Pith. sign in

REVIEW 6 major objections 7 minor 32 references

Hashing for Protein Structure Similarity Search

T0 review · 6 major / 7 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read The paper claims that protein structures can be hashed into binary codes that make similarity search over six times more memory-efficient and over four times faster than existing vector methods, while achieving state-of-the-art accuracy.

desk verdict A genuinely useful hashing method for protein search, but the SOTA accuracy claim is one metric short and the empirical reporting needs tightening. read the letter →

arxiv 2411.08286 v1 pith:IFVFUWDA submitted 2024-11-13 cs.LG cs.AIq-bio.QM

Jin Han , Wu-Jun Li This is my paper · ORCID
classification cs.LGcs.AIq-bio.QM
keywords proteinstructuresimilaritysearchhashingbinaryhashcodesgraphneuralnetworkcontrastivelearningTM-scoreHammingdistancealignment-freerepresentation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper proposes POSH, which it calls the first hashing method for protein structure similarity search. The central claim is that a learned binary hash code of fixed length can stand in for a protein structure well enough that searching a database by Hamming distance matches or beats the accuracy of existing alignment-free real-valued vector methods, while using over six times less memory and running over four times faster. This matters because alignment-based search is impractical at the scale of modern structure databases, and even vector-based alignment-free methods need hundreds of gigabytes to index hundreds of millions of predicted structures. POSH trains a graph encoder with hand-crafted backbone features and contrastive learning, with TM-score as the ground-truth similarity signal, so the claim is that a 400-bit code preserves enough structural information for retrieval.

What carries the argument

The central object is the protein structure graph: nodes are amino acids, and edges are the `$k$` nearest neighbors by `$C_\alpha$` distance. The argument is carried by three components working together: hand-crafted node features (sine and cosine of backbone bond angles and dihedral angles) and edge features (Gaussian RBF encodings of interatomic distances among C, `$C_\alpha$`, N, O, and `$C_\beta$` atoms); a structure encoder that alternates node update and edge update message-passing layers, so that both node and edge interactions are modeled; and a hashing-plus-contrastive objective that trains the pooled node representation toward a binary sign code while making Hamming distances track TM-score similarity. A substructure sampling strategy guided by a TM-score floor of 0.9 diversifies positive examples, and a length-scaling term breaks ties among equal Hamming distances. The name of the method, POSH, stands for protein structure hashing.

What would settle it

Build a test set of protein pairs with nearly identical C-alpha traces and backbone angles but different side-chain packing or rewired nonlocal contacts; if POSH's Hamming rankings rate these pairs as similar while their TM-scores are low, the graph features rather than the hashing objective are the bottleneck, and the claim that the full method preserves structural similarity fails.

Watch

Extended reading notes

Core claim

On its own terms, the paper establishes that protein structure similarity can be compressed into short binary codes without sacrificing retrieval quality. POSH builds a graph per protein whose nodes are amino acids and whose edges connect each residue to its `$k$` nearest neighbors by `$C_\alpha$` distance; raw node features are sine and cosine encodings of backbone bond and dihedral angles, and raw edge features are RBF-encoded distances between backbone and `$C_\beta$` atoms. A graph neural network with alternating node and edge message-passing layers produces node representations, which are pooled and linearly mapped to a real-valued code; a hashing loss pulls this code toward its sign-binarized form, and an InfoNCE contrastive loss pushes hash codes of TM-score-similar proteins together and dissimilar ones apart. With the resulting 400-bit codes, POSH reports accuracy above all compared baselines on SCOPe (AUROC 0.9906, AUPRC 0.6853) and on the independent ind_PDB set (AUROC 0.9699, AUPRC 0.4719), with a memory compression ratio of roughly 1955 times relative to storing raw structures, and search time more than four times faster than the best real-valued baseline.

Load-bearing premise

The load-bearing premise is that the k-nearest-neighbor graph built from C-alpha distances plus backbone bond and dihedral angles contains enough information for the graph encoder to learn hash codes that reproduce TM-score similarity; if side chains, nonlocal contacts, or domain organization are essential to structural similarity, the hand-crafted graph caps accuracy regardless of the hashing layer.

Editorial extensions

If this is right

  • If the central claim holds, structure databases the size of the AlphaFold database can be indexed in about 11 GB of binary codes instead of hundreds of gigabytes, putting billion-scale search on a single machine.
  • A model trained on single-domain SCOPe structures transfers to the multi-domain ind_PDB set with top accuracy, indicating the learned hash codes capture reusable structural features rather than dataset-specific artifacts.
  • Code length controls a clean accuracy-cost trade-off: accuracy rises with code length from 64 to 512 bits, so deployment can pick the shortest code that meets a target recall.
  • Because Hamming distance ties are broken by protein length, retrieval rankings in the top-k regime improve, which is the regime users actually see.
  • The ablation results imply that edge features, substructure sampling, and distance scaling all contribute independently, so the reported gains are not coming from hashing alone.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A testable extension the paper leaves open is whether the same binary-code pipeline can be trained from predicted pairwise similarity scores that are cheaper than full alignment, since the current training set-up inherits the cost of alignment-based labels.
  • The graph representation uses only backbone and C-beta atoms; an inference is that adding side-chain contact maps or long-range residue-residue contacts could push accuracy higher on multi-domain proteins, where the paper itself reports a drop relative to single-domain SCOPe.
  • If the contrastive hash codes really track TM-score, they could also serve as a filtering pre-step before alignment: a shortlist of Hamming-near proteins could be re-ranked by the alignment procedure, reducing the costly alignment step to candidate proteins only.
  • The reported compression ratio of 1955 times suggests that far shorter codes may suffice for clustering large structure databases; the paper's code-length experiments only go down to 64 bits.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

6 major / 7 minor

Summary. The paper proposes POSH, a learned binary hashing method for protein structure similarity search (PSSS). POSH represents each protein as a k-nearest-neighbor graph over backbone atoms, extracts hand-crafted node and edge features, and learns a graph encoder with node and edge update layers. The encoder output is binarized via a hashing loss, and training uses an InfoNCE contrastive loss with a sampling strategy that includes a TM-score-guided substructure augmentation. At test time, proteins are encoded into binary codes and searched by Hamming distance with a length-scaling factor. Experiments on SCOPe, ind_PDB, and AlphaFold-scale databases claim state-of-the-art accuracy and substantial memory/time savings over alignment-free baselines.

Significance. If substantiated, the paper would make a useful contribution by introducing a hashing-based representation for PSSS that replaces real-valued vectors with binary codes, enabling more efficient large-scale search. The method is internally coherent, and the graph encoder with edge features is a reasonable design choice. The authors also contribute a substructure sampling strategy and a distance-scaling rule for Hamming search, which are sensible and likely transferable to other structure-based retrieval tasks. However, the central accuracy claim is not fully supported by the reported experiments: the paper claims state-of-the-art accuracy while acknowledging a Top-1 deficit on SCOPe, and the comparisons lack error bars and rerun baselines. The efficiency claims are plausible but need more careful benchmarking to rule out implementation artifacts.

major comments (6)
  1. [Abstract and Section 1 vs. Table 1] The abstract and introduction claim that POSH can 'outperform other methods to achieve state-of-the-art accuracy,' but Table 1 shows POSH Top-1 on SCOPe is 0.7242 versus GraSR's 0.7282. Section 4.2 acknowledges this by saying POSH trails on Top-1 and is only 'comparable' there. This is a direct contradiction of an unqualified claim. The accuracy claim needs to be qualified to the specific metrics where POSH actually leads, or the abstract should be rephrased. As written, the central claim of state-of-the-art accuracy is not supported by the paper's own table.
  2. [Section 4.1, Baseline comparison with GraSR] The results for GraSR are 'directly copied from its original paper' while POSH is trained and evaluated under the authors' own pipeline. The paper states that the same dataset and filtering criteria are used, but it does not demonstrate that the evaluation protocol is identical, including the definition of similar pairs, the Top-k calculation, the database/query split, and the post-processing of rankings. Without rerunning GraSR under the exact same evaluation code, the comparison may systematically favor POSH. The authors should either rerun GraSR (and ideally the other learning-based baseline) using the same evaluation harness, or provide a detailed point-by-point protocol match and, if discrepancies exist, quantify their effect.
  3. [Tables 1, 2, 4, and 5, statistical reliability] All accuracy tables report single point estimates with no error bars, confidence intervals, or number of seeds. This is a concern because the differences between POSH and GraSR are modest on several metrics (e.g., AUROC 0.9906 vs 0.9823, Top-5 0.7225 vs 0.7101 on SCOPe). Given stochastic mini-batch sampling, substructure sampling, and random initialization, these gaps could fall within run-to-run variance. The authors should report mean and standard deviation over at least three independent training runs, and ideally a paired significance test, before claiming consistent superiority.
  4. [Section 4.2, Figure 4, time cost benchmark] The time cost comparison lacks implementation details for the baselines. It is not stated whether all methods use exact linear scan, whether the baseline vector search is implemented with optimized BLAS routines or a straightforward loop, what hardware and software versions are used, or how the database vectors are stored and loaded. The reported speedup of 'more than four times' over GraSR may be sensitive to these implementation choices. The authors should specify the search algorithm and implementation environment for every method, and ideally release the benchmarking code so that the efficiency claim is reproducible.
  5. [Related Works and Section 1, 'first hashing method' claim] The paper cites Foldseek [18] but does not benchmark it or discuss it as a related approach. Foldseek also uses a discrete representation of protein structures (a 3Di alphabet) and is a state-of-the-art fast structure search tool. The claim that POSH is 'the first hashing method for PSSS' requires a clear distinction from Foldseek's discrete encoding, which could be viewed as a form of symbol-level hashing. At minimum, Foldseek should be included in the accuracy and efficiency comparisons, or the authors should explicitly argue why it is outside the scope of the claim. Without this, the novelty and SOTA positioning are unclear.
  6. [Section 4.2 and Table 3, AlphaFold database experiments] The paper reports time and memory results for databases up to the size of the AlphaFold database (approximately 200 million proteins), but it does not explain how the binary codes or real-valued vectors for these databases are obtained. If the authors actually encoded 200 million structures, the computational cost of that step should be reported. If the numbers are extrapolated or simulated, the extrapolation method must be described. This is necessary to judge whether the claimed speedup and memory savings are measured or projected, and whether the comparison is fair.
minor comments (7)
  1. [Section 3.2, Eq. (1)] The kNN graph construction may produce a directed graph if the nearest-neighbor relation is not symmetric; the aggregation uses the neighbor set N_i without clarifying whether the graph is symmetrized. Please state whether the graph is made undirected.
  2. [Section 3.3, Eq. (3)] The term (yt1)^2 uses '1' without definition. Please define 1 as the all-ones vector of appropriate dimension.
  3. [Section 4.3, Table 4] The ablation study does not state which dataset is used (SCOPe or ind_PDB). Please specify the dataset in the table caption or text.
  4. [Appendix B.2, Table 5] The text says 'with the growing code length, the accuracy of the models consistently improves,' but POSH-400* has AUROC 0.9708, which is lower than POSH-256's 0.9720. This is an internal inconsistency; either the statement should be qualified or the explanation for the non-monotonic AUROC should be given.
  5. [Section 3.4, substructure sampling] The procedure for precomputing the minimum sampling length per structure to satisfy TM(P, Ps) >= alpha is not described. Please provide the algorithm or a precise description of how this length is found.
  6. [References] Several references are incomplete: [18] lacks volume and page information, and [20] is cited as an arXiv preprint although a published version exists. Please update all references to their final published versions.
  7. [Figure 1] The architecture diagram is dense and does not label the node update, edge update, hashing layer, and contrastive loss components in the figure itself. Adding direct labels would improve readability.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity; the paper's claims are supported by independent empirical evaluation against an external TM-score benchmark.

full rationale

I walked the paper's claimed derivation chain. The central claims are that POSH is the first hashing method for PSSS, that binary codes reduce time and memory versus real-valued vectors, and that POSH achieves state-of-the-art accuracy. These are empirical claims, and the paper's evaluation uses external ground truth: TM-score computed by TM-align, with training positives defined by Eq. (7) and evaluation similarity defined by the same TM-score threshold. This is standard supervised learning against an external benchmark, not a self-definitional reduction, because the predicted quantity is the binary hash code and the ranking produced by Hamming distance, not the TM-score itself. The only self-citation is Eq. (3), the hashing loss drawn from the corresponding author's earlier work [21]; this is a component objective, not a load-bearing premise, and the accuracy/efficiency claims do not reduce to it. The citation to [27,28] for edge updating is not self-citation and concerns an architectural choice rather than the central prediction. I found no fitted parameter renamed as a prediction, no imported uniqueness theorem, and no known result merely renamed. The reported Top-1 shortfall on SCOPe and the copied GraSR baseline are potential evidence-quality concerns, but they are not circularity. Therefore the derivation is self-contained with respect to circularity.

Assumptions & free parameters 9 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new physical entities and no first-principles derivation; the ledger consists of chosen hyperparameters and domain assumptions that the empirical claims depend on. The key assumptions are that TM-score is the correct ground-truth similarity measure and that a kNN backbone graph is sufficient to learn similarity-preserving binary codes.

free parameters (9)
  • Hash code length d = 400
    Chosen to match the 400-dimensional real-valued vectors of DeepFold and GraSR; directly controls the memory and speed claims.
  • Similarity threshold rho = 0.9
    Used in Eq. (7) to define positive and negative training pairs; controls the difficulty of the contrastive task.
  • Substructure TM-score threshold alpha = 0.9
    Used in Eq. (8) to set the minimum sampling length for positive substructures; balances diversity against similarity to the query.
  • Hash loss weight lambda = 0.5
    Balances the contrastive loss and the hashing loss in Eq. (5).
  • Hash regularization gamma = 0.2
    Strength of the skew penalty in Eq. (3).
  • InfoNCE temperature tau = 0.07
    Controls the concentration of the contrastive distribution in Eq. (4).
  • Encoder depth L = 6
    Number of node and edge update layers in the structure encoder; chosen by hand.
  • kNN graph degree k = not reported
    Determines graph sparsity and the edge set in Section 3.1; no value is given in the paper.
  • RBF kernel count and bandwidth = not reported
    Determines edge feature dimensionality in Section 3.1; unspecified, affects the representation.
assumptions (4)
  • domain assumption TM-score, as computed by TM-align, is the correct ground-truth measure of protein structure similarity.
    Training labels (Eq. 7) and all evaluation metrics (Section 4.1) use TM-score; if this measure is wrong for downstream applications, the reported accuracy does not transfer.
  • domain assumption A kNN graph over C-alpha distances with local backbone angle features is a sufficient representation of protein structure for learning similarity-preserving codes.
    Section 3.1 defines the entire input representation from these features; the encoder cannot recover information excluded at graph construction.
  • ad hoc to paper After supervised training, sign binarization and Hamming distance preserve TM-score similarity well enough for retrieval.
    This is the central design choice of POSH; the paper validates it empirically but offers no guarantee, and the distance scaling in Eq. (6) is an additional ad hoc correction.
  • domain assumption Updating edge features during message passing is necessary for good performance.
    The paper invokes prior results (refs 27, 28) and its own ablation, but the edge update layer is a design choice rather than a derived requirement.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Hashing for Protein Structure Similarity Search." pith.science (2026). https://pith.science/paper/IFVFUWDA

@misc{pith2026241108286,
  author       = {Pith},
  title        = {Pith review of: Hashing for Protein Structure Similarity Search},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IFVFUWDA}},
  note         = {Machine review of arXiv:2411.08286}
}
abstract

Protein structure similarity search (PSSS), which tries to search proteins with similar structures, plays a crucial role across diverse domains from drug design to protein function prediction and molecular evolution. Traditional alignment-based PSSS methods, which directly calculate alignment on the protein structures, are highly time-consuming with high memory cost. Recently, alignment-free methods, which represent protein structures as fixed-length real-valued vectors, are proposed for PSSS. Although these methods have lower time and memory cost than alignment-based methods, their time and memory cost is still too high for large-scale PSSS, and their accuracy is unsatisfactory. In this paper, we propose a novel method, called $\underline{\text{p}}$r$\underline{\text{o}}$tein $\underline{\text{s}}$tructure $\underline{\text{h}}$ashing (POSH), for PSSS. POSH learns a binary vector representation for each protein structure, which can dramatically reduce the time and memory cost for PSSS compared with real-valued vector representation based methods. Furthermore, in POSH we also propose expressive hand-crafted features and a structure encoder to well model both node and edge interactions in proteins. Experimental results on real datasets show that POSH can outperform other methods to achieve state-of-the-art accuracy. Furthermore, POSH achieves a memory saving of more than six times and speed improvement of more than four times, compared with other methods.

Figures

Figures reproduced from arXiv: 2411.08286 by the authors.

Figure 1
Figure 1. The architecture of POSH Furthermore, all existing alignment-free methods adopt real-valued vectors for feature representation, which still have high time and memory cost for large-scale datasets. Hashing Hashing has been widely used for efficient search in many areas [21–25]. Hashing uses a hash function to map each data sample to a binary vector (or called hash code) while preserving the similarity in the original… view at source ↗
Figure 2
Figure 2. Illustration of bond angles and dihedral angles. The letter R denotes the side chain of the [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. POSH for protein structure similarity search. (a) In the training phase, the distance between [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Time cost of searching in databases of different sizes Time Cost We show the time cost of searching in databases of different sizes for SGM, SSEF, GraSR and POSH in [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Comprehensive performance comparison [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 24 canonical work pages

  1. [18]

    Fast and accurate protein structure search with foldseek

    Michel van Kempen, Stephanie S Kim, Charlotte Tumescheit, Milot Mirdita, Jeongjae Lee, Cameron LM Gilchrist, Johannes Söding, and Martin Steinegger. Fast and accurate protein structure search with foldseek. Nature Biotechnology, 2023

    work page 2023

  2. [1]

    Tm-align: a protein structure alignment algorithm based on the tm-score

    Yang Zhang and Jeffrey Skolnick. Tm-align: a protein structure alignment algorithm based on the tm-score. Nucleic Acids Research, 33(7):2302–2309, 2005

    work page 2005

  3. [2]

    Matt: local flexibility aids protein multiple structure alignment

    Matthew Menke, Bonnie Berger, and Lenore Cowen. Matt: local flexibility aids protein multiple structure alignment. PLoS Computational Biology, 4(1):e10, 2008. 9

    work page 2008

  4. [3]

    Protein structure alignment by incremental combinato- rial extension (ce) of the optimal path

    Ilya N Shindyalov and Philip E Bourne. Protein structure alignment by incremental combinato- rial extension (ce) of the optimal path. Protein Engineering, 11(9):739–747, 1998

    work page 1998

  5. [4]

    Flexible structure alignment by chaining aligned fragment pairs allowing twists

    Yuzhen Ye and Adam Godzik. Flexible structure alignment by chaining aligned fragment pairs allowing twists. Bioinformatics, 19(suppl_2):ii246–ii255, 2003

    work page 2003

  6. [5]

    Protein structure alignment beyond spatial proximity

    Sheng Wang, Jianzhu Ma, Jian Peng, and Jinbo Xu. Protein structure alignment beyond spatial proximity. Scientific Reports, 3(1):1–7, 2013

    work page 2013

  7. [6]

    The protein threading problem with sequence amino acid interaction preferences is np-complete

    Richard H Lathrop. The protein threading problem with sequence amino acid interaction preferences is np-complete. Protein Engineering, Design and Selection, 7(9):1059–1068, 1994

    work page 1994

  8. [7]

    Fox, Steven E

    Naomi K. Fox, Steven E. Brenner, and John-Marc Chandonia. Scope: Structural classification of proteins - extended, integrating SCOP and ASTRAL data and classification of new structures. Nucleic Acids Research, 42(Database-Issue):304–309, 2014

    work page 2014

Show all 32 references
  1. [8]

    Highly accurate protein structure prediction with alphafold

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

    2021

  2. [9]

    Accurate prediction of protein structures and interactions using a three-track neural network

    Minkyung Baek, Frank DiMaio, Ivan Anishchenko, Justas Dauparas, Sergey Ovchinnikov, Gyu Rie Lee, Jue Wang, Qian Cong, Lisa N Kinch, R Dustin Schaeffer, et al. Accurate prediction of protein structures and interactions using a three-track neural network. Science, 373(6557):871–...

    2021

  3. [10]

    Language models of protein sequences at the scale of evolution enable accurate structure prediction

    Zeming Lin, Halil Akin, Roshan Rao, Brian Hie, Zhongkai Zhu, Wenting Lu, Allan dos Santos Costa, Maryam Fazel-Zarandi, Tom Sercu, Sal Candido, et al. Language models of protein sequences at the scale of evolution enable accurate structure prediction. BioRxiv, 2022

    2022

  4. [11]

    Alphafold protein structure database: massively expanding the structural coverage of protein-sequence space with high-accuracy models

    Mihaly Varadi, Stephen Anyango, Mandar Deshpande, Sreenath Nair, Cindy Natassia, Galabina Yordanova, David Yuan, Oana Stroe, Gemma Wood, Agata Laydon, et al. Alphafold protein structure database: massively expanding the structural coverage of protein-sequence space with high-a...

    2022

  5. [12]

    Automatic classification of protein structure by using gauss integrals

    Peter Røgen and Boris Fain. Automatic classification of protein structure by using gauss integrals. Proceedings of the National Academy of Sciences, 100(1):119–124, 2003

    2003

  6. [13]

    Secondary structure spatial conformation footprint: a novel method for fast protein structure comparison and classification

    Elena Zotenko, Dianne P O’Leary, and Teresa M Przytycka. Secondary structure spatial conformation footprint: a novel method for fast protein structure comparison and classification. BMC Structural Biology, 6:1–12, 2006

    2006

  7. [14]

    Learning structural motif representations for efficient protein structure search

    Yang Liu, Qing Ye, Liwei Wang, and Jian Peng. Learning structural motif representations for efficient protein structure search. Bioinformatics, 34(17):i773–i780, 2018

    2018

  8. [15]

    Fast protein structure comparison through effective representation learning with contrastive graph neural networks

    Chunqiu Xia, Shi-Hao Feng, Ying Xia, Xiaoyong Pan, and Hong-Bin Shen. Fast protein structure comparison through effective representation learning with contrastive graph neural networks. PLoS Computational Biology, 18(3):e1009986, 2022

    2022

  9. [16]

    Scoring function for automated assessment of protein structure template quality

    Yang Zhang and Jeffrey Skolnick. Scoring function for automated assessment of protein structure template quality. Proteins: Structure, Function, and Bioinformatics, 57(4):702–710, 2004

    2004

  10. [17]

    Critically assessing the state-of-the-art in protein structure prediction

    DT Jones. Critically assessing the state-of-the-art in protein structure prediction. The Pharma- cogenomics Journal, 1(2):126–134, 2001

    2001

  11. [19]

    Convolutional lstm network: A machine learning approach for precipitation nowcasting

    Xingjian Shi, Zhourong Chen, Hao Wang, Dit-Yan Yeung, Wai-Kin Wong, and Wang-chun Woo. Convolutional lstm network: A machine learning approach for precipitation nowcasting. Advances in Neural Information Processing Systems, 2015. 10

    2015

  12. [20]

    Semi-supervised classification with graph convolutional networks

    Thomas N Kipf and Max Welling. Semi-supervised classification with graph convolutional networks. arXiv preprint arXiv:1609.02907, 2016

    2016 arXiv

  13. [21]

    Feature learning based deep supervised hashing with pairwise labels

    Wu-Jun Li, Sheng Wang, and Wang-Cheng Kang. Feature learning based deep supervised hashing with pairwise labels. In International Joint Conference on Artificial Intelligence, 2016

    2016

  14. [22]

    Hashnet: Deep learning to hash by continuation

    Zhangjie Cao, Mingsheng Long, Jianmin Wang, and Philip S Yu. Hashnet: Deep learning to hash by continuation. In International Conference on Computer Vision, 2017

    2017

  15. [23]

    Hashing based answer selection

    Dong Xu and Wu-Jun Li. Hashing based answer selection. In Association for the Advancement of Artificial Intelligence, 2020

    2020

  16. [24]

    Efficient passage retrieval with hashing for open-domain question answering

    Ikuya Yamada, Akari Asai, and Hannaneh Hajishirzi. Efficient passage retrieval with hashing for open-domain question answering. In Association for Computational Linguistics, 2021

    2021

  17. [25]

    Learning efficient hash codes for fast graph-based data similarity retrieval.IEEE Transactions on Image Processing, 30:6321–6334, 2021

    Jinbao Wang, Shuo Xu, Feng Zheng, Ke Lu, Jingkuan Song, and Ling Shao. Learning efficient hash codes for fast graph-based data similarity retrieval.IEEE Transactions on Image Processing, 30:6321–6334, 2021

    2021

  18. [26]

    Robust deep learning–based protein sequence design using proteinmpnn

    Justas Dauparas, Ivan Anishchenko, Nathaniel Bennett, Hua Bai, Robert J Ragotte, Lukas F Milles, Basile IM Wicky, Alexis Courbet, Rob J de Haas, Neville Bethel, et al. Robust deep learning–based protein sequence design using proteinmpnn. Science, 378(6615):49–56, 2022

    2022

  19. [27]

    Pifold: Toward effective and efficient protein inverse folding

    Zhangyang Gao, Cheng Tan, and Stan Z Li. Pifold: Toward effective and efficient protein inverse folding. arXiv preprint arXiv:2209.12643, 2022

    2022 arXiv

  20. [28]

    Lozano, Payel Das, and Jian Tang

    Zuobai Zhang, Minghao Xu, Arian Rokkum Jamasb, Vijil Chenthamarakshan, Aurélie C. Lozano, Payel Das, and Jian Tang. Protein representation learning by geometric structure pretraining. In International Conference on Learning Representations, 2023

    2023

  21. [29]

    Representation learning with contrastive predictive coding

    Aaron van den Oord, Yazhe Li, and Oriol Vinyals. Representation learning with contrastive predictive coding. arXiv preprint arXiv:1807.03748, 2018

    2018 arXiv

  22. [30]

    Contrastive representation learning for 3d protein structures

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

    2022 arXiv

  23. [31]

    How significant is a protein structure similarity with tm-score= 0.5? Bioinformatics, 26(7):889–895, 2010

    Jinrui Xu and Yang Zhang. How significant is a protein structure similarity with tm-score= 0.5? Bioinformatics, 26(7):889–895, 2010

    2010

  24. [32]

    The protein data bank

    Helen M Berman, John Westbrook, Zukang Feng, Gary Gilliland, Talapady N Bhat, Helge Weissig, Ilya N Shindyalov, and Philip E Bourne. The protein data bank. Nucleic Acids Research, 28(1):235–242, 2000. A Implementation Details In our implementation, we set the value of γ to 0.2...

    2000

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.