixed synchronous coupling, as used in prior work, introduces mismatched intermediate states and high-variance supervision signals that degrade the physical realism of outputs. The authors introduce GeoCoupling to learn intrinsic geodesic couplings that align the trajectories of these processes. Experiments on structure-based drug design and unconditional protein design show these learned couplings produce molecules with better physical validity and diversity than synchronous or random baselines. A sympathetic reader would care because modality consistency is essential for generating usable biomolecules in drug discovery and protein engineering.

Core claim

The central claim is that inappropriate temporal coupling of marginal generative processes is the primary overlooked source of inconsistent intermediate states and high-variance supervision in multimodal biomolecular co-design, and that optimizing these couplings through intrinsic geodesic alignment yields biomolecules with improved physical validity and diversity across structure-based drug design and unconditional protein design tasks.

What carries the argument

GeoCoupling, a framework that learns temporal couplings between heterogeneous modalities by aligning their generative trajectories along intrinsic geodesics.

If this is right

  • Biomolecules generated with learned couplings show higher physical validity than those from fixed synchronous coupling.
  • The same couplings increase design diversity while preserving sequence-structure consistency.
  • The framework applies equally to conditional tasks such as structure-based drug design and to unconditional protein generation.
  • Learned couplings outperform both synchronous and randomly chosen couplings in the reported benchmarks.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same coupling-optimization principle could be tested in other multimodal generative settings such as text-image or audio-video models.
  • Analytic derivations of optimal couplings might replace the current learned approach in some cases.
  • Modality consistency in generation may depend more on trajectory alignment than on explicit joint modeling alone.

Load-bearing premise

That the choice of temporal coupling is the main overlooked cause of high-variance supervision and inconsistent states that harm modality consistency.

What would settle it

If controlled experiments using the learned couplings produce no measurable gains in physical validity or diversity metrics over synchronous and random baselines on the same tasks and models, the claim would be falsified.

Figures

Figures reproduced from arXiv: 2606.01628 by Hao Zhou, Keyue Qiu, Wei-Ying Ma, Xintong Wang, Zhilong Zhang.

Figure 1
Figure 1. Figure 1: Decoupling Multimodal Generation via Learned Geodesics. A. Temporal Product Manifold: Visualization of coupling strategies on the joint noise space. While Synchronous coupling (gray) forces an isotropic diagonal path and Random coupling (red) suffers from high variance , GeoCoupling (blue) discovers a learned geodesic that navigates the optimal transport path. B. Conceptual illustration for the intermediat… view at source ↗
Figure 2
Figure 2. Figure 2: Co-designability vs. diversity trade-off (upper-right corner is better). 100 200 300 400 500 Protein Length 1 2 3 4 5 6 7 scRMSD ( ) Ours Ours (post-hoc) MultiFlow 100 200 300 400 500 Protein Length 72 74 76 78 80 82 84 pLDDT ( ) Ours Ours (post-hoc) MultiFlow 100 200 300 400 500 Protein Length 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Co-designability ( ) Ours Ours (post-hoc) MultiFlow [PITH_FULL_IMAGE:figures/full_fi… view at source ↗
Figure 3
Figure 3. Figure 3: Performance Comparison Across Protein Lengths. We evaluate scRMSD, pLDDT, and Co-designability across lengths 100-500. GeoCoupling maintains robustness even in OOD regimes (Length ≥ 400), whereas baselines degrade significantly. 4.2. Protein Design Dataset. We follow Campbell et al. (2024) and adopt the same training data, which consists of 18,694 monomeric protein structures obtained from Yim et al. (2023… view at source ↗
Figure 4
Figure 4. Figure 4: Comparison of the coupling learned by GeoCou￾pling against baselines on different modalities, showing a structure￾leading geodesic for biomolecules, prioritizing geometric context before semantic sequence decoding. Computational Efficiency of Bayesian Optimization. We measure the wall-clock time for optimization meth￾ods with and without Bayesian Optimization via Gaussian Process, i.e., a brute-force appro… view at source ↗
Figure 5
Figure 5. Figure 5: Visualization of the learned coupling against baselines [PITH_FULL_IMAGE:figures/full_fig_p021_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Planar distribution of the 1000 sampled points during the first fixed-coupling stage. The x-axis and y-axis represent Continuous Time and Discrete Time, respectively, with points colored by their corresponding Total Loss. The plot exhibits broad, space-filling coverage across the entire [0, 1] × [0, 1] search space, confirming that the Gaussian Process (GP) initialization effectively explores the parameter… view at source ↗
read the original abstract

Biomolecules such as proteins and small-molecule ligands play a central role in biological systems, arising from the tight interplay between sequence and three-dimensional structure. Recent generative models for biomolecular co-design aim to capture this interplay by jointly modeling coupled modalities. However, existing approaches largely adopt a parallel execution of marginal generative processes, implicitly enforcing fixed synchronous coupling. We argue that a critical but overlooked degree of freedom lies in how these marginal processes are temporally coupled during training and generation, where inappropriate coupling can introduce high-variance supervision and inconsistent intermediate states, affecting modality consistency. To address this, we introduce GeoCoupling, a systematic framework that optimizes for temporal couplings between heterogeneous modalities. Empirical results across structure-based drug design and unconditional protein design demonstrate the learned couplings consistently outperform synchronous and randomly coupled baselines, yielding biomolecules with improved physical validity and diversity.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 3 minor

Summary. The paper argues that multimodal biomolecular co-design models suffer from suboptimal fixed synchronous temporal coupling between modalities, which introduces high-variance supervision and inconsistent states; it introduces GeoCoupling, a framework that learns intrinsic geodesic couplings to optimize this degree of freedom, and reports that the learned couplings outperform synchronous and random baselines on physical validity and diversity metrics in structure-based drug design and unconditional protein design tasks.

Significance. If the empirical results hold, the work identifies and directly tests a previously overlooked modeling choice (temporal coupling) in multimodal generative models for biomolecules, with potential to improve consistency and diversity in co-design applications. The construction of explicit controls for the proposed mechanism strengthens the interpretability of the findings.

minor comments (3)
  1. [Abstract] The abstract states that learned couplings 'consistently outperform' baselines but provides no quantitative metrics, dataset sizes, or statistical details; the full manuscript should ensure these appear in the results section with error bars or significance tests.
  2. [Methods] Clarify the precise definition of 'intrinsic geodesic coupling' and how it is optimized (e.g., loss function or parameterization) early in the methods, as the current description risks remaining high-level for readers unfamiliar with the geometric framing.
  3. [Figures] Figure captions and axis labels should explicitly state the metrics used for validity and diversity (e.g., which validity criteria for small molecules or proteins) to allow direct comparison with prior work.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their supportive review, accurate summary of the work, and recommendation for minor revision. The referee correctly identifies the overlooked modeling choice of temporal coupling and the value of explicit controls in our experiments.

Circularity Check

0 steps flagged

No significant circularity; empirical comparison stands on external validation

full rationale

The paper's core contribution is an empirical demonstration that a learned temporal coupling (GeoCoupling) outperforms synchronous and random baselines on validity and diversity metrics in two biomolecular design tasks. The argument directly constructs and tests the claimed degree of freedom (temporal coupling choice) via explicit controls. No derivation chain, uniqueness theorem, or self-citation load-bearing premise is present in the abstract or described structure; the result is not forced by definition or by renaming a fitted quantity as a prediction. The claim is falsifiable against the reported baselines and does not reduce to its own inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review supplies no explicit free parameters, axioms, or invented entities.

pith-pipeline@v0.9.1-grok · 5683 in / 939 out tokens · 16539 ms · 2026-06-28T11:56:23.283334+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

228 extracted references · 15 canonical work pages · 1 internal anchor

  1. [1]

    Cell , year=

    A generative deep learning approach to de novo antibiotic design , author=. Cell , year=

  2. [2]

    The Twelfth International Conference on Learning Representations , year=

    Unified generative modeling of 3d molecules with bayesian flow networks , author=. The Twelfth International Conference on Learning Representations , year=

  3. [3]

    Science advances , volume=

    The human proteoform project: defining the human proteome , author=. Science advances , volume=. 2021 , publisher=

    2021

  4. [4]

    Nature , volume=

    A wealth of discovery built on the Human Genome Project—by the numbers , author=. Nature , volume=. 2021 , publisher=

    2021

  5. [5]

    Nature protocols , volume=

    A practical guide to large-scale docking , author=. Nature protocols , volume=. 2021 , publisher=

    2021

  6. [6]

    arXiv preprint arXiv:2403.07902 , year=

    DecompDiff: diffusion models with decomposed priors for structure-based drug design , author=. arXiv preprint arXiv:2403.07902 , year=

    arXiv

  7. [7]

    arXiv preprint arXiv:2303.03543 , year=

    3d equivariant diffusion for target-aware molecule generation and affinity prediction , author=. arXiv preprint arXiv:2303.03543 , year=

    arXiv

  8. [8]

    2024 , publisher=

    Schneuing, Arne and Harris, Charles and Du, Yuanqi and Didi, Kieran and Jamasb, Arian and Igashov, Ilia and Du, Weitao and Gomes, Carla and Blundell, Tom L and Lio, Pietro and others , journal=. 2024 , publisher=

    2024

  9. [9]

    Nature , volume=

    Ultra-large library docking for discovering new chemotypes , author=. Nature , volume=. 2019 , publisher=

    2019

  10. [10]

    Nature , volume=

    Structure-based drug design , author=. Nature , volume=. 1996 , publisher=

    1996

  11. [11]

    Drug Discovery Today , pages=

    Caught between a ROCK and a hard place: current challenges in structure-based drug design , author=. Drug Discovery Today , pages=. 2024 , publisher=

    2024

  12. [12]

    Highly accurate protein structure prediction with

    Jumper, John and Evans, Richard and Pritzel, Alexander and Green, Tim and Figurnov, Michael and Ronneberger, Olaf and Tunyasuvunakool, Kathryn and Bates, Russ and. Highly accurate protein structure prediction with. Nature , volume=. 2021 , publisher=

    2021

  13. [13]

    Accurate structure prediction of biomolecular interactions with

    Abramson, Josh and Adler, Jonas and Dunger, Jack and Evans, Richard and Green, Tim and Pritzel, Alexander and Ronneberger, Olaf and Willmore, Lindsay and Ballard, Andrew J and Bambrick, Joshua and others , journal=. Accurate structure prediction of biomolecular interactions with. 2024 , publisher=

    2024

  14. [14]

    RSC Medicinal Chemistry , volume=

    Target 2035--an update on private sector contributions , author=. RSC Medicinal Chemistry , volume=. 2023 , publisher=

    2035

  15. [15]

    Nature Reviews Chemistry , volume=

    CACHE (Critical Assessment of Computational Hit-finding Experiments): A public--private partnership benchmarking initiative to enable the development of computational methods for hit-finding , author=. Nature Reviews Chemistry , volume=. 2022 , publisher=

    2022

  16. [16]

    Nature reviews Drug discovery , volume=

    Unexplored therapeutic opportunities in the human genome , author=. Nature reviews Drug discovery , volume=. 2018 , publisher=

    2018

  17. [17]

    Wiley Interdisciplinary Reviews: Computational Molecular Science , volume=

    Understanding drug-likeness , author=. Wiley Interdisciplinary Reviews: Computational Molecular Science , volume=. 2011 , publisher=

    2011

  18. [18]

    Advanced drug delivery reviews , volume=

    The application of in silico drug-likeness predictions in pharmaceutical research , author=. Advanced drug delivery reviews , volume=. 2015 , publisher=

    2015

  19. [19]

    Patrick , title =

    Walters, W. Patrick , title =. Journal of Medicinal Chemistry , volume =. 2019 , doi =

    2019

  20. [20]

    and Reymond, Jean-Louis , title =

    Ruddigkeit, Lars and van Deursen, Ruud and Blum, Lorenz C. and Reymond, Jean-Louis , title =. Journal of Chemical Information and Modeling , volume =. 2012 , doi =

    2012

  21. [21]

    British journal of pharmacology , volume=

    Principles of early drug discovery , author=. British journal of pharmacology , volume=. 2011 , publisher=

    2011

  22. [22]

    Nature Machine Intelligence , pages=

    PocketFlow is a data-and-knowledge-driven structure-based molecular generative model , author=. Nature Machine Intelligence , pages=. 2024 , publisher=

    2024

  23. [23]

    Proteins: Structure, Function, and Bioinformatics , volume=

    Salt bridges: Geometrically specific, designable interactions , author=. Proteins: Structure, Function, and Bioinformatics , volume=. 2011 , publisher=

    2011

  24. [24]

    Advanced drug delivery reviews , volume=

    Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings , author=. Advanced drug delivery reviews , volume=. 1997 , publisher=

    1997

  25. [25]

    Current Opinion in Structural Biology , author =

    Deep learning approaches for de novo drug design:. Current Opinion in Structural Biology , author =. 2022 , pages =. doi:10.1016/j.sbi.2021.10.001 , abstract =

    work page doi:10.1016/j.sbi.2021.10.001 2022

  26. [26]

    ChemBioChem , author =

    Structure‐based. ChemBioChem , author =. 2023 , pages =. doi:10.1002/cbic.202200776 , language =

    work page doi:10.1002/cbic.202200776 2023

  27. [27]

    Drug Discovery Today , author =

    Docking-based generative approaches in the search for new drug candidates , volume =. Drug Discovery Today , author =. 2023 , pages =. doi:10.1016/j.drudis.2022.103439 , language =

    work page doi:10.1016/j.drudis.2022.103439 2023

  28. [28]

    Current Opinion in Structural Biology , author =

    Integrating structure-based approaches in generative molecular design , volume =. Current Opinion in Structural Biology , author =. 2023 , pages =. doi:10.1016/j.sbi.2023.102559 , abstract =

    work page doi:10.1016/j.sbi.2023.102559 2023

  29. [29]

    Cell Reports Medicine , author =

    Deep generative molecular design reshapes drug discovery , volume =. Cell Reports Medicine , author =. 2022 , pages =. doi:10.1016/j.xcrm.2022.100794 , abstract =

    work page doi:10.1016/j.xcrm.2022.100794 2022

  30. [30]

    Pharmaceuticals , volume=

    The role of ai in drug discovery: challenges, opportunities, and strategies , author=. Pharmaceuticals , volume=. 2023 , publisher=

    2023

  31. [31]

    arXiv preprint arXiv:2404.19230 , year=

    Deep Lead Optimization: Leveraging Generative AI for Structural Modification , author=. arXiv preprint arXiv:2404.19230 , year=

    arXiv

  32. [32]

    Wiley Interdisciplinary Reviews: Computational Molecular Science , volume=

    Generative models for molecular discovery: Recent advances and challenges , author=. Wiley Interdisciplinary Reviews: Computational Molecular Science , volume=. 2022 , publisher=

    2022

  33. [33]

    International conference on machine learning , pages=

    Deep unsupervised learning using nonequilibrium thermodynamics , author=. International conference on machine learning , pages=. 2015 , organization=

    2015

  34. [34]

    Advances in neural information processing systems , volume=

    Generative modeling by estimating gradients of the data distribution , author=. Advances in neural information processing systems , volume=

  35. [35]

    Advances in neural information processing systems , volume=

    Denoising diffusion probabilistic models , author=. Advances in neural information processing systems , volume=

  36. [36]

    Advances in neural information processing systems , volume=

    Variational diffusion models , author=. Advances in neural information processing systems , volume=

  37. [37]

    arXiv preprint arXiv:2211.15089 , year=

    Continuous diffusion for categorical data , author=. arXiv preprint arXiv:2211.15089 , year=

    Pith/arXiv arXiv

  38. [38]

    Advances in Neural Information Processing Systems , volume=

    Understanding diffusion objectives as the elbo with simple data augmentation , author=. Advances in Neural Information Processing Systems , volume=

  39. [39]

    International Conference on Learning Representations , year=

    Score-Based Generative Modeling through Stochastic Differential Equations , author=. International Conference on Learning Representations , year=

  40. [40]

    Advances in neural information processing systems , volume=

    Diffusion models beat gans on image synthesis , author=. Advances in neural information processing systems , volume=

  41. [41]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    High-resolution image synthesis with latent diffusion models , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  42. [42]

    International Conference on Machine Learning , pages=

    GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models , author=. International Conference on Machine Learning , pages=. 2022 , organization=

    2022

  43. [43]

    International conference on machine learning , pages=

    Learning transferable visual models from natural language supervision , author=. International conference on machine learning , pages=. 2021 , organization=

    2021

  44. [44]

    International Conference on Learning Representations , year=

    GeoDiff: A Geometric Diffusion Model for Molecular Conformation Generation , author=. International Conference on Learning Representations , year=

  45. [45]

    International Conference on Learning Representations (ICLR) , year=

    DiffDock: Diffusion Steps, Twists, and Turns for Molecular Docking , author =. International Conference on Learning Representations (ICLR) , year=

  46. [46]

    Nature Machine Intelligence , volume=

    A geometric deep learning approach to predict binding conformations of bioactive molecules , author=. Nature Machine Intelligence , volume=. 2021 , publisher=

    2021

  47. [47]

    The Eleventh International Conference on Learning Representations , year=

    Uni-Mol: A Universal 3D Molecular Representation Learning Framework , author=. The Eleventh International Conference on Learning Representations , year=

  48. [48]

    Journal of molecular biology , volume=

    Development and validation of a genetic algorithm for flexible docking , author=. Journal of molecular biology , volume=. 1997 , publisher=

    1997

  49. [49]

    Journal of cheminformatics , volume=

    ProLIF: a library to encode molecular interactions as fingerprints , author=. Journal of cheminformatics , volume=. 2021 , publisher=

    2021

  50. [50]

    Nature biotechnology , volume=

    MMseqs2 enables sensitive protein sequence searching for the analysis of massive data sets , author=. Nature biotechnology , volume=. 2017 , publisher=

    2017

  51. [51]

    International conference on machine learning , pages=

    Equivariant diffusion for molecule generation in 3d , author=. International conference on machine learning , pages=. 2022 , organization=

    2022

  52. [52]

    arXiv preprint arXiv:2210.05274 , year=

    Equivariant 3d-conditional diffusion models for molecular linker design , author=. arXiv preprint arXiv:2210.05274 , year=

    arXiv

  53. [53]

    International Conference on Machine Learning , pages=

    Geometric latent diffusion models for 3d molecule generation , author=. International Conference on Machine Learning , pages=. 2023 , organization=

    2023

  54. [54]

    International Conference on Machine Learning , pages=

    MolDiff: Addressing the Atom-Bond Inconsistency Problem in 3D Molecule Diffusion Generation , author=. International Conference on Machine Learning , pages=. 2023 , organization=

    2023

  55. [55]

    Lipman, Yaron and Chen, Ricky T. Q. and Ben-Hamu, Heli and Nickel, Maximilian and Le, Matt , month = feb, year =. Flow. doi:10.48550/arXiv.2210.02747 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2210.02747

  56. [56]

    Advances in Neural Information Processing Systems , volume=

    Equivariant Flow Matching with Hybrid Probability Transport for 3D Molecule Generation , author=. Advances in Neural Information Processing Systems , volume=

  57. [57]

    arXiv preprint arXiv:2403.15441 , year=

    Unified Generative Modeling of 3D Molecules via Bayesian Flow Networks , author=. arXiv preprint arXiv:2403.15441 , year=

    arXiv

  58. [58]

    The Twelfth International Conference on Learning Representations , year=

    Training-free Multi-objective Diffusion Model for 3D Molecule Generation , author=. The Twelfth International Conference on Learning Representations , year=

  59. [59]

    The eleventh international conference on learning representations , year=

    Equivariant energy-guided sde for inverse molecular design , author=. The eleventh international conference on learning representations , year=

  60. [60]

    The Eleventh International Conference on Learning Representations , year=

    DiGress: Discrete Denoising diffusion for graph generation , author=. The Eleventh International Conference on Learning Representations , year=

  61. [61]

    NeurIPS , year=

    Zero-Shot 3D Drug Design by Sketching and Generating , author=. NeurIPS , year=

  62. [62]

    and Masuda, Tomohide and Sunseri, Jocelyn and Jia, Andrew and Iovanisci, Richard B

    Francoeur, Paul G. and Masuda, Tomohide and Sunseri, Jocelyn and Jia, Andrew and Iovanisci, Richard B. and Snyder, Ian and Koes, David R. , title =. Journal of Chemical Information and Modeling , volume =. 2020 , doi =

    2020

  63. [63]

    , title =

    Trott, Oleg and Olson, Arthur J. , title =. Journal of Computational Chemistry , volume =. doi:https://doi.org/10.1002/jcc.21334 , url =. https://onlinelibrary.wiley.com/doi/pdf/10.1002/jcc.21334 , year =

    work page doi:10.1002/jcc.21334

  64. [64]

    Glide: a new approach for rapid, accurate docking and scoring. 2. Enrichment factors in database screening , author=. Journal of medicinal chemistry , volume=. 2004 , publisher=

    2004

  65. [65]

    Frontiers in pharmacology , volume=

    Molecular sets (MOSES): a benchmarking platform for molecular generation models , author=. Frontiers in pharmacology , volume=. 2020 , publisher=

    2020

  66. [66]

    ICLR2022 Machine Learning for Drug Discovery , year=

    Fragment-based ligand generation guided by geometric deep learning on protein-ligand structures , author=. ICLR2022 Machine Learning for Drug Discovery , year=

  67. [67]

    International Conference on Learning Representations , year=

    Molecule Generation For Target Protein Binding with Structural Motifs , author=. International Conference on Learning Representations , year=

  68. [68]

    Forty-first International Conference on Machine Learning , year=

    Unifying Bayesian Flow Networks and Diffusion Models through Stochastic Differential Equations , author=. Forty-first International Conference on Machine Learning , year=

  69. [69]

    Structure-based

    Schneuing, Arne and Du, Yuanqi and Harris, Charles and Jamasb, Arian and Igashov, Ilia and Du, Weitao and Blundell, Tom and Lió, Pietro and Gomes, Carla and Welling, Max and Bronstein, Michael and Correia, Bruno , month = oct, year =. Structure-based

  70. [70]

    Lu, Shuqi and Yao, Lin and Chen, Xi and Zheng, Hang and He, Di and Ke, Guolin , month = feb, year =

  71. [71]

    International Conference on Machine Learning , year=

    Generating 3D Molecules for Target Protein Binding , author=. International Conference on Machine Learning , year=

  72. [72]

    2022 , editor =

    Peng, Xingang and Luo, Shitong and Guan, Jiaqi and Xie, Qi and Peng, Jian and Ma, Jianzhu , booktitle =. 2022 , editor =

    2022

  73. [73]

    Generating

    Masuda, Tomohide and Ragoza, Matthew and Koes, David Ryan , month = nov, year =. Generating

  74. [74]

    Advances in Neural Information Processing Systems , volume=

    A. Advances in Neural Information Processing Systems , volume=. 2021 , author =

    2021

  75. [75]

    Current Opinion in Structural Biology , author =

    Structure-based drug design with geometric deep learning , volume =. Current Opinion in Structural Biology , author =. 2023 , pages =. doi:10.1016/j.sbi.2023.102548 , language =

    work page doi:10.1016/j.sbi.2023.102548 2023

  76. [76]

    The Eleventh International Conference on Learning Representations , year=

    3D Equivariant Diffusion for Target-Aware Molecule Generation and Affinity Prediction , author=. The Eleventh International Conference on Learning Representations , year=

  77. [77]

    2023 , editor =

    Guan, Jiaqi and Zhou, Xiangxin and Yang, Yuwei and Bao, Yu and Peng, Jian and Ma, Jianzhu and Liu, Qiang and Wang, Liang and Gu, Quanquan , booktitle =. 2023 , editor =

    2023

  78. [78]

    Learning

    Zhang, Zaixi and Liu, Qi , month = may, year =. Learning

  79. [79]

    Nature Computational Science , author =

    Learning on topological surface and geometric structure for. Nature Computational Science , author =. 2023 , file =. doi:10.1038/s43588-023-00530-2 , language =

    work page doi:10.1038/s43588-023-00530-2 2023

  80. [80]

    arXiv preprint arXiv:2308.07413 , year=

    Benchmarking Generated Poses: How Rational is Structure-based Drug Design with Generative Models? , author=. arXiv preprint arXiv:2308.07413 , year=

    arXiv

Showing first 80 references.