Pith. sign in

REVIEW 4 major objections 5 minor 64 references

Curriculum Learning for Biological Sequence Prediction: The Case of De Novo Peptide Sequencing

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read RefineNovo claims a curriculum-plus-refinement scheme makes non-autoregressive peptide sequencing the best on two nine-species benchmarks.

desk verdict RefineNovo is a plausible incremental improvement over PrimeNovo, but the state-of-the-art claim outruns the evidence: a key baseline is missing, there are no error bars, and the masking schedule is ambiguous. read the letter →

arxiv 2506.13485 v1 pith:2RHPDVSZ submitted 2025-06-16 q-bio.BM cs.LG

Xiang Zhang , Jiaqi Wei , Zijie Qiu , Sheng Xu , Nanqing Dong , Zhiqiang Gao , Siqi Sun This is my paper · ORCID
classification q-bio.BMcs.LG
keywords denovopeptidesequencingnon-autoregressivetransformercurriculumlearningCTClossiterativerefinementmassspectrometrydifficultyannealingproteinsequenceprediction
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 argues that the main obstacle to non-autoregressive (NAT) peptide sequencing is not the architecture but the CTC training objective, which is unstable and frequently fails to converge. It introduces RefineNovo, which combines a structured protein-sequence curriculum with an inference-time self-refinement loop. The curriculum samples the model's own most probable CTC alignment of the true peptide, reveals those tokens through a dynamically annealed mask, and thereby eases the model from nearly supervised prediction to full de novo decoding. On the 9-species-V1 and 9-species-V2 benchmarks the paper reports average amino acid precision of 0.800 and 0.907 and peptide recall of 0.653 and 0.790, surpassing all previous methods, and a reduction of NAT training failures by more than 90 percent.

What carries the argument

The load-bearing object is the CTC oracle-path curriculum mask. Given the target peptide $A$ and a CTC path $y'$ sampled as the most probable valid alignment of $A$, the model computes an oracle token at each position by taking the label at the midpoint of $y'$ and blanking where the path parity requires; the fraction of positions masked is $\rho=\alpha(1-\text{acc}(A,y_{\text{argmax}}))$, so early training shows many true tokens and later training masks more. This converts the NAT objective from independent prediction $P(A|I)$ into conditioned prediction $P(A|\rho(A,y'),I)$, shrinking the effective search space, and it is what makes the embedding layer able to consume arbitrary CTC paths. At inference the same embedding layer is fed the argmax path repeatedly, and the precise-mass-control knapsack decoder enforces the precursor-mass constraint.

What would settle it

Run RefineNovo at inference with the first pass replaced by a random or deliberately corrupted CTC path and measure whether subsequent refinement degrades; also train a control that conditions on the model's own argmax path during training instead of the oracle path. If the control matches or beats RefineNovo, the oracle-path curriculum is not the source of the improvement, and if corrupted conditioning propagates errors, the refinement transfer assumption is false.

Watch

Extended reading notes

Core claim

The central claim is that a CTC-trained NAT decoder becomes both more stable to train and more accurate at inference if it is conditioned, during training, on a partially revealed oracle CTC path rather than on an empty decoder input. The method computes the most probable valid CTC alignment of the ground-truth peptide, extracts the amino-acid tokens at the midpoint of that alignment, blanks out positions according to a mask ratio that anneals with the model's own argmax accuracy, and feeds the resulting masked sequence into the embedding layer. The same embedding layer is then reused at inference: the argmax-decoded CTC path from the previous pass is embedded and fed back, yielding $y^{(i)}=\arg\max P(\cdot|I,\text{EmbeddingLayer}(y^{(i-1)}))$, repeated a few times before the precise-mass-control dynamic program decodes the final sequence. The paper reports that this combination achieves state-of-the-art results on both benchmark versions across most of the nine species, with ablation experiments attributing the gain to the interaction of curriculum masking, difficulty annealing, and iterative refinement.

Load-bearing premise

The method trains the decoder on partially revealed oracle CTC paths, but at inference it feeds back the model's own greedy argmax path; the whole accuracy gain rests on the assumption that the model learns to condition on arbitrary CTC paths as well as it conditions on correct ones.

Editorial extensions

If this is right

  • On the 9-species-V1 benchmark, RefineNovo leads in amino acid precision and peptide recall for eight of nine species (averages 0.800 and 0.653), making it the direct reference point for future de novo sequencing work.
  • On the stricter 9-species-V2 benchmark the model reports the best amino acid precision for all nine species and peptide recall for eight, with averages 0.907 and 0.790.
  • The curriculum reduces NAT training failures by more than 90 percent: in 20 sampled MassIVE-KB training runs, PrimeNovo failed 18 times while RefineNovo failed once.
  • Iterative refinement saturates quickly: peptide recall rises from 0.728 at one pass to 0.736 at three passes and then plateaus, so the accuracy gain can be obtained at three decoding iterations.
  • The ablation shows fixed masking at ratio 0.7 hurts (0.733 amino acid precision) whereas dynamic annealing plus curriculum plus refinement reaches 0.800, so the adaptive difficulty schedule is doing real work.

Reading between the lines

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

  • The refinement module actually recycles the CTC path rather than the reduced peptide, which suggests the method is learning to search alignment space; a direct comparison against masked-token diffusion refinement of the peptide sequence would isolate whether alignment-level recycling is the mechanism.
  • Because the curriculum uses the model's own argmax accuracy as the difficulty meter, the same recipe should transfer to other CTC-trained NAT tasks such as speech recognition or handwriting recognition, where loss instability is also reported.
  • The NovoBench domain-shift result (0.09 peptide score on 7-species yeast for both PrimeNovo and RefineNovo) implies the gains concentrate in the training distribution; a testable extension is to retrain on mixed-instrument data and check whether the oracle-path curriculum closes the distribution gap.
  • The embedding layer trained on oracle CTC paths accepts arbitrary pseudo-labels, so one could feed unlabeled spectra decoded by the model itself and fine-tune semi-supervised, using the curriculum mask as a denoising objective.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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. This paper proposes RefineNovo, a non-autoregressive Transformer (NAT) for de novo peptide sequencing built on the authors' earlier PrimeNovo backbone. The two claimed contributions are (i) a CTC-path-based curriculum learning strategy with adaptive difficulty annealing, and (ii) an iterative self-refinement module used at inference. The model is trained on MassIVE-KB and evaluated on the 9-species-V1 and 9-species-V2 benchmarks, reporting average amino acid precision of 0.800 and 0.907 and peptide recall of 0.653 and 0.790, respectively, and claiming to outperform all previous methods. The paper also reports a more than 90% reduction in NAT training failures based on 20 randomly selected training subsets.

Significance. If the reported numbers hold, RefineNovo would set a new state of the art for NAT-based de novo peptide sequencing, and the curriculum strategy would be a practically valuable contribution to stabilizing CTC training in parallel sequence models. The paper has real strengths: it evaluates against external public benchmarks and multiple baseline families, provides an ablation of the three proposed components, reports a training-stability case study, and releases code and models. Those strengths are offset by the fact that the headline gains over PrimeNovo are small (0.012 AA precision on V1, 0.016 on V2), every reported number is a single point estimate with no error bars, and the iterative refinement module is used at inference under a different input distribution from the one it was trained on. The central empirical claim is therefore plausible but not yet established at the standard expected for a benchmark-leading result.

major comments (4)
  1. [3.3, Algorithm 1, Appendix A] The curriculum schedule is internally inconsistent and cannot be reproduced as written. In Algorithm 1, the variable named 'acc' is computed as acc = 1 − same_num/seq_lens, which is an error rate, and then ρ_ratio is set to α(1−acc), so ρ_ratio equals α times the token matching rate; that gives low masking early and higher masking later, consistent with the 'easy-to-hard' prose description. The prose, however, defines acc as the model's prediction accuracy and states ρ_ratio = α(1−acc), which would give a hard-to-easy schedule. The Appendix A code further complicates matters: it sets keep_prob = ((seq_lens − same_num)/seq_lens)·peek_factor and masks with probability 1−keep_prob. This equals α·matching_rate only when α=1; for any other α, the code's mask probability is 1−α(1−matching_rate), not α·matching_rate. The value of α is never reported. Please resolve the naming conflict, state the intended mask probability explicitly, and report the value of α used in the experiments.
  2. [3.3, 3.4, Table 5] The iterative refinement module trains on oracle CTC paths constructed from ground-truth target tokens (Algorithm 1 lines 7–9 and Appendix A: oracle = tgt_tokens.gather(-1, oracle_pos)), while at inference Eq. (5) feeds the model's own argmax path y^(i−1) into the same embedding layer. The decoder is therefore never trained on model-generated paths, so the self-refinement procedure is an out-of-distribution use of the embedding layer whenever the argmax path deviates from the oracle alignment. Table 5 reports only the iteration curve (AA precision 0.848 to 0.854, peptide recall 0.728 to 0.736) with no error bars and no comparison to a same-checkpoint single-pass baseline; the ablation in Table 4 attributes only +0.009 AA precision to the refinement component. A controlled comparison is needed between refinement from (a) the oracle path, (b) the model's argmax path, and (c) no refinement, with multiple seeds and with the agreement rate between argmax paths and oracle alignments reported.
  3. [Tables 1, 2, 4, 5] All benchmark results are single point estimates with no standard deviations, seeds, or significance tests. The headline differences over PrimeNovo are small: 0.800 vs 0.788 AA precision and 0.653 vs 0.638 peptide recall on V1, and 0.907 vs 0.891 AA precision and 0.790 vs 0.777 peptide recall on V2. The ablation differences are of the same small magnitude. Without variance estimates, per-species paired tests, or at least multiple runs, the abstract's claim that RefineNovo 'outperforms all previous methods' is not statistically supported. Please report mean and standard deviation over at least three seeds and, where possible, per-species paired comparisons against PrimeNovo.
  4. [4.2, Training Success Rate and Case Study] The claim of a more than 90% reduction in training failures rests on 20 random subsets from MassIVE-KB, with PrimeNovo failing in 18 runs and RefineNovo failing once. The manuscript does not define the failure criterion, describe how the subsets were drawn, or report the random seeds; moreover, Figure 4 states that the case-study plots come from different datasets, so the displayed curves are not a same-data comparison. This claim should be made reproducible by specifying the failure criterion (e.g., loss NaN, validation collapse, divergence threshold) and by reporting the same-dataset comparisons for both models.
minor comments (5)
  1. [3.2, Eq. (1)] The sinusoidal encoding formula is malformed; the expression as printed is missing operators and parentheses and cannot be parsed. Please rewrite it in a standard form.
  2. [3.3, Eq. (3)] Equation (3) writes L = P(A|ρ(A,y),I), but the quantity being optimized should be a loss, i.e., the negative log probability of the target under the conditional input, not the probability itself.
  3. [Appendix B, Table 5] The caption of Table 5 says 'Effect of different beam sizes,' but the rows report the number of iterative refinement steps. The caption should be corrected.
  4. [3.3 and Figure 3] The main text says candidate CTC paths are re-ranked during training to select the oracle path, while Figure 3's inference phase mentions 'Argmax CTC Path Sampling' without describing whether reranking is used at inference. Please clarify the inference-time path selection procedure.
  5. [4.1, Model Details] The text states an initial learning rate of 4e-4 that is 'gradually increased to the target peak within the first epoch,' which is ambiguous about whether the initial value or the peak is 4e-4; please specify the warm-up schedule.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the headline accuracy claims are measured against external benchmark datasets and baselines.

full rationale

No load-bearing step reduces to its inputs. The headline accuracies (AA precision 0.800 on 9-species-V1 and 0.907 on 9-species-V2; peptide recall 0.653 and 0.790) are measured on public benchmark datasets against external baselines (Peaks, DeepNovo, PointNovo, Casanovo, AdaNovo, CasanovoV2, PrimeNovo), so the central claim is externally falsifiable rather than defined into existence. The curriculum module and iterative refinement are training/inference procedures whose outputs are evaluated against ground-truth peptides; the oracle-token leakage in Algorithm 1 is a training target, not the evaluation metric. The authors reuse their own PrimeNovo backbone, CTC recipe, and PMC unit, and cite that prior work, but PrimeNovo is a published, independently benchmarked system and the comparisons are made on the same public test sets, so this is normal self-citation rather than circularity. The train-inference gap (oracle CTC path during training versus argmax path during inference, Eq. 5 and Algorithm 2) is a potential validity weakness, but it is not circular: the refinement output is still a model prediction scored against ground truth. The Figure 4 note that validation does not involve looking at true tokens is a transparency statement, not a circular reduction. No step was found in which the claimed result is equivalent to its input by construction.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The central empirical claim does not rest on a new mathematical derivation. It depends on inherited domain assumptions plus an unreported peeking factor alpha and an ambiguous masking schedule. No new physical entities are introduced.

free parameters (2)
  • peek factor alpha = not reported
    Controls the curriculum masking ratio via rho_ratio = alpha * (1 - acc). The value is not given in the paper, yet the ablation shows that a fixed mask ratio of 0.7 degrades performance, so the schedule is load-bearing.
  • number of iterative refinement steps N = 3
    Selected from Table 5 after performance plateaus at three iterations; affects inference cost and final accuracy.
assumptions (5)
  • ad hoc to paper The most probable CTC path under the model, conditioned on the true target, is a useful oracle for curriculum masking.
    Algorithm 1 selects oracle positions from best_aligns and builds the masked input from those tokens. The whole curriculum rests on this oracle being a stable training signal.
  • ad hoc to paper The masking ratio formula rho_ratio = alpha * (1 - acc) produces an easy-to-hard difficulty schedule.
    The formula is ambiguous because 'acc' is called accuracy in the text but computed as an error rate in Algorithm 1. No derivation or sensitivity analysis is provided for the schedule.
  • domain assumption De novo peptide sequencing can be modeled as a sequence-to-sequence task where CTC alignment paths over blank tokens are the correct latent variable.
    Standard assumption in CTC-based NAT peptide models, inherited from PrimeNovo and prior NAT translation work.
  • domain assumption Precursor mass is a hard constraint satisfiable by knapsack-style dynamic programming with fixed amino acid masses.
    The PMC unit assumes each amino acid has a known mass and that the precursor mass bounds the total peptide mass; this is invoked at decoding in Appendix D.
  • domain assumption Models pretrained on MassIVE-KB transfer to the 9-species V1 and V2 benchmark distributions.
    The model is trained on MassIVE-KB and evaluated on the 9-species datasets without per-dataset retraining; the paper does not quantify the distribution shift between training and test data.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Curriculum Learning for Biological Sequence Prediction: The Case of De Novo Peptide Sequencing." pith.science (2026). https://pith.science/paper/2RHPDVSZ

@misc{pith2026250613485,
  author       = {Pith},
  title        = {Pith review of: Curriculum Learning for Biological Sequence Prediction: The Case of De Novo Peptide Sequencing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2RHPDVSZ}},
  note         = {Machine review of arXiv:2506.13485}
}
read the original abstract

Peptide sequencing-the process of identifying amino acid sequences from mass spectrometry data-is a fundamental task in proteomics. Non-Autoregressive Transformers (NATs) have proven highly effective for this task, outperforming traditional methods. Unlike autoregressive models, which generate tokens sequentially, NATs predict all positions simultaneously, leveraging bidirectional context through unmasked self-attention. However, existing NAT approaches often rely on Connectionist Temporal Classification (CTC) loss, which presents significant optimization challenges due to CTC's complexity and increases the risk of training failures. To address these issues, we propose an improved non-autoregressive peptide sequencing model that incorporates a structured protein sequence curriculum learning strategy. This approach adjusts protein's learning difficulty based on the model's estimated protein generational capabilities through a sampling process, progressively learning peptide generation from simple to complex sequences. Additionally, we introduce a self-refining inference-time module that iteratively enhances predictions using learned NAT token embeddings, improving sequence accuracy at a fine-grained level. Our curriculum learning strategy reduces NAT training failures frequency by more than 90% based on sampled training over various data distributions. Evaluations on nine benchmark species demonstrate that our approach outperforms all previous methods across multiple metrics and species.

Figures

Figures reproduced from arXiv: 2506.13485 by the authors.

Figure 1
Figure 1. Comparison between (a) Vanilla NAT and (b) our pro￾posed NAT peptide sequencing architectures, which integrate cur￾riculum learning and a self-refining module. 1. Introduction Peptide sequencing via tandem mass spectrometry plays a pivotal role in proteomics research, with significant impli￾cations for fundamental and applied studies in chemistry, biology, medicine, and pharmacology (Aebersold & Mann, 2003; Ng et al… view at source ↗
Figure 2
Figure 2. Overview of the protein identification workflow. Protein samples are digested into peptides, which are then analyzed using mass spectrometry to generate MS/MS spectra. These spectra are subsequently used in de novo peptide sequencing to determine the peptide sequences. predicted sequences as input, models can iteratively refine their predictions, improving generation quality. However, traditional NAT models (Zhang e… view at source ↗
Figure 3
Figure 3. The architecture of RefineNovo. (A) The training phase of RefineNovo begins with a Spectrum Encoder that processes the input spectra. Following encoding, the Non-Autoregressive Peptide Decoder leverages the encoded spectra and blank tokens to generate a probability matrix. To predict potential sequences, CTC path sampling identifies all candidate paths, which are then re-ranked for optimal selection. The model emplo… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: A case study showing three types of training failures that frequently happened during the training of NAT peptide sequencing models. RefineNovo can successfully mitigate training problems in traditional NAT models. Note that the plots show training on different dataset…

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

64 extracted references · 26 canonical work pages

  1. [1]

    J., Bambrick, J., et al

    Abramson, J., Adler, J., Dunger, J., Evans, R., Green, T., Pritzel, A., Ronneberger, O., Willmore, L., Ballard, A. J., Bambrick, J., et al. Accurate structure prediction of biomolecular interactions with alphafold 3. Nature, pp.\ 1--3, 2024

    2024

  2. [2]

    and Mann, M

    Aebersold, R. and Mann, M. Mass spectrometry-based proteomics. Nature, 422 0 (6928): 0 198--207, 2003

    2003

  3. [3]

    Non-Autoregressive Translation by Learning Target Categorical Codes

    Bao, Y., Huang, S., Xiao, T., Wang, D., Dai, X., and Chen, J. Non-autoregressive translation by learning target categorical codes. arXiv preprint arXiv:2103.11405, 2021

    work page Pith review arXiv 2021

  4. [4]

    J., and Cheng, J

    Chen, C., Hou, J., Tanner, J. J., and Cheng, J. Bioinformatics methods for mass spectrometry-based proteomics data analysis. International journal of molecular sciences, 21 0 (8): 0 2873, 2020

    2020

  5. [5]

    Context-Aware Cross-Attention for Non-Autoregressive Translation

    Ding, L., Wang, L., Wu, D., Tao, D., and Tu, Z. Context-aware cross-attention for non-autoregressive translation. arXiv preprint arXiv:2011.00770, 2020

    work page Pith review arXiv 2011

  6. [6]

    Progressive Multi-Granularity Training for Non-Autoregressive Translation

    Ding, L., Wang, L., Liu, X., Wong, D. F., Tao, D., and Tu, Z. Progressive multi-granularity training for non-autoregressive translation. arXiv preprint arXiv:2106.05546, 2021 a

    work page Pith review arXiv 2021

  7. [7]

    Rejuvenating Low-Frequency Words: Making the Most of Parallel Data in Non-Autoregressive Translation

    Ding, L., Wang, L., Liu, X., Wong, D. F., Tao, D., and Tu, Z. Rejuvenating low-frequency words: Making the most of parallel data in non-autoregressive translation. arXiv preprint arXiv:2106.00903, 2021 b

    work page Pith review arXiv 2021

  8. [8]

    B., Williams, W., Beljouw, S

    Eloff, K., Kalogeropoulos, K., Morell, O., Mabona, A., Jespersen, J. B., Williams, W., Beljouw, S. P. v., Skwark, M., Laustsen, A. H., Brouns, S. J., et al. De novo peptide sequencing with instanovo: Accurate, database-free peptide identification for large scale proteomics experiments. bioRxiv, pp.\ 2023--08, 2023

    work page 2023

Show all 64 references
  1. [9]

    Deep multimodal networks for m-type star classification with paired spectrum and photometric image

    Gao, J., Chen, J., Wei, J., Jiang, B., and Luo, A.-L. Deep multimodal networks for m-type star classification with paired spectrum and photometric image. Publications of the Astronomical Society of the Pacific, 135 0 (1046): 0 044503, 2023

    2023

  2. [10]

    Mask-predict: Parallel decoding of conditional masked language models

    Ghazvininejad, M., Levy, O., Liu, Y., and Zettlemoyer, L. Mask-predict: Parallel decoding of conditional masked language models. arXiv preprint arXiv:1904.09324, 2019

    1904 arXiv

  3. [11]

    Aligned cross entropy for non-autoregressive machine translation

    Ghazvininejad, M., Karpukhin, V., Zettlemoyer, L., and Levy, O. Aligned cross entropy for non-autoregressive machine translation. In International Conference on Machine Learning, pp.\ 3515--3523. PMLR, 2020

    2020

  4. [12]

    and Jaitly, N

    Graves, A. and Jaitly, N. Towards end-to-end speech recognition with recurrent neural networks. In International conference on machine learning, pp.\ 1764--1772. PMLR, 2014

    2014

  5. [13]

    Connectionist temporal classification: labelling unsegmented sequence data with recurrent neural networks

    Graves, A., Fern \'a ndez, S., Gomez, F., and Schmidhuber, J. Connectionist temporal classification: labelling unsegmented sequence data with recurrent neural networks. In Proceedings of the 23rd international conference on Machine learning, pp.\ 369--376, 2006

    2006

  6. [14]

    O., and Socher, R

    Gu, J., Bradbury, J., Xiong, C., Li, V. O., and Socher, R. Non-autoregressive neural machine translation. arXiv preprint arXiv:1711.02281, 2017

    2017 arXiv

  7. [15]

    Non-autoregressive neural machine translation with enhanced decoder input

    Guo, J., Tan, X., He, D., Qin, T., Xu, L., and Liu, T.-Y. Non-autoregressive neural machine translation with enhanced decoder input. In Proceedings of the AAAI conference on artificial intelligence, volume 33, pp.\ 3723--3730, 2019

    2019

  8. [16]

    Jointly masked sequence-to-sequence model for non-autoregressive neural machine translation

    Guo, J., Xu, L., and Chen, E. Jointly masked sequence-to-sequence model for non-autoregressive neural machine translation. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pp.\ 376--385, 2020

    2020

  9. [17]

    J., Oktay, D., Lin, Z., Verkuil, R., Tran, V

    Hayes, T., Rao, R., Akin, H., Sofroniew, N. J., Oktay, D., Lin, Z., Verkuil, R., Tran, V. Q., Deaton, J., Wiggert, M., et al. Simulating 500 million years of evolution with a language model. bioRxiv, pp.\ 2024--07, 2024

    2024

  10. [18]

    R., Mou, L., and Li, L

    Huang, C., Zhou, H., Za \" ane, O. R., Mou, L., and Li, L. Non-autoregressive translation with layer-wise prediction and deep supervision. In Proceedings of the AAAI conference on artificial intelligence, pp.\ 10776--10784, 2022 a

    2022

  11. [19]

    Directed acyclic transformer for non-autoregressive machine translation

    Huang, F., Zhou, H., Liu, Y., Li, H., and Huang, M. Directed acyclic transformer for non-autoregressive machine translation. In International Conference on Machine Learning, pp.\ 9410--9428. PMLR, 2022 b

    2022

  12. [20]

    S., Perez, F., and Volkovs, M

    Huang, X. S., Perez, F., and Volkovs, M. Improving non-autoregressive translation models without distillation. In International Conference on Learning Representations, 2022 c

    2022

  13. [21]

    Contranovo: a contrastive learning approach to enhance de novo peptide sequencing

    Jin, Z., Xu, S., Zhang, X., Ling, T., Dong, N., Ouyang, W., Gao, Z., Chang, C., and Sun, S. Contranovo: a contrastive learning approach to enhance de novo peptide sequencing. arxiv. arXiv preprint arXiv:2312.11584, 2023

    2023 arXiv

  14. [22]

    Contranovo: A contrastive learning approach to enhance de novo peptide sequencing

    Jin, Z., Xu, S., Zhang, X., Ling, T., Dong, N., Ouyang, W., Gao, Z., Chang, C., and Sun, S. Contranovo: A contrastive learning approach to enhance de novo peptide sequencing. In Proceedings of the AAAI Conference on Artificial Intelligence, pp.\ 144--152, 2024

    2024

  15. [23]

    Highly accurate protein structure prediction with alphafold

    Jumper, J., Evans, R., Pritzel, A., Green, T., Figurnov, M., Ronneberger, O., Tunyasuvunakool, K., Bates, R., Z \' dek, A., Potapenko, A., et al. Highly accurate protein structure prediction with alphafold. Nature, 596 0 (7873): 0 583--589, 2021

    2021

  16. [24]

    W., Chuangsuwanich, E., and Sriswasdi, S

    Karunratanakul, K., Tang, H.-Y., Speicher, D. W., Chuangsuwanich, E., and Sriswasdi, S. Uncovering Thousands of New Peptides with Sequence-Mask-Search Hybrid De Novo Peptide Sequencing Framework. Molecular & cellular proteomics : MCP, 18 0 (12): 0 2478--2491, dec 2019. ISSN 15...

    2019

  17. [25]

    Kingma, D. P. and Ba, J. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2014

    2014 arXiv

  18. [26]

    Deep learning

    LeCun, Y., Bengio, Y., and Hinton, G. Deep learning. nature, 521 0 (7553): 0 436--444, 2015

    2015

  19. [27]

    Evolutionary-scale prediction of atomic-level protein structure with a language model

    Lin, Z., Akin, H., Rao, R., Hie, B., Zhu, Z., Lu, W., Smetanin, N., Verkuil, R., Kabeli, O., Shmueli, Y., et al. Evolutionary-scale prediction of atomic-level protein structure with a language model. Science, 379 0 (6637): 0 1123--1130, 2023

    2023

  20. [28]

    Accurate de novo peptide sequencing using fully convolutional neural networks

    Liu, K., Ye, Y., Li, S., and Tang, H. Accurate de novo peptide sequencing using fully convolutional neural networks. Nature Communications, 14 0 (1): 0 7974, 2023

    2023

  21. [29]

    A character-level length-control algorithm for non-autoregressive sentence summarization

    Liu, P., Zhang, X., and Mou, L. A character-level length-control algorithm for non-autoregressive sentence summarization. Advances in Neural Information Processing Systems, 35: 0 29101--29112, 2022

    2022

  22. [30]

    Peaks: powerful software for peptide de novo sequencing by tandem mass spectrometry

    Ma, B., Zhang, K., Hendrie, C., Liang, C., Li, M., Doherty-Kirby, A., and Lajoie, G. Peaks: powerful software for peptide de novo sequencing by tandem mass spectrometry. Rapid communications in mass spectrometry, 17 0 (20): 0 2337--2342, 2003

    2003

  23. [31]

    Flowseq: Non-autoregressive conditional sequence generation with generative flow

    Ma, X., Zhou, C., Li, X., Neubig, G., and Hovy, E. Flowseq: Non-autoregressive conditional sequence generation with generative flow. arXiv preprint arXiv:1909.02480, 2019

    1909 arXiv

  24. [32]

    Mitigating the missing-fragmentation problem in de novo peptide sequencing with a two-stage graph-based deep learning model

    Mao, Z., Zhang, R., Xin, L., and Li, M. Mitigating the missing-fragmentation problem in de novo peptide sequencing with a two-stage graph-based deep learning model. Nature Machine Intelligence, 5 0 (11): 0 1250--1260, 2023

    2023

  25. [33]

    Muth, T., Hartkopf, F., Vaudel, M., and Renard, B. Y. A potential golden age to come—current tools, recent use cases, and future avenues for de novo sequencing in proteomics. Proteomics, 18 0 (18): 0 1700150, 2018

    2018

  26. [34]

    Ng, C. C. A., Zhou, Y., and Yao, Z.-P. Algorithms for de-novo sequencing of peptides by tandem mass spectrometry: A review. Analytica Chimica Acta, pp.\ 341330, 2023

    2023

  27. [35]

    Glancing transformer for non-autoregressive neural machine translation

    Qian, L., Zhou, H., Bao, Y., Wang, M., Qiu, L., Zhang, W., Yu, Y., and Li, L. Glancing transformer for non-autoregressive neural machine translation. arXiv preprint arXiv:2008.07905, 2020

    2008 arXiv

  28. [36]

    H., Xin, L., Chen, X., Li, M., Shan, B., and Ghodsi, A

    Qiao, R., Tran, N. H., Xin, L., Chen, X., Li, M., Shan, B., and Ghodsi, A. Computationally instrument-resolution-independent de novo peptide sequencing for high-resolution devices. Nature Machine Intelligence, 3 0 (5): 0 420--425, 2021

    2021

  29. [37]

    Universal biological sequence reranking for improved de novo peptide sequencing

    Qiu, Z., Wei, J., Zhang, X., Xu, S., Zou, K., Jin, Z., Gao, Z., Dong, N., and Sun, S. Universal biological sequence reranking for improved de novo peptide sequencing. arXiv preprint arXiv:2505.17552, 2025

    2025 arXiv

  30. [38]

    Non-autoregressive machine translation with latent alignments

    Saharia, C., Chan, W., Saxena, S., and Norouzi, M. Non-autoregressive machine translation with latent alignments. arXiv preprint arXiv:2004.07437, 2020

    2004 arXiv

  31. [39]

    Simple and effective masked diffusion language models

    Sahoo, S., Arriola, M., Schiff, Y., Gokaslan, A., Marroquin, E., Chiu, J., Rush, A., and Kuleshov, V. Simple and effective masked diffusion language models. Advances in Neural Information Processing Systems, 37: 0 130136--130184, 2024

    2024

  32. [40]

    Savinov, N., Chung, J., Binkowski, M., Elsen, E., and Oord, A. v. d. Step-unrolled denoising autoencoders for text generation. arXiv preprint arXiv:2112.06749, 2021

    2021 arXiv

  33. [41]

    Minimizing the bag-of-ngrams difference for non-autoregressive neural machine translation

    Shao, C., Zhang, J., Feng, Y., Meng, F., and Zhou, J. Minimizing the bag-of-ngrams difference for non-autoregressive neural machine translation. In Proceedings of the AAAI conference on artificial intelligence, volume 34, pp.\ 198--205, 2020

    2020

  34. [42]

    One reference is not enough: Diverse distillation with reference selection for non-autoregressive translation

    Shao, C., Wu, X., and Feng, Y. One reference is not enough: Diverse distillation with reference selection for non-autoregressive translation. arXiv preprint arXiv:2205.14333, 2022

    2022 arXiv

  35. [43]

    Latent-variable non-autoregressive neural machine translation with deterministic inference using a delta posterior

    Shu, R., Lee, J., Nakayama, H., and Cho, K. Latent-variable non-autoregressive neural machine translation with deterministic inference using a delta posterior. In Proceedings of the aaai conference on artificial intelligence, volume 34, pp.\ 8846--8853, 2020

    2020

  36. [44]

    Alignart: Non-autoregressive neural machine translation by jointly learning to estimate alignment and translate

    Song, J., Kim, S., and Yoon, S. Alignart: Non-autoregressive neural machine translation by jointly learning to estimate alignment and translate. arXiv preprint arXiv:2109.06481, 2021

    2021 arXiv

  37. [45]

    Insertion transformer: Flexible sequence generation via insertion operations

    Stern, M., Chan, W., Kiros, J., and Uszkoreit, J. Insertion transformer: Flexible sequence generation via insertion operations. In International Conference on Machine Learning, pp.\ 5976--5985. PMLR, 2019

    2019

  38. [46]

    H., Zhang, X., Xin, L., Shan, B., and Li, M

    Tran, N. H., Zhang, X., Xin, L., Shan, B., and Li, M. De novo peptide sequencing by deep learning. Proceedings of the National Academy of Sciences, 114 0 (31): 0 8247--8252, 2017

    2017

  39. [47]

    S., Cha, S

    Wang, M., Wang, J., Carver, J., Pullman, B. S., Cha, S. W., and Bandeira, N. Assembling the community-scale discoverable human proteome. Cell systems, 7 0 (4): 0 412--421, 2018

    2018

  40. [48]

    Non-autoregressive machine translation with auxiliary regularization

    Wang, Y., Tian, F., He, D., Qin, T., Zhai, C., and Liu, T.-Y. Non-autoregressive machine translation with auxiliary regularization. In Proceedings of the AAAI conference on artificial intelligence, volume 33, pp.\ 5377--5384, 2019

    2019

  41. [50]

    Xia, J., Chen, S., Zhou, J., Ling, T., Du, W., Liu, S., and Li, S. Z. Adanovo: Adaptive De Novo peptide sequencing with conditional mutual information, 2024 b . URL https://arxiv.org/abs/2403.07013

    2024 arXiv

  42. [51]

    Xia, J., Liu, S., Zhou, J., Chen, S., Xiang, H., Liu, Z., Liu, Y., and Li, S. Z. Bridging the gap between database search and de novo peptide sequencing with searchnovo. bioRxiv, pp.\ 2024--10, 2024 c

    2024

  43. [52]

    A survey on non-autoregressive generation for neural machine translation and beyond

    Xiao, Y., Wu, L., Guo, J., Li, J., Zhang, M., Qin, T., and Liu, T.-y. A survey on non-autoregressive generation for neural machine translation and beyond. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45 0 (10): 0 11407--11427, 2023

    2023

  44. [53]

    pnovo 3: precise de novo peptide sequencing using a learning-to-rank framework

    Yang, H., Chi, H., Zeng, W.-F., Zhou, W.-J., and He, S.-M. pnovo 3: precise de novo peptide sequencing using a learning-to-rank framework. Bioinformatics, 35 0 (14): 0 i183--i190, 2019

    2019

  45. [54]

    Introducing -helixnovo for practical large-scale de novo peptide sequencing

    Yang, T., Ling, T., Sun, B., Liang, Z., Xu, F., Huang, X., Xie, L., He, Y., Li, L., He, F., et al. Introducing -helixnovo for practical large-scale de novo peptide sequencing. Briefings in Bioinformatics, 25 0 (2): 0 bbae021, 2024

    2024

  46. [55]

    Yilmaz, M., Fondrie, W., Bittremieux, W., Oh, S., and Noble, W. S. De novo mass spectrometry peptide sequencing with a transformer model. In International Conference on Machine Learning, pp.\ 25514--25522. PMLR, 2022

    2022

  47. [56]

    E., Bittremieux, W., Melendez, C

    Yilmaz, M., Fondrie, W. E., Bittremieux, W., Melendez, C. F., Nelson, R., Ananth, V., Oh, S., and Noble, W. S. Sequence-to-sequence translation from mass spectra to peptides with a transformer model. Nature communications, 15 0 (1): 0 6427, 2024

    2024

  48. [57]

    Ttida: Controllable generative data augmentation via text-to-text and text-to-image models

    Yin, Y., Kaddour, J., Zhang, X., Nie, Y., Liu, Z., Kong, L., and Liu, Q. Ttida: Controllable generative data augmentation via text-to-text and text-to-image models. arXiv preprint arXiv:2304.08821, 2023

    2023 arXiv

  49. [58]

    -primenovo: An accurate and efficient non-autoregressive deep learning model for de novo peptide sequencing

    Zhang, X., Ling, T., Jin, Z., Xu, S., Gao, Z., Sun, B., Qiu, Z., Dong, N., Wang, G., Wang, G., et al. -primenovo: An accurate and efficient non-autoregressive deep learning model for de novo peptide sequencing. bioRxiv, pp.\ 2024--05, 2024

    2024

  50. [59]

    -primenovo: an accurate and efficient non-autoregressive deep learning model for de novo peptide sequencing

    Zhang, X., Ling, T., Jin, Z., Xu, S., Gao, Z., Sun, B., Qiu, Z., Wei, J., Dong, N., Wang, G., et al. -primenovo: an accurate and efficient non-autoregressive deep learning model for de novo peptide sequencing. Nature Communications, 16 0 (1): 0 267, 2025

    2025

  51. [60]

    Structure-informed language models are protein designers

    Zheng, Z., Deng, Y., Xue, D., Zhou, Y., Ye, F., and Gu, Q. Structure-informed language models are protein designers. In International conference on machine learning, pp.\ 42317--42338. PMLR, 2023

    2023

  52. [61]

    and Keung, P

    Zhou, J. and Keung, P. Improving non-autoregressive neural machine translation with monolingual data. arXiv preprint arXiv:2005.00932, 2020

    2005 arXiv

  53. [62]

    Zhou, J., Chen, S., Xia, J., Liu, S., Ling, T., Du, W., Liu, Y., Yin, J., and Li, S. Z. Novobench: Benchmarking deep learning-based de novo peptide sequencing methods in proteomics. arXiv preprint arXiv:2406.11906, 2024

    2024 arXiv

  54. [63]

    pdeep: predicting ms/ms spectra of peptides with deep learning

    Zhou, X.-X., Zeng, W.-F., Chi, H., Luo, C., Liu, C., Zhan, J., He, S.-M., and Zhang, Z. pdeep: predicting ms/ms spectra of peptides with deep learning. Analytical chemistry, 89 0 (23): 0 12690--12697, 2017

    2017

  55. [64]

    Non-autoregressive neural machine translation with consistency regularization optimized variational framework

    Zhu, M., Wang, J., and Yan, C. Non-autoregressive neural machine translation with consistency regularization optimized variational framework. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Te...

    2022

  56. [65]

    write newline

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

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.