pith. sign in

arxiv: 2606.19374 · v1 · pith:B3JQ2E2Unew · submitted 2026-06-12 · 💻 cs.LG · cs.AI

Protein Representation Learning with Secondary-Structure and Energy-Filtered Hydrogen-Bond Graphs

Mohamed Mouhajir , Limei Wang , El Houcine Bergou , Hajar El Hammouti , Lamiae Azizi , Dongqi Fu This is my paper

Pith reviewed 2026-06-27 04:58 UTC · model grok-4.3

classification 💻 cs.LG cs.AI
keywords protein representation learninggraph neural networkssecondary structurehydrogen bondsprotein foldinginductive biasstructural motifs
0
0 comments X

The pith

Incorporating secondary structure assignments and energy-filtered hydrogen-bond edges improves protein graph representations.

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

Proteins fold into shapes stabilized by secondary structures like helices and sheets held by hydrogen bonds, yet many graph models for proteins use only sequence order or nearby atoms. This paper builds graphs where each residue node carries its secondary structure label and edges connect only residues linked by sufficiently strong hydrogen bonds. The resulting graph neural network learns representations that better reflect the principles of protein folding and stability. Evaluation on common protein benchmarks shows consistent gains over previous graph methods. The approach also produces representations whose connectivity matches known structural biology patterns.

Core claim

By augmenting residue-level node features with secondary structure assignments and constructing graph edges exclusively from energy-filtered hydrogen-bond interactions, the model captures both recurring local motifs and long-range stabilizing couplings that govern protein stability and function, leading to improved performance on protein benchmarks and greater alignment with biological motifs.

What carries the argument

Secondary-structure-augmented graph neural network with edges defined by energy-filtered hydrogen-bond interactions

If this is right

  • Consistent improvements over existing graph-based methods on protein benchmarks
  • Enhanced biological interpretability of the learned graph representations
  • Learned connectivity aligns with established structural motifs

Where Pith is reading between the lines

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

  • Removing the energy filter on hydrogen bonds should reduce performance toward levels seen with proximity-based graphs
  • The same node and edge construction could be tested on related tasks such as mutation stability prediction
  • The topology may transfer to modeling interactions between proteins and small molecules

Load-bearing premise

Secondary-structure assignments and energy-filtered hydrogen-bond interactions supply a graph topology that is meaningfully more informative for stability and function than sequence adjacency or geometric proximity alone.

What would settle it

Train the same GNN architecture on identical protein data but with three different edge sets (sequence adjacency, geometric proximity, and the proposed energy-filtered hydrogen bonds), then measure whether the hydrogen-bond version produces statistically significant gains on a held-out stability or function prediction benchmark.

Figures

Figures reproduced from arXiv: 2606.19374 by Dongqi Fu, El houcine Bergou, Hajar El Hammouti, Lamiae Azizi, Limei Wang, Mohamed Mouhajir.

Figure 1
Figure 1. Figure 1: Common secondary-structure elements such as [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Comparison of graph construction strategies. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Examples of 3D protein structures from the FOLD dataset annotated by DSSP. Colors indicate DSSP secondary-structure [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
read the original abstract

Graph-based representations are widely used in protein modeling, yet many existing approaches rely primarily on sequence adjacency or geometric proximity, which only partially reflect the principles governing protein folding. Proteins instead adopt complex three-dimensional conformations organized around secondary structure elements, such as $\alpha$-helices and $\beta$-sheets, which encode recurring local motifs and stabilizing hydrogen-bond interactions. In this work, we introduce a secondary-structure-aware graph neural network for protein representation learning. Residue-level node representations are augmented with secondary structure assignments, and graph edges are constructed from hydrogen-bond interactions filtered by their energetic strength. This design enables the model to capture both local structural context and long-range couplings that are central to protein stability and function. We evaluate the proposed approach on commonly used protein benchmarks and observe consistent improvements over existing graph-based methods. In addition, the resulting graph representations offer enhanced biological interpretability, as the learned connectivity aligns with established structural motifs. These findings suggest that incorporating secondary structure and energy-filtered hydrogen-bond topology provides an effective inductive bias for protein representation learning. The code is released at https://github.com/mohamedmohamed2021/SSProNet

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 / 2 minor

Summary. The manuscript introduces SSProNet, a graph neural network for protein representation learning. Residue nodes are augmented with secondary-structure assignments, and edges are defined via energetically filtered hydrogen-bond interactions rather than sequence adjacency or geometric proximity alone. The model is evaluated on standard protein benchmarks, where it reports consistent gains over prior graph-based methods and improved alignment with known structural motifs. Code is released publicly.

Significance. If the empirical gains hold under rigorous controls, the work supplies a biologically motivated inductive bias that directly encodes secondary-structure elements and stabilizing hydrogen bonds, which are central to folding and function. Public code release strengthens reproducibility.

minor comments (2)
  1. [Abstract] Abstract: the claim of 'consistent improvements' should be accompanied in the results section by explicit baseline comparisons, dataset sizes, and statistical significance tests so that the magnitude of gains can be assessed directly.
  2. Methods: the precise energy threshold and filtering procedure for hydrogen-bond edges should be stated with a reference to the underlying energy function or software used.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive assessment of our work on SSProNet and the recommendation for minor revision. The report does not list any specific major comments, so we have no points to address point-by-point. We are happy to incorporate any minor suggestions that may arise during the revision process.

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper proposes a graph construction for protein representation learning that augments nodes with secondary-structure assignments and defines edges via energy-filtered hydrogen-bond interactions. This is presented as an inductive bias choice and is evaluated empirically on external protein benchmarks with reported gains over sequence-adjacency and geometric baselines. No equations, parameter fits, or self-citations are described that reduce any claimed result to an input by construction. The central claim rests on benchmark comparisons that are independent of the method definition itself.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The approach rests on domain assumptions from structural biology rather than new mathematical axioms or invented entities.

axioms (2)
  • domain assumption Secondary structure assignments accurately capture recurring local motifs that influence folding.
    Used to augment node representations.
  • domain assumption Hydrogen-bond energy calculations reliably identify stabilizing long-range interactions.
    Used to filter graph edges.

pith-pipeline@v0.9.1-grok · 5753 in / 1026 out tokens · 24808 ms · 2026-06-27T04:58:18.417020+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

51 extracted references · 6 canonical work pages

  1. [1]

    Functional evolution of two subtly different (similar) folds.BMC structural biology, 1(1):5, 2001

    Vishal Agrawal and Radha KV Kishan. Functional evolution of two subtly different (similar) folds.BMC structural biology, 1(1):5, 2001

    2001

  2. [2]

    A minimal sequence code for switching protein structure and function.Proceedings of the National Academy of Sciences, 106(50):21149–21154, 2009

    Patrick A Alexander, Yanan He, Yihong Chen, John Orban, and Philip N Bryan. A minimal sequence code for switching protein structure and function.Proceedings of the National Academy of Sciences, 106(50):21149–21154, 2009

    2009

  3. [3]

    Learning to represent programs with graphs.arXiv preprint arXiv:1711.00740, 2017

    Miltiadis Allamanis, Marc Brockschmidt, and Mahmoud Khademi. Learning to represent programs with graphs.arXiv preprint arXiv:1711.00740, 2017

    Pith/arXiv arXiv 2017

  4. [4]

    Graphqa: protein model quality assessment using graph convolutional networks.Bioinformatics, 37(3):360–366, 2021

    Federico Baldassarre, David Menéndez Hurtado, Arne Elofsson, and Hossein Azizpour. Graphqa: protein model quality assessment using graph convolutional networks.Bioinformatics, 37(3):360–366, 2021

    2021

  5. [5]

    Network biology: understanding the cell’s functional organization.Nature reviews genetics, 5(2):101–113, 2004

    Albert-Laszlo Barabasi and Zoltan N Oltvai. Network biology: understanding the cell’s functional organization.Nature reviews genetics, 5(2):101–113, 2004

    2004

  6. [6]

    Mace: Higher order equivariant message passing neural networks for fast and accurate force fields

    Ilyes Batatia, David P Kovacs, Gregor Simm, Christoph Ortner, and Gabor Csanyi. Mace: Higher order equivariant message passing neural networks for fast and accurate force fields. In S. Koyejo, S. Mohamed, A. Agarwal, D. Belgrave, K. Cho, and A. Oh, editors,Advances in Neural Information Processing Systems, volume 35, pages 11423–11436. Curran Associates, ...

    2022

  7. [7]

    Berman, John Westbrook, Zukang Feng, Gary Gilliland, T

    Helen M. Berman, John Westbrook, Zukang Feng, Gary Gilliland, T. N. Bhat, Helge Weissig, Ilya N. Shindyalov, and Philip E. Bourne. The protein data bank. Nucleic Acids Research, 28(1):235–242, January 2000. doi: 10.1093/nar/28.1.235

    work page doi:10.1093/nar/28.1.235 2000

  8. [8]

    Bekkers, and Max Welling

    Johannes Brandstetter, Rob Hesselink, Elise van der Pol, Erik J. Bekkers, and Max Welling. Geometric and physical quantities improve E(3) equivariant message passing.CoRR, abs/2110.02905, 2021

    arXiv 2021

  9. [9]

    Geometric deep learning: going beyond euclidean data.IEEE Signal Processing Magazine, 34(4):18–42, 2017

    Michael M Bronstein, Joan Bruna, Yann LeCun, Arthur Szlam, and Pierre Van- dergheynst. Geometric deep learning: going beyond euclidean data.IEEE Signal Processing Magazine, 34(4):18–42, 2017

    2017

  10. [10]

    Dana, Aleksandras Gutmanas, Nidhi Tyagi, Guoying Qi, Claire O’Donovan, Maria Martin, and Sameer Velankar

    Jose M. Dana, Aleksandras Gutmanas, Nidhi Tyagi, Guoying Qi, Claire O’Donovan, Maria Martin, and Sameer Velankar. Sifts: updated structure inte- gration with function, taxonomy and sequences resource allows 40-fold increase in coverage of structure-based annotations for proteins.Nucleic Acids Research, 47(D1):D482–D489, January 2019. doi: 10.1093/nar/gky1114

    work page doi:10.1093/nar/gky1114 2019

  11. [11]

    Networks, crowds, and markets.Econ

    David Easley and Jon Kleinberg. Networks, crowds, and markets.Econ. Theory, 26:1–28, 2010

    2010

  12. [12]

    Continuous- discrete convolution for geometry-sequence modeling in proteins

    Hehe Fan, Zhangyang Wang, Yi Yang, and Mohan Kankanhalli. Continuous- discrete convolution for geometry-sequence modeling in proteins. InThe Eleventh International Conference on Learning Representations, 2022

    2022

  13. [13]

    Matthias Fey and Jan E. Lenssen. Fast graph representation learning with PyTorch geometric. InICLR Workshop on Representation Learning on Graphs and Manifolds, 2019

    2019

  14. [14]

    Protein interface prediction using graph convolutional networks.Advances in neural information processing systems, 30, 2017

    Alex Fout, Jonathon Byrd, Basir Shariat, and Asa Ben-Hur. Protein interface prediction using graph convolutional networks.Advances in neural information processing systems, 30, 2017

    2017

  15. [15]

    DPPIN: A biological repository of dynamic protein- protein interaction network data

    Dongqi Fu and Jingrui He. DPPIN: A biological repository of dynamic protein- protein interaction network data. InIEEE International Conference on Big Data, Big Data 2022, Osaka, Japan, December 17-20, 2022, pages 5269–5277. IEEE,

    2022

  16. [16]

    URL https://doi.org/10.1109/ BigData55660.2022.10020904

    doi: 10.1109/BIGDATA55660.2022.10020904. URL https://doi.org/10.1109/ BigData55660.2022.10020904

    work page doi:10.1109/bigdata55660.2022.10020904 2022

  17. [17]

    Torvik, and Jingrui He

    Dongqi Fu, Liri Fang, Ross Maciejewski, Vetle I. Torvik, and Jingrui He. Meta- learned metrics over multi-evolution temporal graphs. InKDD ’22: The 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, Washington, DC, USA, August 14 - 18, 2022, pages 367–377. ACM, 2022. doi: 10.1145/3534678. 3539313. URL https://doi.org/10.1145/3534678.3539313

    work page doi:10.1145/3534678 2022

  18. [18]

    Graph u-nets

    Hongyang Gao and Shuiwang Ji. Graph u-nets. Ininternational conference on machine learning, pages 2083–2092. PMLR, 2019

    2083

  19. [19]

    Topology-aware graph pooling networks

    Hongyang Gao, Yi Liu, and Shuiwang Ji. Topology-aware graph pooling networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(12):4512–4518, 2021

    2021

  20. [20]

    Good: A graph out-of- distribution benchmark.Advances in Neural Information Processing Systems, 35: 2059–2073, 2022

    Shurui Gui, Xiner Li, Limei Wang, and Shuiwang Ji. Good: A graph out-of- distribution benchmark.Advances in Neural Information Processing Systems, 35: 2059–2073, 2022

    2059

  21. [21]

    Dssp 4: Fair annotation of protein secondary structure.Protein Science, 2025

    Maarten L Hekkelman, Daniel Álvarez Salmoral, Anastassis Perrakis, and Rob- bie P Joosten. Dssp 4: Fair annotation of protein secondary structure.Protein Science, 2025

    2025

  22. [22]

    Contrastive representation learning for 3d protein structures.arXiv preprint arXiv:2205.15675, 2022

    Pedro Hermosilla and Timo Ropinski. Contrastive representation learning for 3d protein structures.arXiv preprint arXiv:2205.15675, 2022

    arXiv 2022

  23. [23]

    Intrinsic-extrinsic convolution and pooling for learning on 3d protein structures

    Pedro Hermosilla, Marco Schäfer, Matěj Lang, Gloria Fackelmann, Pere Pau Vázquez, Barbora Kozlíková, Michael Krone, Tobias Ritschel, and Timo Ropinski. Intrinsic-extrinsic convolution and pooling for learning on 3d protein structures. arXiv preprint arXiv:2007.06252, 2020

    arXiv 2007

  24. [24]

    Jie Hou, Badri Adhikari, and Jianlin Cheng. Deepsf: deep convolutional neural network for mapping protein sequences to folds.Bioinformatics, 34(8):1295–1303, Protein Representation Learning with Secondary-Structure and Energy-Filtered Hydrogen-Bond Graphs BIOKDD ’26, August 10, 2026, Jeju, Korea 2018

    2026

  25. [25]

    Learning complete protein representation by dynamically coupling of sequence and structure.Advances in Neural Information Processing Systems, 37:137673–137697, 2024

    Bozhen Hu, Cheng Tan, Jun Xia, Yue Liu, Lirong Wu, Jiangbin Zheng, Yongjie Xu, Yufei Huang, and Stan Z Li. Learning complete protein representation by dynamically coupling of sequence and structure.Advances in Neural Information Processing Systems, 37:137673–137697, 2024

    2024

  26. [26]

    Open graph benchmark: Datasets for machine learning on graphs.Advances in neural information processing systems, 33:22118–22133, 2020

    Weihua Hu, Matthias Fey, Marinka Zitnik, Yuxiao Dong, Hongyu Ren, Bowen Liu, Michele Catasta, and Jure Leskovec. Open graph benchmark: Datasets for machine learning on graphs.Advances in neural information processing systems, 33:22118–22133, 2020

    2020

  27. [27]

    Learning from protein structure with geometric vector perceptrons.arXiv preprint arXiv:2009.01411, 2020

    Bowen Jing, Stephan Eismann, Patricia Suriana, Raphael JL Townshend, and Ron Dror. Learning from protein structure with geometric vector perceptrons.arXiv preprint arXiv:2009.01411, 2020

    arXiv 2009

  28. [28]

    Highly accurate protein structure prediction with alphafold

    John Jumper, Richard Evans, Alexander Pritzel, Tim Green, Michael Figurnov, Olaf Ronneberger, Kathryn Tunyasuvunakool, Russ Bates, Augustin Žídek, Anna Potapenko, et al. Highly accurate protein structure prediction with alphafold. nature, 596(7873):583–589, 2021

    2021

  29. [29]

    Kingma and Jimmy Ba

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

    2015

  30. [30]

    Kipf and Max Welling

    Thomas N. Kipf and Max Welling. Semi-supervised classification with graph convolutional networks.ICLR, 2017

    2017

  31. [31]

    Structural patterns in globular proteins

    Michael Levitt and Cyrus Chothia. Structural patterns in globular proteins. Nature, 261(5561):552–558, 1976

    1976

  32. [32]

    Directed weight neural networks for protein structure representation learning.arXiv preprint arXiv:2201.13299, 2022

    Jiahan Li. Directed weight neural networks for protein structure representation learning.arXiv preprint arXiv:2201.13299, 2022

    arXiv 2022

  33. [33]

    Spherical message passing for 3d molecular graphs

    Yi Liu, Limei Wang, Meng Liu, Yuchao Lin, Xuan Zhang, Bora Oztekin, and Shuiwang Ji. Spherical message passing for 3d molecular graphs. InInternational Conference on Learning Representations (ICLR), 2022

    2022

  34. [34]

    Pdb-wide collection of binding data: current status of the pdbbind database.Bioinformatics, 31(3):405–412, February 2015

    Zhihai Liu, Yan Li, Li Han, Jie Li, Jie Liu, Zhixiong Zhao, Wei Nie, Yuchen Liu, and Renxiao Wang. Pdb-wide collection of binding data: current status of the pdbbind database.Bioinformatics, 31(3):405–412, February 2015. doi: 10.1093/ bioinformatics/btu626

    2015

  35. [35]

    One transformer can understand both 2d & 3d molecular data.arXiv preprint arXiv:2210.01765, 2022

    Shengjie Luo, Tianlang Chen, Yixian Xu, Shuxin Zheng, Tie-Yan Liu, Liwei Wang, and Di He. One transformer can understand both 2d & 3d molecular data.arXiv preprint arXiv:2210.01765, 2022

    arXiv 2022

  36. [36]

    Weisfeiler and leman go neural: Higher-order graph neural networks

    Christopher Morris, Martin Ritzert, Matthias Fey, William L Hamilton, Jan Eric Lenssen, Gaurav Rattan, and Martin Grohe. Weisfeiler and leman go neural: Higher-order graph neural networks. InProceedings of the AAAI conference on artificial intelligence, volume 33, pages 4602–4609, 2019

    2019

  37. [37]

    Omelchenko, Michael Y

    Mikhail V. Omelchenko, Michael Y. Galperin, Yuri I. Wolf, et al. Non-homologous isofunctional enzymes: A systematic analysis of alternative solutions in enzyme evolution.Biology Direct, 5:31, 2010. doi: 10.1186/1745-6150-5-31

    work page doi:10.1186/1745-6150-5-31 2010

  38. [38]

    Pytorch: An imperative style, high-performance deep learning library

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, Alban Des- maison, Andreas Kopf, Edward Yang, Zachary DeVito, Martin Raison, Alykhan Tejani, Sasank Chilamkurthy, Benoit Steiner, Lu Fang, Junjie Bai, and Soumith Chintala. Pytorch: An imperative style, high-per...

    2019

  39. [39]

    Springer Science & Business Media, 2013

    Georg E Schulz and R Heiner Schirmer.Principles of protein structure. Springer Science & Business Media, 2013

    2013

  40. [40]

    Dynamic edge-conditioned filters in convolutional neural networks on graphs

    Martin Simonovsky and Nikos Komodakis. Dynamic edge-conditioned filters in convolutional neural networks on graphs. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 3693–3702, 2017

    2017

  41. [41]

    Multi-scale representation learning on proteins.Advances in Neural Information Processing Systems, 34, 2021

    Vignesh Ram Somnath, Charlotte Bunne, and Andreas Krause. Multi-scale representation learning on proteins.Advances in Neural Information Processing Systems, 34, 2021

    2021

  42. [42]

    Anderson, Stephan Eismann, Risi Kondor, Russ Alt- man, and Ron O

    Raphael John Lamarre Townshend, Martin Vögele, Patricia Adriana Suriana, Alexander Derry, Alexander Powers, Yianni Laloudakis, Sidhika Balachandar, Bowen Jing, Brandon M. Anderson, Stephan Eismann, Risi Kondor, Russ Alt- man, and Ron O. Dror. ATOM3D: Tasks on molecules in three dimensions. In Thirty-fifth Conference on Neural Information Processing System...

    2021

  43. [43]

    Learning hierar- chical protein representations via complete 3d graph networks.ICLR, 2022

    Limei Wang, Haoran Liu, Yi Liu, Jerry Kurtin, and Shuiwang Ji. Learning hierar- chical protein representations via complete 3d graph networks.ICLR, 2022

    2022

  44. [44]

    The pdbbind database: Collection of binding affinities for protein–ligand complexes with known three-dimensional structures.Journal of Medicinal Chemistry, 47(12): 2977–2980, 2004

    Renxiao Wang, Xueliang Fang, Yipin Lu, and Shaomeng Wang. The pdbbind database: Collection of binding affinities for protein–ligand complexes with known three-dimensional structures.Journal of Medicinal Chemistry, 47(12): 2977–2980, 2004. doi: 10.1021/jm030580l

    work page doi:10.1021/jm030580l 2004

  45. [45]

    A theoretically-principled sparse, connected, and rigid graph repre- sentation of molecules

    Shih-Hsin Wang, Yuhao Huang, Justin M Baker, Yuan-En Sun, Qi Tang, and Bao Wang. A theoretically-principled sparse, connected, and rigid graph repre- sentation of molecules. InThe Thirteenth International Conference on Learning Representations, 2025

    2025

  46. [46]

    Advanced graph and se- quence neural networks for molecular property prediction and drug discovery

    Zhengyang Wang, Meng Liu, Youzhi Luo, Zhao Xu, Yaochen Xie, Limei Wang, Lei Cai, Qi Qi, Zhuoning Yuan, Tianbao Yang, et al. Advanced graph and se- quence neural networks for molecular property prediction and drug discovery. Bioinformatics, 38(9):2579–2586, 2022

    2022

  47. [47]

    Moleculenet: a bench- mark for molecular machine learning.Chemical science, 9(2):513–530, 2018

    Zhenqin Wu, Bharath Ramsundar, Evan N Feinberg, Joseph Gomes, Caleb Ge- niesse, Aneesh S Pappu, Karl Leswing, and Vijay Pande. Moleculenet: a bench- mark for molecular machine learning.Chemical science, 9(2):513–530, 2018

    2018

  48. [48]

    Task-agnostic graph explanations.Advances in neural information processing systems, 35:12027–12039, 2022

    Yaochen Xie, Sumeet Katariya, Xianfeng Tang, Edward Huang, Nikhil Rao, Karthik Subbian, and Shuiwang Ji. Task-agnostic graph explanations.Advances in neural information processing systems, 35:12027–12039, 2022

    2022

  49. [49]

    Periodic graph transformers for crystal material property prediction.Advances in Neural Information Processing Systems, 35:15066–15080, 2022

    Keqiang Yan, Yi Liu, Yuchao Lin, and Shuiwang Ji. Periodic graph transformers for crystal material property prediction.Advances in Neural Information Processing Systems, 35:15066–15080, 2022

    2022

  50. [50]

    Graphfm: Improving large-scale gnn training via feature momentum

    Haiyang Yu, Limei Wang, Bokun Wang, Meng Liu, Tianbao Yang, and Shuiwang Ji. Graphfm: Improving large-scale gnn training via feature momentum. In International conference on machine learning, pages 25684–25701. PMLR, 2022

    2022

  51. [51]

    Protein representation learning by geometric structure pretraining.ICLR, 2023

    Zuobai Zhang, Minghao Xu, Arian Jamasb, Vijil Chenthamarakshan, Aurelie Lozano, Payel Das, and Jian Tang. Protein representation learning by geometric structure pretraining.ICLR, 2023

    2023