Pith. sign in

REVIEW 4 major objections 5 minor 39 references

Pan-protein Design Learning Enables Task-adaptive Generalization for Low-resource Enzyme Design

T0 review · 4 major / 5 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 By aligning protein structures with pretrained protein language model representations, CrossDesign transfers sequence knowledge into structure-based design, outperforming larger baselines on enzyme design and zero-shot mutation effect…

desk verdict Solid inverse folding paper with strong enzyme benchmarks, but the ProteinGym zero-shot fitness claim rests on an unvalidated scoring rule. read the letter →

arxiv 2411.17795 v1 pith:5QPILLGE submitted 2024-11-26 q-bio.QM cs.AIcs.LG

Jiangbin Zheng , Ge Wang , Han Zhang , Stan Z. Li This is my paper · ORCID
classification q-bio.QMcs.AIcs.LG
keywords inversefoldingproteindesignenzymelanguagemodelscross-modalalignmentzero-shotmutationeffectpredictionlow-resourcelearningtransfer
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 sets out to show that the data bottleneck in computational protein design can be broken by borrowing the knowledge already captured in pretrained protein language models. It presents CrossDesign, which aligns a structure-to-sequence model with a frozen language model during training, then discards the language model at inference. On the CATH benchmark the model reports a perplexity of 3.67 (a lower-is-better measure of sequence uncertainty) and 57.11% amino-acid recovery (the fraction of native residues reproduced), above ProteinMPNN, PiFold, and ESM-IF. On out-of-domain enzyme datasets its recovery reaches about 60%, and on ProteinGym it reports the best average matching rank among compared inverse folding and language models. The payoff, if the method is right, is a practical route to enzyme design when structure-sequence pairs are scarce.

What carries the argument

The load-bearing component is the asymmetric auto-encoder formed by the frozen pretrained protein language model (PPLM), the linear alignment unit (LAU), and the shared decoder in the auxiliary stream. During training the LAU maps PPLM output into the same dimension as the temporal graph transformer's alignment features, and the inter-modality alignment loss (InterMA) applies knowledge distillation between those two representations at temperature τ=8; a second distillation, cross-layer consistency (CLC), does the same between early geometric features and late contextual features in the main stream. This forces the structure encoder to internalize sequence-context knowledge without needing the PPLM at inference. The main stream's transformation-enhanced GVP (tGVP) makes the representation invariant to rigid-body transformations, and the shared triangular-mask decoder enables both autoregressive sampling for de novo sequence design and non-autoregressive parallel scoring for mutation effects.

What would settle it

Compute the Spearman correlation between the NAR cross-entropy score and experimentally measured fitness for each of the 217 ProteinGym assays; if the average rank of CrossDesign falls below the ESM-IF baseline, or if a sizable fraction of assays show near-zero or negative correlation, the zero-shot fitness claim is refuted. Alternatively, test the scoring rule against a single designed enzyme mutant in the lab and compare predicted and measured activity.

Watch

Extended reading notes

Core claim

The central claim is that pretrained protein language models contain implicit structural and functional knowledge that can be transferred to an inverse folding model through an auxiliary supervised stream, rather than by fine-tuning the language model itself. CrossDesign contains a main structure-to-sequence stream built from transformation-enhanced geometric vector perceptrons, a temporal graph transformer, and a shared autoregressive decoder, plus an auxiliary stream that feeds the native sequence into a frozen PPLM and maps its representation back into the same feature space. Two knowledge-distillation losses—cross-layer consistency within the structure stream and inter-modality alignment between streams—push the structural features to mimic the language model's contextual features. The model is trained end-to-end with an exponential cross-entropy loss, and at inference only the structure stream runs. With non-autoregressive decoding, the same model scores mutant sequences against the wildtype structure, which the paper treats as a zero-shot fitness predictor.

Load-bearing premise

The zero-shot fitness claim rests on treating the cross-entropy between the model's mutant-sequence probabilities under the wildtype structure and the wildtype sequence as a universal fitness proxy, with no per-assay calibration or wet-lab validation.

Editorial extensions

If this is right

  • CrossDesign reports the best perplexity and recovery among the compared models on CATH, Ts50, Ts500, and the enzyme benchmarks, including the first reported recovery above 60% on Ts50.
  • The framework produces consistent out-of-domain enzyme design results, with fold-averaged recovery around 60% and up to about 64% on PET hydrolases.
  • Zero-shot mutation effect scoring with non-autoregressive decoding achieves an average matching rank of 0.445 on ProteinGym, ahead of compared language models and inverse folding baselines.
  • Training with the PPLM stream and distilling into the structure stream gives low-resource structure models an advantage over models trained on much larger structure-sequence corpora.
  • The shared decoder with AR and NAR modes lets one trained model serve both conditional design and mutation effect prediction without task-specific retraining.

Reading between the lines

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

  • If the cross-entropy scoring rule is taken at face value, the method should transfer to other property prediction tasks, such as antibody affinity or thermostability, where per-position sequence likelihood under a structure may correlate with function; this is not tested in the paper.
  • The alignment strategy suggests a general recipe: any frozen sequence-only model could be used to regularize a structure-based predictor in low-data regimes, not just protein language models for inverse folding.
  • A straightforward stress test would be to train CrossDesign on progressively smaller fractions of CATH to see whether the PPLM alignment preserves its advantage as data shrinks; the paper does not report such a curve.
  • The fitness predictor is uncalibrated; adding per-assay calibration or an ensemble over multiple PPLMs could raise or clarify the rank, but the paper does not explore it.
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

4 major / 5 minor

Summary. The manuscript introduces CrossDesign, an encoder-decoder protein inverse-folding framework that couples a structure-to-sequence stream with a frozen pretrained protein language model (PPLM) stream through cross-modal alignment, cross-layer consistency, and a shared decoder. The model supports autoregressive decoding for conditioning design and non-autoregressive decoding for mutation-effect scoring. The authors evaluate CrossDesign on CATH, Ts50/Ts500, enzyme-specific datasets (EnzPetDB and EnzFoldDB), and ProteinGym, reporting state-of-the-art perplexity and amino-acid recovery on CATH, improved out-of-domain recovery on Ts50/Ts500, AAR scores around 60% on enzyme folds, and a best matching rank of 0.445 on ProteinGym zero-shot fitness prediction. The paper also contributes curated enzyme structure-sequence benchmarks for PET hydrolases and fold-aware enzyme design.

Significance. If the claims hold, the paper is a useful contribution to low-resource enzyme design. The architectural idea of transferring sequence-level knowledge from a PPLM into a structure-conditioned inverse-folding model is coherent, and the evaluation strategy is laudably external: the headline CATH, Ts50/Ts500, enzyme, and ProteinGym results are not obtained by fitting the model to the test labels. The protein-design evaluations use standard benchmarks and the zero-shot fitness score is not trained on ProteinGym labels, which reduces circularity concerns. The enzyme benchmark collection is also a practical asset. However, the paper's central fitness-prediction claim rests on a non-standard scoring rule that is not validated against any baseline, the main quantitative claims lack error bars and statistical tests, and the absence of training details, code, and data prevents independent verification. These gaps are fixable within a revision, but they currently leave the strength of the claims overstated.

major comments (4)
  1. [§III.C.3, Eq. (8)] The zero-shot fitness score is not the model's likelihood of the mutant sequence, which is the standard inverse-folding proxy, but the cross-entropy between the model's mutant-conditioned output distribution and the wildtype sequence: Effectmut = CE(Prob_mut, Y_wild). The paper provides no theoretical justification, no per-assay Spearman correlations, and no comparison against standard alternatives such as ESM-IF's mutant log-likelihood or a simple substitution-score baseline. Without such controls, the reported ProteinGym rank of 0.445 does not demonstrate that the score captures fitness rather than sequence similarity or model artifacts. This is load-bearing because mutation-effect prediction is a stated headline contribution and the abstract claims the model 'excels in fitness prediction.' The limitation section, admitted lack of wet-lab validation, also omits this missing computational validation of the scoring rule itself.
  2. [§III.C.3 and abstract] The paper reports 'best matching rank (0.445)' on ProteinGym, but never defines what 'matching rank' means, nor how it relates to the Spearman ρ values cited for VESPA (0.437) and ESM-IF (0.422). ProteinGym's standard evaluation metric is per-assay Spearman's rank correlation, averaged or median-aggregated; if 'matching rank' is a different aggregation, the comparison to baseline ρ values is not meaningful. Additionally, the introduction says ProteinGym spans 'over 178 standardized deep mutational scanning assays,' while Section III.C.3 says 'all 217 ProteinGym proteins'; the discrepancy needs to be resolved, as it changes the scope of the claim.
  3. [§III.B, Table I] The headline CATH and Ts50/Ts500 results are reported as single numbers with no error bars, no repeated-seed variance, and no statistical significance tests, so the statement that CrossDesign outperforms ProteinMPNN, PiFold, and ESM-IF 'by a significant margin' is unsupported. This is especially important because the margins on some Ts500 recovery values are small (60.90 vs. 60.42 for PiFold). The paper also says 'no canonical training and validation sets exist for Ts50 or Ts500,' but does not specify which checkpoint or training protocol was used to produce the Ts50/Ts500 rows, which is needed to interpret those generalization claims.
  4. [§II, Eq. (2)–Eq. (5)] Reproducibility is compromised by omitted training details that are load-bearing for the method's identity. The loss weights for LexpCE, LCLC, LInterMA, and LasyAE are not given; the specific PPLM is not named; and architecture hyperparameters such as hidden dimension d, number of TGT layers, decoder layers, and number of attention heads are absent. Without these, a reader cannot tell whether the reported gains come from the cross-modal alignment, the exponential CE, the auxiliary reconstruction loss, or simply from scaling, and the claimed contribution cannot be reimplemented. The authors should provide the full training configuration, code, or at least a detailed hyperparameter table.
minor comments (5)
  1. [§II.3, Eq. (2)] The summation in Eq. (2) starts at b=0, but batch indices conventionally run from 1 to B; please correct the range or explain the indexing.
  2. [Figure 2 caption] The caption contains stray text '输出aar_enzyme.pdf', which appears to be an artifact; it should be removed.
  3. [§III.C, Figure 2] The acronym AAR is used in tables and figures but is never defined; please spell out 'amino-acid recovery' at first use.
  4. [§III.A, Enzyme datasets] EnzFoldDB is described as having 691 enzymes with 'each fold has an equal number of enzymes,' but 691 is not divisible by 10; the fold sizes in Figure 2 are 69 or 70, so the text should say 'approximately equal.'
  5. [§III.A, ProteinGym] The sentence 'encompasses a broad collection of over 178 standardized deep mutational scanning assays, spanning millions of mutated sequences' is inconsistent with the later '217 ProteinGym proteins'; please use a single consistent count and cite the exact ProteinGym version.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central claims are anchored by external benchmarks and zero-shot evaluations, and no result is forced by construction or by self-citation.

full rationale

The paper's main performance claims are tested on external benchmarks that are not used to fit the model: CATH test splits are independent of the CATH training split, Ts50/Ts500 are explicitly out-of-domain sets with no canonical training set, EnzPetDB and EnzFoldDB are newly collected enzyme datasets evaluated by amino-acid recovery, and ProteinGym is an external zero-shot mutation-effect benchmark. The model is trained with native-sequence reconstruction losses (expCE, Eq. 2; asyAE, Eq. 5) and auxiliary alignment losses (Eqs. 3-4), none of which consume ProteinGym labels or enzyme test-set labels. The fitness scoring rule in Eq. 8, Effectmut = CE(Prob_mut, Y_wild), is a modeling choice: it defines a zero-shot proxy from the mutant-conditioned output distribution against the wildtype sequence. Although this proxy is non-standard and the paper does not validate it against simpler baselines or report per-assay correlations, it is not circular because the ProteinGym fitness labels are external and the score is not derived from, fitted to, or defined as those labels. The paper's own limitation statement that mutation-effect verification lacks wet-lab validation is an acknowledged correctness/evidentiary gap, not a circular reduction. The numerous self-citations (e.g., MetaEnzyme, MMDesign, and related prior work by the authors) are not used as load-bearing justification for the current claims; the architecture is described directly and evaluated against independent baselines and data. No equation in the paper reduces by construction to its own input, no fitted parameter is renamed as a prediction, and no uniqueness theorem is imported from the authors to force a choice. Therefore no circular step meets the quoted-reduction standard, and the appropriate finding is no significant circularity.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

The central contribution is a training procedure rather than a new physical entity, so no invented entities are proposed. The free parameters are mostly undisclosed hyperparameters, and the domain assumptions listed above are load-bearing for the design and fitness claims.

free parameters (3)
  • loss weighting coefficients for LexpCE, LCLC, LInterMA, LasyAELoss = not reported
    The total objective is a weighted sum of four losses, but no weights are listed; these coefficients are chosen by hand and affect all reported results.
  • distillation temperature tau = 8
    Set to 8 in CLC and InterMA losses (Eqs. 3 and 4); a hyperparameter selected without ablation.
  • architecture hyperparameters (hidden dimension d, TGT layers, decoder layers) = not reported
    Dimensions and layer counts are not specified; they are tuned choices that influence performance.
assumptions (3)
  • domain assumption Pretrained protein language model embeddings carry transferable structural and functional knowledge that helps structure-to-sequence design even when structural data are scarce.
    The whole PPLM supervised stream in Section II.2 assumes that aligning PPLM sequence features with structure features improves design; no causal or quantitative validation of this assumption is provided beyond the final benchmark numbers.
  • domain assumption The mutation effect score Effectmut = CE(Prob_mut, Y_wild) computed from the NAR decoder is a valid zero-shot proxy for protein fitness across diverse deep mutational scanning assays.
    Adopted in Eq. 8 with no empirical calibration or comparison to alternative scoring schemes; if this premise fails, the ProteinGym superiority claim collapses.
  • domain assumption The datasets EnzPetDB and EnzFoldDB are correctly curated and representative of enzyme families, with no train/test leakage between enzyme folds and ProteinGym.
    Dataset construction is described in Section III.A but no data release or leakage analysis is given, and ProteinGym evaluation is called zero-shot.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Pan-protein Design Learning Enables Task-adaptive Generalization for Low-resource Enzyme Design." pith.science (2026). https://pith.science/paper/5QPILLGE

@misc{pith2026241117795,
  author       = {Pith},
  title        = {Pith review of: Pan-protein Design Learning Enables Task-adaptive Generalization for Low-resource Enzyme Design},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5QPILLGE}},
  note         = {Machine review of arXiv:2411.17795}
}
read the original abstract

Computational protein design (CPD) offers transformative potential for bioengineering, but current deep CPD models, focused on universal domains, struggle with function-specific designs. This work introduces a novel CPD paradigm tailored for functional design tasks, particularly for enzymes-a key protein class often lacking specific application efficiency. To address structural data scarcity, we present CrossDesign, a domain-adaptive framework that leverages pretrained protein language models (PPLMs). By aligning protein structures with sequences, CrossDesign transfers pretrained knowledge to structure models, overcoming the limitations of limited structural data. The framework combines autoregressive (AR) and non-autoregressive (NAR) states in its encoder-decoder architecture, applying it to enzyme datasets and pan-proteins. Experimental results highlight CrossDesign's superior performance and robustness, especially with out-of-domain enzymes. Additionally, the model excels in fitness prediction when tested on large-scale mutation data, showcasing its stability.

Figures

Figures reproduced from arXiv: 2411.17795 by the authors.

Figure 1
Figure 1. (a). The proposed CrossDesign framework contains a Structure-to-Sequence Stream (Str2Seq Stream) and an auxiliary PPLM Supervised Stream [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. AAR of enzyme design. Orange for EnzPetDB; Green and blue for [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 10 canonical work pages

  1. [1]

    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

  2. [2]

    Alphadesign: A graph protein design method and benchmark on alphafolddb

    Zhangyang Gao, Cheng Tan, Stan Li, et al. Alphadesign: A graph protein design method and benchmark on alphafolddb. arXiv preprint arXiv:2202.01079, 2022

    arXiv 2022

  3. [3]

    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

    arXiv 2022

  4. [4]

    Self-mutual distillation learning for continuous sign language recognition

    Aiming Hao, Yuecong Min, and Xilin Chen. Self-mutual distillation learning for continuous sign language recognition. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 11303– 11312, 2021

    work page 2021

  5. [5]

    Learning inverse folding from millions of predicted structures

    Chloe Hsu, Robert Verkuil, Jason Liu, Zeming Lin, Brian Hie, Tom Sercu, Adam Lerer, and Alexander Rives. Learning inverse folding from millions of predicted structures. bioRxiv, 2022

    work page 2022

  6. [6]

    Learning complete protein representation by deep coupling of sequence and structure

    Bozhen Hu, Cheng Tan, Jun Xia, Jiangbin Zheng, Yufei Huang, Lirong Wu, Yue Liu, Yongjie Xu, and Stan Z Li. Learning complete protein representation by deep coupling of sequence and structure. bioRxiv, pages 2023–07, 2023

    2023

  7. [7]

    Protein language models and structure prediction: Connection and progression

    Bozhen Hu, Jun Xia, Jiangbin Zheng, Cheng Tan, Yufei Huang, Yongjie Xu, and Stan Z Li. Protein language models and structure prediction: Connection and progression. arXiv preprint arXiv:2211.16742 , 2022

    arXiv 2022

  8. [8]

    Protein 3d graph structure learning for robust structure-based protein property prediction

    Yufei Huang, Siyuan Li, Lirong Wu, Jin Su, Haitao Lin, Odin Zhang, Zihan Liu, Zhangyang Gao, Jiangbin Zheng, and Stan Z Li. Protein 3d graph structure learning for robust structure-based protein property prediction. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 12662–12670, 2024

    2024

Show all 39 references
  1. [9]

    Data-efficient protein 3d geometric pretraining via refinement of diffused protein structure decoy

    Yufei Huang, Lirong Wu, Haitao Lin, Jiangbin Zheng, Ge Wang, and Stan Z Li. Data-efficient protein 3d geometric pretraining via refinement of diffused protein structure decoy. arXiv preprint arXiv:2302.10888 , 2023

    2023 arXiv

  2. [10]

    Generative models for graph-based protein design

    John Ingraham, Vikas Garg, Regina Barzilay, and Tommi Jaakkola. Generative models for graph-based protein design. Advances in neural information processing systems , 32, 2019

    2019

  3. [11]

    Learning from protein structure with geometric vector perceptrons

    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

    2009 arXiv

  4. [12]

    Highly accurate protein structure prediction with alphafold

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

    2021

  5. [13]

    Technical approaches to chinese sign language processing: a review

    Suhail Muhammad Kamal, Yidong Chen, Shaozi Li, Xiaodong Shi, and Jiangbin Zheng. Technical approaches to chinese sign language processing: a review. IEEE Access, 7:96926–96935, 2019

    2019

  6. [14]

    Direct prediction of profiles of sequences compatible with a protein structure by neural networks with fragment-based local and energy-based nonlocal profiles

    Zhixiu Li, Yuedong Yang, Eshel Faraggi, Jian Zhan, and Yaoqi Zhou. Direct prediction of profiles of sequences compatible with a protein structure by neural networks with fragment-based local and energy-based nonlocal profiles. Proteins: Structure, Function, and Bioinformatics ...

    2014

  7. [15]

    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

  8. [16]

    Machine learning-aided engineering of hydrolases for pet depolymerization

    Hongyuan Lu, Daniel J Diaz, Natalie J Czarnecki, Congzhi Zhu, Wantae Kim, Raghav Shroff, Daniel J Acosta, Bradley R Alexander, Hannah O Cole, Yan Zhang, et al. Machine learning-aided engineering of hydrolases for pet depolymerization. Nature, 604(7907):662–667, 2022

    2022

  9. [17]

    Language models enable zero-shot prediction of the effects of mutations on protein function

    Joshua Meier, Roshan Rao, Robert Verkuil, Jason Liu, Tom Sercu, and Alex Rives. Language models enable zero-shot prediction of the effects of mutations on protein function. Advances in Neural Information Processing Systems, 34:29287–29303, 2021

    2021

  10. [18]

    Proteingym: Large-scale benchmarks for protein design and fitness prediction

    Pascal Notin, Aaron W Kollasch, Daniel Ritter, Lood van Niekerk, Steffanie Paul, Hansen Spinner, Nathan Rollins, Ada Shaw, Ruben Weitzman, Jonathan Frazer, et al. Proteingym: Large-scale benchmarks for protein design and fitness prediction. bioRxiv, pages 2023–12, 2023

    2023

  11. [19]

    Spin2: Predicting sequence profiles from protein structures using deep neural networks

    James O’Connell, Zhixiu Li, Jack Hanson, Rhys Heffernan, James Lyons, Kuldip Paliwal, Abdollah Dehzangi, Yuedong Yang, and Yaoqi Zhou. Spin2: Predicting sequence profiles from protein structures using deep neural networks. Proteins: Structure, Function, and Bioinformatics, 86(...

    2018

  12. [20]

    Learning transferable visual models from natural language supervision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervision. In International conference on machine learning , p...

    2021

  13. [21]

    Hierarchical text-conditional image generation with clip latents

    Aditya Ramesh, Prafulla Dhariwal, Alex Nichol, Casey Chu, and Mark Chen. Hierarchical text-conditional image generation with clip latents. arXiv preprint arXiv:2204.06125 , 2022

    2022 arXiv

  14. [22]

    Transformer protein language models are unsupervised structure learners

    Roshan Rao, Joshua Meier, Tom Sercu, Sergey Ovchinnikov, and Alexander Rives. Transformer protein language models are unsupervised structure learners. Biorxiv, 2020

    2020

  15. [23]

    Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences

    Alexander Rives, Joshua Meier, Tom Sercu, Siddharth Goyal, Zeming Lin, Jason Liu, Demi Guo, Myle Ott, C Lawrence Zitnick, Jerry Ma, et al. Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences. Proceedings of the National ...

    2021

  16. [24]

    Generative de novo protein design with global context

    Cheng Tan, Zhangyang Gao, Jun Xia, and Stan Z Li. Generative de novo protein design with global context. arXiv preprint arXiv:2204.10673 , 2022

    2022 arXiv

  17. [25]

    A document-level neural machine translation model with dynamic caching guided by theme-rheme information

    Yiqi Tong, Jiangbin Zheng, Hongkang Zhu, Yidong Chen, and Xiaodong Shi. A document-level neural machine translation model with dynamic caching guided by theme-rheme information. In Proceedings of the 28th International Conference on Computational Linguistics , pages 4385– 4395, 2020

    2020

  18. [26]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need. Advances in neural information processing systems , 30, 2017

    2017

  19. [27]

    Fuzzy slic: Fuzzy simple linear iterative clustering

    Chong Wu, Jiangbin Zheng, Zhenan Feng, Houwang Zhang, Le Zhang, Jiawang Cao, and Hong Yan. Fuzzy slic: Fuzzy simple linear iterative clustering. IEEE Transactions on Circuits and Systems for Video Technology, 31(6):2114–2124, 2020

    2020

  20. [28]

    Discognn: A sample-efficient framework for self-supervised graph representation learning

    Jun Xia, Shaorong Chen, Yue Liu, Zhangyang Gao, Jiangbin Zheng, Xihong Yang, and Stan Z Li. Discognn: A sample-efficient framework for self-supervised graph representation learning. In 2024 IEEE 40th International Conference on Data Engineering (ICDE) , pages 2876–

    2024

  21. [29]

    Understand- ing the limitations of deep models for molecular property prediction: Insights and solutions

    Jun Xia, Lecheng Zhang, Xiao Zhu, Yue Liu, Zhangyang Gao, Bozhen Hu, Cheng Tan, Jiangbin Zheng, Siyuan Li, and Stan Z Li. Understand- ing the limitations of deep models for molecular property prediction: Insights and solutions. Advances in Neural Information Processing Systems...

    2024

  22. [30]

    Enhancing neural sign language translation by highlighting the facial expression information

    Jiangbin Zheng, Yidong Chen, Chong Wu, Xiaodong Shi, and Suhail Muhammad Kamal. Enhancing neural sign language translation by highlighting the facial expression information. Neurocomputing, 464:462–472, 2021

    2021

  23. [31]

    Mmdesign: Multi- modality transfer learning for generative protein design

    Jiangbin Zheng, Siyuan Li, Yufei Huang, Zhangyang Gao, Cheng Tan, Bozhen Hu, Jun Xia, Ge Wang, and Stan Z Li. Mmdesign: Multi- modality transfer learning for generative protein design. arXiv preprint arXiv:2312.06297, 2023

    2023 arXiv

  24. [32]

    Leveraging graph-based cross-modal information fusion for neural sign language translation

    Jiangbin Zheng, Siyuan Li, Cheng Tan, Chong Wu, Yidong Chen, and Stan Z Li. Leveraging graph-based cross-modal information fusion for neural sign language translation. arXiv preprint arXiv:2211.00526, 2022

    2022 arXiv

  25. [33]

    Progressive multi-modality learning for inverse protein folding

    Jiangbin Zheng and Stan Z Li. Progressive multi-modality learning for inverse protein folding. In 2024 IEEE International Conference on Multimedia and Expo (ICME) , pages 1–6. IEEE, 2024

    2024

  26. [34]

    Lightweight contrastive protein structure-sequence transformation

    Jiangbin Zheng, Ge Wang, Yufei Huang, Bozhen Hu, Siyuan Li, Cheng Tan, Xinwen Fan, and Stan Z Li. Lightweight contrastive protein structure-sequence transformation. arXiv preprint arXiv:2303.11783 , 2023

    2023 arXiv

  27. [35]

    Cvt-slr: Contrastive visual-textual transformation for sign language recognition with variational alignment

    Jiangbin Zheng, Yile Wang, Cheng Tan, Siyuan Li, Ge Wang, Jun Xia, Yidong Chen, and Stan Z Li. Cvt-slr: Contrastive visual-textual transformation for sign language recognition with variational alignment. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern ...

    2023

  28. [36]

    Using context-to-vector with graph retrofitting to improve word embeddings

    Jiangbin Zheng, Yile Wang, Ge Wang, Jun Xia, Yufei Huang, Guojiang Zhao, Yue Zhang, and Stan Li. Using context-to-vector with graph retrofitting to improve word embeddings. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Vol- ume 1: ...

    2022

  29. [37]

    Metaenzyme: Meta pan-enzyme learning for task-adaptive redesign

    Jiangbin Zheng, Han Zhang, Qianqing Xu, An-Ping Zeng, and Stan Z Li. Metaenzyme: Meta pan-enzyme learning for task-adaptive redesign. In Proceedings of the 32nd ACM International Conference on Multimedia , pages 1721–1730, 2024

    2024

  30. [38]

    An improved sign language translation model with explainable adaptations for processing long sign sentences

    Jiangbin Zheng, Zheng Zhao, Min Chen, Jing Chen, Chong Wu, Yidong Chen, Xiaodong Shi, and Yiqi Tong. An improved sign language translation model with explainable adaptations for processing long sign sentences. Computational Intelligence and Neuroscience , 2020, 2020

    2020

  31. [39]

    C2slr: Consistency-enhanced continuous sign language recognition

    Ronglai Zuo and Brian Mak. C2slr: Consistency-enhanced continuous sign language recognition. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 5131–5140, 2022

    2022

Pith tools

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