Pith. sign in

REVIEW 3 major objections 6 minor 84 references

Artificial Intelligence in Spectroscopy: Advancing Chemistry from Prediction to Generation and Beyond

T0 review · 3 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A unified roadmap for AI in spectroscopy, from prediction to generation

desk verdict A useful survey with a central taxonomy table that needs fixing before it can be trusted as a roadmap. read the letter →

arxiv 2502.09897 v1 pith:SO7BUP6S submitted 2025-02-14 cs.AI cs.LG

Kehan Guo , Yili Shen , Gisela Abigail Gonzalez-Montiel , Yue Huang , Yujun Zhou , Mihir Surve , Zhichun Guo , Prayel Das
show 3 more authors
Nitesh V Chawla Olaf Wiest Xiangliang Zhang
This is my paper
classification cs.AIcs.LG
keywords spectroscopymachinelearningforwardprobleminversefoundationmodelsstructureelucidationmassspectrometryNMR
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

This survey argues that the scattered applications of machine learning to spectroscopy—mass spectrometry, NMR, IR, Raman, and UV-Vis—can be organized into one coherent framework built on two complementary tasks: the forward problem of predicting a spectrum from a molecular structure, and the inverse problem of inferring a molecular structure from a spectrum. It reviews the field's history, from early pattern recognition to modern transformers and foundation models, and classifies representative architectures by input representation and output type. If the classification holds, researchers gain a common vocabulary, a way to locate methods for their modality, and a clear picture of where the field is heading. The survey also releases an open-source repository of papers and datasets to make the roadmap reproducible.

What carries the argument

The central organizing device is the forward/inverse task distinction, mapped onto a two-space diagram: the molecule space (structures, fingerprints, SMILES) and the spectrum space (vectors, sequences, images). The survey's Table 1 is the load-bearing index: it classifies dozens of recent papers by task type (classification, regression, generation, reasoning), input representation (vector, 2D graph, 3D graph, SMILES, spectrum), output, and dataset, for both the forward and inverse problems. That table, together with the historical timeline, is what turns a scattered literature into a roadmap and lets the authors identify challenges such as data quality, multimodal integration, and scalability, along with emerging directions like synthetic data generation and foundation models.

What would settle it

A reader could audit Table 1 against each cited paper's own stated task; one concrete check already suggested by the paper's text is that entry [50], listed as an inverse NMR task regressing logD from an NMR vector, is described in Section 3.2 as a forward extension predicting a physicochemical property. If a substantial fraction of entries show such mismatches between the table and the text, the survey's central organizing claim fails.

Watch

Extended reading notes

Core claim

On the paper's own terms, the contribution is a unified review of Spectroscopy Machine Learning (SpectraML) that systematically examines state-of-the-art approaches for forward tasks (molecule-to-spectrum prediction) and inverse tasks (spectrum-to-molecule inference) across five major spectroscopic techniques. The authors claim this fills a key gap in the literature because existing surveys focus on a single modality or lack a clear framework for distinguishing forward from inverse problems. They provide a taxonomy of representative neural architectures—graph-based, transformer-based, and emerging foundation models—and frame the field's evolution as a progression from predictive analytics to generative and reasoning-based systems. They further contend that unified frameworks and cross-modal integration can address both forward and inverse problems simultaneously, enabling few-shot and zero-shot learning on scarce spectral data.

Load-bearing premise

The roadmap's value depends on its literature classification being accurate and complete; if the table's forward and inverse assignments are wrong for many papers, the map misleads the researchers who rely on it.

Editorial extensions

If this is right

  • Researchers in chemistry and ML gain a shared terminology, so a method for predicting NMR shifts can be compared directly with a method for predicting IR spectra as instances of the forward problem.
  • The taxonomy highlights under-explored modality–task combinations and could direct new work toward techniques that have fewer published methods.
  • The emphasis on unified frameworks suggests that future models should be trained simultaneously on multiple spectral modalities rather than per-technique, enabling few- and zero-shot transfer to rare or novel compounds.
  • The release of a curated paper-and-dataset repository gives the community a starting point for benchmarking and reproducing results.
  • If foundation models live up to the survey's account, spectrum-to-molecule inference could shift from hand-crafted pipelines to automated reasoning with uncertainty quantification.

Reading between the lines

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

  • The forward/inverse taxonomy could be tested as a predictive scheme: papers published after the survey should be classifiable into the same boxes; a growing number of uncategorizable papers would indicate the field is outgrowing the framework.
  • The survey's Table 1 could be converted into a living benchmark where the misclassification rate is tracked across versions, turning the review into a continuously updated map of the field.
  • The same forward/inverse framing may transfer to neighboring structure–property problems beyond spectroscopy, such as predicting chromatographic retention times or electron density maps, where similar spectrum–structure mappings appear.
  • The survey's assumption that foundation models will unify forward and inverse tasks implies a concrete testable prediction: a single pretrained model should match or beat per-task specialist models on both a spectral prediction benchmark and a structure elucidation benchmark, which no current work in the surveyed set demonstrates.
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

3 major / 6 minor

Summary. The paper surveys machine learning applied to spectroscopic and spectrometric data (SpectraML), proposing a unified framework built on forward tasks (molecule-to-spectrum) and inverse tasks (spectrum-to-molecule) across MS, NMR, IR, Raman, and UV-Vis. It reviews data representations, representative neural architectures, the historical progression from classic ML to generative models and foundation models, and current challenges and opportunities. It also releases a GitHub repository of cited papers and datasets and claims to fill a gap left by single-modality surveys.

Significance. If the taxonomy and curation are reliable, this survey is a useful orientation resource for an interdisciplinary audience. Its strengths are the breadth of coverage from early pattern recognition through 2024 foundation-model work, the explicit forward/inverse framing with an acknowledgment that other communities use opposite terminology, a compact timeline (Figure 1), a summary table organized by input representation (Table 1), and the accompanying repository. The paper does not make mathematical claims that require proof, so its value rests on accurate literature classification, completeness, and clarity of the roadmap. The central risk is that the table and text are inconsistent in exactly the places that a reader would use to locate methods.

major comments (3)
  1. [§3.2, §3.3, Table 1] Table 1 is internally inconsistent with the paper's own text and with the cited papers, and since the table is the paper's primary summary and roadmap artifact, this is a load-bearing issue. Specific examples: [50] is placed in the inverse NMR subsection as 'NMR vector -> LogD value', while Section 3.2 explicitly names logD prediction as an extension of the forward problem; [21] appears in the inverse table as 'MS vector -> Reaction classification', but Section 3.2 discusses reaction-behavior classification as a forward extension; [79] is listed under 'NMR Spectral Representations' even though its input is 'MS Seq' and Section 3.3 discusses it as an MS/MS de novo sequencing method; [74] appears twice, once as forward SMILES-to-EI-MS prediction and once as inverse 'MS vector -> Intensity values', the latter contradicting both Section 3.2 and the cited paper's actual contribution; [76] is listed under inverse spectrum-to-molecule despite having output 'High-resolution image', which does not match the paper's own definition of the inverse problem in Section 2.1; and the row for [63] describes it as 'Peptide-spectrum matches', which does not match the cited paper's actual topic of predicting tensorial properties and molecular spectra. The table also omits works explicitly discussed in Section 3.2, such as [16], [17], [56], and [73]. I ask the authors to perform a full verification pass of Table 1 against the text and the cited references before revision.
  2. [Abstract, §1] The paper claims to 'systematically examine' the literature and to provide a 'comprehensive' review, but it does not report any literature-search methodology. The authors do not state which databases were searched, the time window covered, inclusion and exclusion criteria, or how the 84 references and the rows of Table 1 were selected. This makes the coverage claim difficult to assess and also makes it hard for readers to update or extend the roadmap. I recommend adding a short methodology paragraph, or at minimum an explicit scope statement, describing the search and selection process.
  3. [§2.1, Table 1] The forward/inverse dichotomy as defined in Section 2.1 is not sufficient for the entries actually included in Table 1. Under the paper's definitions, forward is molecule-to-spectrum and inverse is spectrum-to-molecule, but several entries are spectrum-to-property ([50]), spectrum-to-image ([76]), or spectrum-to-intensity ([74] in the inverse row), which fit neither category. Rather than forcing such tasks into the two existing buckets, the authors should either restrict the table to tasks that satisfy their definitions or introduce explicit additional categories (e.g., spectrum-to-property, spectrum-to-spectrum) and apply them consistently in both the text and the table.
minor comments (6)
  1. [Abstract] The sentence 'To foster reproducible research. We also release an open-source repository' is grammatically incomplete; the period after 'research' should be a comma or the sentence should be restructured.
  2. [References, Appendix] There are several proofreading issues, including 'trakcling' and 'desgin' in reference [66], 'sammple' in Section 6.4, and 'loose energy' where 'lose energy' is intended.
  3. [§2.1] The definition of spectrometry as 'measuring chemical interaction to gain insight into molecular structures and properties' is imprecise; mass spectrometry measures the mass-to-charge ratio of gas-phase ions. Please align this sentence with standard definitions to avoid confusing readers new to the field.
  4. [Throughout] Raman spectroscopy is listed as one of the five covered modalities, but it receives almost no modality-specific treatment beyond the IR/Raman background in the appendix and brief mentions in Sections 3.2 and 4.2. Consider adding a short paragraph on Raman data formats and representative methods, or explicitly state that Raman is treated through its overlap with IR-based vibrational spectroscopy.
  5. [§3.4, §4.2] Several claims about foundation models and reasoning capabilities are supported only by benchmark or dataset papers (e.g., [3], [31], [68]); the authors should clarify which references are actual foundation-model systems and which are benchmarks, datasets, or application studies.
  6. [§1, Repository] The GitHub repository is presented as a contribution, but the paper does not describe its curation, versioning, or relationship to Table 1. A brief statement of maintenance policy and repository contents would help readers trust and reuse the resource.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found; the survey's central taxonomy and literature review are built from external cited work, not from its own fitted parameters or self-referential derivations.

full rationale

This paper is a survey with no derivations, fitted parameters, or predictive claims of its own. Its contribution is a literature-based taxonomy (forward molecule-to-spectrum tasks versus inverse spectrum-to-molecule tasks) and a roadmap organized around cited external works. The few self-citations (e.g., [31], [34], [38], [41], [82]) appear only as examples of foundation-model benchmarks, molecular representation learning, or trustworthiness considerations; they are not used to define the taxonomy, justify a uniqueness theorem, or force any conclusion. The forward/inverse distinction is explicitly anchored to an external community convention with citation [53], and the paper even notes that the opposite convention exists in [4], showing that the taxonomy is not definitionally self-serving. The internal inconsistencies in Table 1 (e.g., [50] listed as an inverse NMR task while Section 3.2 describes it as a forward extension, and [74] appearing in both forward and inverse tables) are accuracy or organization concerns, not circularity: they do not make any 'prediction' reduce to its input by construction. Because the central claim is a literature organization rather than a derived result, no circular step can be exhibited, and the appropriate score is 0.

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

The survey introduces no free parameters, derived quantities, or invented entities. It is a descriptive review that rests on the accuracy of its literature selection and classification, which are assumed rather than demonstrated.

assumptions (3)
  • domain assumption The cited papers are accurately reported and representative of the field
    The survey's taxonomy and roadmap depend on the accuracy and completeness of its literature selection, as stated in Section 3 and Table 1.
  • domain assumption The forward/inverse framing is a valid organizing principle for spectroscopy ML
    The paper defines forward and inverse problems in Section 2.1, acknowledging that some communities invert the terminology [4]; the value of the roadmap depends on this framing.
  • domain assumption The GitHub repository is functional and matches the cited papers
    The survey claims to support reproducible research via the repository; this is an empirical claim about an external resource not verifiable from the preprint.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Artificial Intelligence in Spectroscopy: Advancing Chemistry from Prediction to Generation and Beyond." pith.science (2026). https://pith.science/paper/SO7BUP6S

@misc{pith2026250209897,
  author       = {Pith},
  title        = {Pith review of: Artificial Intelligence in Spectroscopy: Advancing Chemistry from Prediction to Generation and Beyond},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SO7BUP6S}},
  note         = {Machine review of arXiv:2502.09897}
}
read the original abstract

The rapid advent of machine learning (ML) and artificial intelligence (AI) has catalyzed major transformations in chemistry, yet the application of these methods to spectroscopic and spectrometric data, referred to as Spectroscopy Machine Learning (SpectraML), remains relatively underexplored. Modern spectroscopic techniques (MS, NMR, IR, Raman, UV-Vis) generate an ever-growing volume of high-dimensional data, creating a pressing need for automated and intelligent analysis beyond traditional expert-based workflows. In this survey, we provide a unified review of SpectraML, systematically examining state-of-the-art approaches for both forward tasks (molecule-to-spectrum prediction) and inverse tasks (spectrum-to-molecule inference). We trace the historical evolution of ML in spectroscopy, from early pattern recognition to the latest foundation models capable of advanced reasoning, and offer a taxonomy of representative neural architectures, including graph-based and transformer-based methods. Addressing key challenges such as data quality, multimodal integration, and computational scalability, we highlight emerging directions such as synthetic data generation, large-scale pretraining, and few- or zero-shot learning. To foster reproducible research, we also release an open-source repository containing recent papers and their corresponding curated datasets (https://github.com/MINE-Lab-ND/SpectrumML_Survey_Papers). Our survey serves as a roadmap for researchers, guiding progress at the intersection of spectroscopy and AI.

Figures

Figures reproduced from arXiv: 2502.09897 by the authors.

Figure 1
Figure 1. Timeline of ML progression and its application to spectroscopic studies. Left: Molecule to Spectrum, [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. (Top) Overview of SpectraML, translating between [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

84 extracted references · 63 canonical work pages

  1. [50]

    Leniak, A., Pietrus, W ., and Kurczab, R. From nmr to ai: designing a novel chemical representation to enhance machine learning predictions of physicochemical properties.Journal of Chemical Information and Modeling, 64(8):3302–3321, 2024

    work page 2024

  2. [74]

    N., Belanger, D., Adams, R

    Wei, J. N., Belanger, D., Adams, R. P ., and Sculley, D. Rapid prediction of electron–ionization mass spectrom- etry using neural networks. ACS central science, 5(4):700–708, 2019

    work page 2019

  3. [79]

    De novo mass spectrometry peptide sequenc- ing with a transformer model

    Yilmaz, M., Fondrie, W ., Bittremieux, W ., Oh, S., and Noble, W . De novo mass spectrometry peptide sequenc- ing with a transformer model. biorxiv, 2022.02. 07.479481, 2022

    work page 2022

  4. [21]

    K.-Y., Beck, A., Alzarieni, K

    Fine, J., Liu, J. K.-Y., Beck, A., Alzarieni, K. Z., Ma, X., Boulos, V . M., Kenttämaa, H. I., and Chopra, G. Graph- based machine learning interprets and predicts diagnostic isomer-selective ion–molecule reactions in tan- dem mass spectrometry. Chemical Science, 11(43):11849–11858, 2020

    work page 2020

  5. [76]

    T ., Athersuch, T ., Xiang, Y., Liu, Z., Jiménez, B., and Ebbels, T

    Yan, Y., Judge, M. T ., Athersuch, T ., Xiang, Y., Liu, Z., Jiménez, B., and Ebbels, T . M. Resolution enhancement of metabolomic j-res nmr spectra using deep learning. Analytical Chemistry, 96(29):11707–11715, 2024

    work page 2024

  6. [63]

    Equivariant message passing for the prediction of tensorial properties and molecular spectra

    Schütt, K., Unke, O., and Gastegger, M. Equivariant message passing for the prediction of tensorial properties and molecular spectra. In International Conference on Machine Learning, pp. 9377–9388. PMLR, 2021

    work page 2021

  7. [16]

    Pure isotropic proton nmr spectra in solids using deep learning

    Cordova, M., Moutzouri, P ., Simões de Almeida, B., Torodii, D., and Emsley, L. Pure isotropic proton nmr spectra in solids using deep learning. Angewandte Chemie International Edition, 62(8):e202216607, 2023

    work page 2023

  8. [17]

    Spectroscopy-guided deep learning predicts solid–liquid surface adsorbate properties in unseen solvents.Journal of the American Chem- ical Society, 146(1):811–823, 2023

    Du, W ., Ma, F ., Zhang, B., Zhang, J., Wu, D., Sharman, E., Jiang, J., and Wang, Y. Spectroscopy-guided deep learning predicts solid–liquid surface adsorbate properties in unseen solvents.Journal of the American Chem- ical Society, 146(1):811–823, 2023

    work page 2023

  9. [56]

    D., Joshi, R

    McNaughton, A. D., Joshi, R. P ., Knutson, C. R., Fnu, A., Luebke, K. J., Malerich, J. P ., Madrid, P . B., and Kumar, N. Machine learning models for predicting molecular uv–vis spectra with quantum mechanical properties. Journal of Chemical Information and Modeling, 63(5):1462–1471, 2023

    work page 2023

  10. [73]

    Multi-scale Sinusoidal Embeddings Enable Learning on High Resolution Mass Spectrometry Data

    Voronov, G., Lightheart, R., Davison, J., Krettler, C. A., Healey, D., and Butler, T . Multi-scale sinusoidal embed- dings enable learning on high resolution mass spectrometry data. arXiv preprint arXiv:2207.02980, 2022

    work page Pith review arXiv 2022

Show all 84 references
  1. [1]

    Alberts, M., Zipoli, F ., and Vaucher, A. C. Learning the language of nmr: Structure elucidation from nmr spectra using transformer models. 2023

    2023

  2. [2]

    Alberts, M., Laino, T ., and Vaucher, A. C. Leveraging infrared spectroscopy for automated structure elucida- tion. Communications Chemistry, 7(1):268, 2024

    2024

  3. [3]

    Unraveling molecular structure: A multimodal spectroscopic dataset for chemistry

    Alberts, M., Schilter, O., Zipoli, F ., Hartrampf, N., and Laino, T . Unraveling molecular structure: A multimodal spectroscopic dataset for chemistry. arXiv preprint arXiv:2407.17492, 2024

    2024 arXiv

  4. [4]

    G., Muhoberac, M., Randolph, C

    Beck, A. G., Muhoberac, M., Randolph, C. E., Beveridge, C. H., Wijewardhane, P . R., Kenttämaa, H. I., and Chopra, G. Recent developments in machine learning for mass spectrometry. ACS Measurement Science Au, 4(3):233–246, 2024

    2024

  5. [5]

    Binev, Y., Marques, M. M. B., and Aires-de Sousa, J. Prediction of 1h nmr coupling constants with associative neural networks trained for chemical shifts. Journal of Chemical Information and Modeling, 47(6):2089–2097, 2007

    2007

  6. [6]

    A., Kozlov, K

    Boiko, D. A., Kozlov, K. S., Burykina, J. V ., Ilyushenkova, V . V ., and Ananikov, V . P . Fully automated uncon- strained analysis of high-resolution mass spectrometry data with machine learning. Journal of the American Chemical Society, 144(32):14590–14606, 2022

    2022

  7. [7]

    A., Adeli, E., Altman, R., Arora, S., von Arx, S., Bernstein, M

    Bommasani, R., Hudson, D. A., Adeli, E., Altman, R., Arora, S., von Arx, S., Bernstein, M. S., Bohg, J., Bosselut, A., Brunskill, E., et al. On the opportunities and risks of foundation models. arXiv preprint arXiv:2108.07258, 2021

    2021 arXiv

  8. [8]

    F ., Young, A., Kretschmer, F ., Samusevich, R., Heirman, J., Wang, F ., Zhang, L., Dührkop, K., et al

    Bushuiev, R., Bushuiev, A., de Jonge, N. F ., Young, A., Kretschmer, F ., Samusevich, R., Heirman, J., Wang, F ., Zhang, L., Dührkop, K., et al. Massspecgym: A benchmark for the discovery and identification of molecules. arXiv preprint arXiv:2410.23326, 2024

    arXiv

  9. [9]

    Interleaved scene graph for interleaved text-and-image generation assessment

    Chen, D., Chen, R., Pu, S., Liu, Z., Wu, Y., Chen, C., Liu, B., Huang, Y., Wan, Y., Zhou, P ., et al. Interleaved scene graph for interleaved text-and-image generation assessment. arXiv preprint arXiv:2411.17188, 2024

    2024 arXiv

  10. [10]

    Gui-world: A dataset for gui-oriented multimodal llm-based agents

    Chen, D., Huang, Y., Wu, S., Tang, J., Chen, L., Bai, Y., He, Z., Wang, C., Zhou, H., Li, Y., et al. Gui-world: A dataset for gui-oriented multimodal llm-based agents. arXiv preprint arXiv:2406.10819, 2024

    2024 arXiv

  11. [11]

    Uncertainty-aware yield prediction with multimodal molec- ular features

    Chen, J., Guo, K., Liu, Z., Isayev, O., and Zhang, X. Uncertainty-aware yield prediction with multimodal molec- ular features. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pp. 8274–8282, 2024

    2024

  12. [12]

    Unveiling the power of language models in chemical research question answering

    Chen, X., Wang, T ., Guo, T ., Guo, K., Zhou, J., Li, H., Song, Z., Gao, X., and Zhang, X. Unveiling the power of language models in chemical research question answering. Communications Chemistry, 8(1):4, 2025

    2025

  13. [13]

    V ., Gelin, M

    Chen, Y., Pios, S. V ., Gelin, M. F ., and Chen, L. Accelerating molecular vibrational spectra simulations with a physically informed deep learning model. Journal of Chemical Theory and Computation, 2024

    2024

  14. [14]

    C., Sievers, C

    Chen, Z., Bononi, F . C., Sievers, C. A., Kong, W .-Y., and Donadio, D. Uv–visible absorption spectra of solvated molecules by quantum chemical machine learning. Journal of Chemical Theory and Computation , 18(8): 4891–4902, 2022

    2022

  15. [15]

    W ., Jin, W ., Rogers, L., Jamison, T

    Coley, C. W ., Jin, W ., Rogers, L., Jamison, T . F ., Jaakkola, T . S., Green, W . H., Barzilay, R., and Jensen, K. F . A graph-convolutional neural network model for the prediction of chemical reactivity. Chemical science, 10(2): 370–377, 2019

    2019

  16. [18]

    A., Petras, Daniel ans Ger- wick, W

    Dührkop, K., Nothias, L.-F ., Fleischauer, M., Reher, R., Ludwig, M., Hoffmann, M. A., Petras, Daniel ans Ger- wick, W . H., Rousu, J., Dorrestein, P . C., and Böcker, S. Systematic classification of unknown metabolites using high-resolution fragmentation mass spectra. Nature ...

    2020

  17. [19]

    E., Gibbons, F

    Elias, J. E., Gibbons, F . D., King, O. D., Roth, F . P ., and Gygi, S. P . Intensity-based protein identification by machine learning from a library of tandem mass spectra. Nature biotechnology, 22(2):214–219, 2004

    2004

  18. [20]

    A., North, N

    Enders, A. A., North, N. M., Fensore, C. M., Velez-Alvarez, J., and Allen, H. C. Functional group identification for ftir spectra using image-based machine learning models. Analytical Chemistry, 93(28):9711–9718, 2021. 10 Artificial Intelligence in Spectroscopy: Advancing Chem...

    2021

  19. [22]

    A., Rajasekar, A

    Fine, J. A., Rajasekar, A. A., Jethava, K. P ., and Chopra, G. Spectral deep learning for prediction and prospective validation of functional groups. Chemical science, 11(18):4618–4630, 2020

    2020

  20. [23]

    A., Bößl, F ., and Wilhelm, M

    Gabriel, W ., Shouman, O., Schröder, E. A., Bößl, F ., and Wilhelm, M. Prospect ptms: Rich labeled tandem mass spectrometry dataset of modified peptides for machine learning in proteomics. InThe Thirty-eight Conference on Neural Information Processing Systems Datasets and Benc...

  21. [24]

    Honestllm: Toward an honest and helpful large language model

    Gao, C., Wu, S., Huang, Y., Chen, D., Zhang, Q., Fu, Z., Wan, Y., Sun, L., and Zhang, X. Honestllm: Toward an honest and helpful large language model. In The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024

    2024

  22. [25]

    Machine learning molecular dynamics for the simulation of infrared spectra

    Gastegger, M., Behler, J., and Marquetand, P . Machine learning molecular dynamics for the simulation of infrared spectra. Chem. Sci., 8:6924–6935, 2017

    2017

  23. [26]

    T ., and Müller, K.-R

    Gastegger, M., Schütt, K. T ., and Müller, K.-R. Machine learning of solvent effects on molecular spectra and reactions. Chemical science, 12(34):11473–11483, 2021

    2021

  24. [27]

    A., Packer, M

    Gerrard, W ., Bratholm, L. A., Packer, M. J., Mulholland, A. J., Glowacki, D. R., and Butts, C. P . Impression – pre- diction of nmr parameters for 3-dimensional chemical structures using machine learning with near quantum chemical accuracy. Chem. Sci., 11:508–515, 2020

    2020

  25. [28]

    Prefix-tree decoding for predicting mass spectra from molecules

    Goldman, S., Bradshaw, J., Xin, J., and Coley, C. Prefix-tree decoding for predicting mass spectra from molecules. Advances in Neural Information Processing Systems, 36:48548–48572, 2023

    2023

  26. [29]

    Goldman, S., Li, J., and Coley, C. W . Generating molecular fragmentation graphs with autoregressive neural networks. Analytical Chemistry, 96(8):3419–3428, 2024

    2024

  27. [30]

    S., Gallegos, L

    Guan, Y., Sowndarya, S. S., Gallegos, L. C., John, P . C. S., and Paton, R. S. Real-time prediction of 1 h and 13 c chemical shifts with dft accuracy using a 3d graph neural network. Chemical Science, 12(36):12012–12026, 2021

    2021

  28. [31]

    V ., Wiest, O., and Zhang, X

    Guo, K., Nan, B., Zhou, Y., Guo, T ., Guo, Z., Surve, M., Liang, Z., Chawla, N. V ., Wiest, O., and Zhang, X. Can llms solve molecule puzzles? a multimodal benchmark for molecular structure elucidation. In The Thirty- eight Conference on Neural Information Processing Systems D...

    2024

  29. [32]

    What can large language models do in chemistry? a comprehensive benchmark on eight tasks

    Guo, T ., Nan, B., Liang, Z., Guo, Z., Chawla, N., Wiest, O., Zhang, X., et al. What can large language models do in chemistry? a comprehensive benchmark on eight tasks. Advances in Neural Information Processing Systems, 36:59662–59688, 2023

    2023

  30. [33]

    Guo, Z., Zhang, C., Yu, W ., Herr, J., Wiest, O., Jiang, M., and Chawla, N. V . Few-shot graph learning for molec- ular property prediction. In Proceedings of the web conference 2021, pp. 2559–2567, 2021

    2021

  31. [34]

    G., Ma, Y., Wiest, O., Zhang, X., Wang, W ., Zhang, C., et al

    Guo, Z., Guo, K., Nan, B., Tian, Y., Iyer, R. G., Ma, Y., Wiest, O., Zhang, X., Wang, W ., Zhang, C., et al. Graph- based molecular representation learning. arXiv preprint arXiv:2207.04869, 2022

    2022 arXiv

  32. [35]

    Ho Manh, L., Chen, V . C. P ., Rosenberger, J., Wang, S., Yang, Y., and Schug, K. A. Prediction of vacuum ultravi- olet/ultraviolet gas-phase absorption spectra using molecular feature representations and machine learning. Journal of Chemical Information and Modeling, 64(14):5...

    2024

  33. [36]

    S., Rotskoff, G

    Hu, F ., Chen, M. S., Rotskoff, G. M., Kanan, M. W ., and Markland, T . E. Accurate and efficient structure elu- cidation from routine one-dimensional nmr spectra using multitask machine learning. ACS Central Science, 2024

    2024

  34. [37]

    Z., et al

    Huang, Y., Shi, J., Li, Y., Fan, C., Wu, S., Zhang, Q., Liu, Y., Zhou, P ., Wan, Y., Gong, N. Z., et al. Metatool benchmark for large language models: Deciding whether to use tools and which to use. arXiv preprint arXiv:2310.03128, 2023

    arXiv

  35. [38]

    Trustgpt: A benchmark for trustworthy and responsible large language models

    Huang, Y., Zhang, Q., Sun, L., et al. Trustgpt: A benchmark for trustworthy and responsible large language models. arXiv preprint arXiv:2306.11507, 2023

    2023 arXiv

  36. [39]

    Trustllm: Trustworthiness in large language models

    Huang, Y., Sun, L., Wang, H., Wu, S., Zhang, Q., Li, Y., Gao, C., Huang, Y., Lyu, W ., Zhang, Y., et al. Trustllm: Trustworthiness in large language models. arXiv preprint arXiv:2401.05561, 2024

    2024 arXiv

  37. [40]

    Obscureprompt: Jailbreaking large language models via obscure input

    Huang, Y., Tang, J., Chen, D., Tang, B., Wan, Y., Sun, L., and Zhang, X. Obscureprompt: Jailbreaking large language models via obscure input. arXiv preprint arXiv:2406.13662, 2024

    2024 arXiv

  38. [41]

    Social science meets llms: How reliable are large language models in social simulations? arXiv preprint arXiv:2410.23426, 2024

    Huang, Y., Yuan, Z., Zhou, Y., Guo, K., Wang, X., Zhuang, H., Sun, W ., Sun, L., Wang, J., Ye, Y., et al. Social science meets llms: How reliable are large language models in social simulations? arXiv preprint arXiv:2410.23426, 2024. 11 Artificial Intelligence in Spectroscopy:...

    arXiv

  39. [42]

    Breaking focus: Contex- tual distraction curse in large language models

    Huang, Y., Wang, Y., Xu, Z., Gao, C., Wu, S., Ye, J., Chen, X., Chen, P .-Y., and Zhang, X. Breaking focus: Contex- tual distraction curse in large language models. arXiv preprint arXiv:2502.01609, 2025

    2025

  40. [43]

    S., Woroch, C

    Huang, Z., Chen, M. S., Woroch, C. P ., Markland, T . E., and Kanan, M. W . A framework for automated structure elucidation from routine nmr spectra. Chemical Science, 12(46):15329–15338, 2021

    2021

  41. [44]

    J., Madotto, A., and Fung, P

    Ji, Z., Lee, N., Frieske, R., Yu, T ., Su, D., Xu, Y., Ishii, E., Bang, Y. J., Madotto, A., and Fung, P . Survey of halluci- nation in natural language generation. ACM Computing Surveys, 55(12):1–38, 2023

    2023

  42. [45]

    Deep imitation learning for molecular inverse problems

    Jonas, E. Deep imitation learning for molecular inverse problems. Advances in neural information processing systems, 32, 2019

    2019

  43. [46]

    and Kuhn, S

    Jonas, E. and Kuhn, S. Rapid prediction of nmr spectral properties with quantified uncertainty. Journal of cheminformatics, 11:1–7, 2019

    2019

  44. [47]

    Predictive modeling of nmr chemical shifts without using atomic- level annotations

    Kang, S., Kwon, Y., Lee, D., and Choi, Y.-S. Predictive modeling of nmr chemical shifts without using atomic- level annotations. Journal of Chemical Information and Modeling, 60(8):3765–3769, 2020

    2020

  45. [48]

    W ., Zhang, C., Reher, R., Wang, M., Alexander, K

    Kim, H. W ., Zhang, C., Reher, R., Wang, M., Alexander, K. L., Nothias, L.-F ., Han, Y. K., Shin, H., Lee, K. Y., Lee, K. H., et al. Deepsat: learning molecular structures from nuclear magnetic resonance data. Journal of Cheminformatics, 15(1):71, 2023

    2023

  46. [49]

    Neural message passing for nmr chemical shift prediction

    Kwon, Y., Lee, D., Choi, Y.-S., Kang, M., and Kang, S. Neural message passing for nmr chemical shift prediction. Journal of chemical information and modeling, 60(4):2024–2030, 2020

    2024

  47. [51]

    Bart: Denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension

    Lewis, M. Bart: Denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension. arXiv preprint arXiv:1910.13461, 2019

    1910 arXiv

  48. [52]

    A machine learning protocol for revealing ion transport mechanisms from dynamic nmr shifts in paramagnetic battery materials

    Lin, M., Xiong, J., Su, M., Wang, F ., Liu, X., Hou, Y., Fu, R., Yang, Y., and Cheng, J. A machine learning protocol for revealing ion transport mechanisms from dynamic nmr shifts in paramagnetic battery materials. Chem. Sci., 13:7863–7872, 2022

    2022

  49. [53]

    Deep learning-assisted spectrum-structure correlation: State-of-the-art and perspectives

    Lu, X.-Y., Wu, H.-P ., Ma, H., Li, H., Li, J., Liu, Y.-T ., Pan, Z.-Y., Xie, Y., Wang, L., Ren, B., and Liu, G.-K. Deep learning-assisted spectrum-structure correlation: State-of-the-art and perspectives. Analytical Chemistry, 96 (20):7959–7975, 2024

    2024

  50. [54]

    Luo, T ., Mo, Z., and Pan, S. J. Fast graph generation via spectral diffusion. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023

    2023

  51. [55]

    Coculture of two developmental stages of a marine-derived aspergillus alliaceus results in the production of the cytotoxic bianthrone allianthrone a

    Mandelare, P ., Adpressa, D., Kaweesa, E., Zakharov, L., and Loesgen, S. Coculture of two developmental stages of a marine-derived aspergillus alliaceus results in the production of the cytotoxic bianthrone allianthrone a. Journal of Natural Products, 81(4):1014–1022, 2018

    2018

  52. [57]

    Moor, M., Banerjee, O., Abad, Z. S. H., Krumholz, H. M., Leskovec, J., Topol, E. J., and Rajpurkar, P . Foundation models for generalist medical artificial intelligence. Nature, 616(7956):259–265, 2023

    2023

  53. [58]

    Efficiently predicting high resolution mass spectra with graph neural networks

    Murphy, M., Jegelka, S., Fraenkel, E., Kind, T ., Healey, D., and Butler, T . Efficiently predicting high resolution mass spectra with graph neural networks. arxiv 2023. arXiv preprint arXiv:2301.11419, 2023

    2023 arXiv

  54. [59]

    An introduction to convolutional neural networks

    O’Shea, K. An introduction to convolutional neural networks. arXiv preprint arXiv:1511.08458, 2015

    2015 arXiv

  55. [60]

    Mass spectra prediction with structural motif-based graph neural networks

    Park, J., Jo, J., and Yoon, S. Mass spectra prediction with structural motif-based graph neural networks. Scien- tific Reports, 14(1):1400, 2024

    2024

  56. [61]

    Ralbovsky, N. M. and Lednev, I. K. Towards development of a novel universal medical diagnostic method: Raman spectroscopy and machine learning. Chemical Society Reviews, 49(20):7428–7453, 2020

    2020

  57. [62]

    and Paliwal, K

    Schuster, M. and Paliwal, K. K. Bidirectional recurrent neural networks.IEEE transactions on Signal Processing, 45(11):2673–2681, 1997

    1997

  58. [64]

    Graph neural networks for learning molecular excitation spectra

    Singh, K., Munchmeyer, J., Weber, L., Leser, U., and Bande, A. Graph neural networks for learning molecular excitation spectra. Journal of Chemical Theory and Computation, 18(7):4408–4417, 2022. 12 Artificial Intelligence in Spectroscopy: Advancing Chemistry from Prediction to...

    2022

  59. [65]

    Sowa, J. K. and Rossky, P . J. A bond-based machine learning model for molecular polarizabilities and a priori raman spectra. Journal of Chemical Theory and Computation, 20(22):10071–10079, 2024

    2024

  60. [66]

    Sridharan, B., Goel, M., and Priyakumar, U. D. Modern machine learning for trackling inverse problems in chemistry: molecular desgin to realization. Chemical Communication, 58(35):5316–5331, 2022

    2022

  61. [67]

    Sridharan, B., Mehta, S., Pathak, Y., and Priyakumar, U. D. Deep reinforcement learning for molecular inverse problem of nuclear magnetic resonance spectra to molecular structure. The Journal of Physical Chemistry Letters, 13(22):4924–4933, 2022

    2022

  62. [68]

    Automation and machine learning augmented by large language models in a catalysis study

    Su, Y., Wang, X., Ye, Y., Xie, Y., Xu, Y., Jiang, Y., and Wang, C. Automation and machine learning augmented by large language models in a catalysis study. Chemical Science, 15(31):12200–12233, 2024

    2024

  63. [69]

    Cross-modal retrieval between 13c nmr spectra and structures based on focused libraries

    Sun, H., Xue, X., Liu, X., Hu, H.-Y., Deng, Y., and Wang, X. Cross-modal retrieval between 13c nmr spectra and structures based on focused libraries. Analytical Chemistry, 96(15):5763–5770, 2024

    2024

  64. [70]

    Large language models for data annotation and synthesis: A survey

    Tan, Z., Li, D., Wang, S., Beigi, A., Jiang, B., Bhattacharjee, A., Karami, M., Li, J., Cheng, L., and Liu, H. Large language models for data annotation and synthesis: A survey. In Al-Onaizan, Y., Bansal, M., and Chen, Y.-N. (eds.), Proceedings of the 2024 Conference on Empiri...

    2024 doi

  65. [71]

    Enhancing chemical reac- tion monitoring with a deep learning model for nmr spectra image matching to target compounds

    Tian, Z., Dai, Y., Hu, F ., Shen, Z., Xu, H., Zhang, H., Xu, J., Hu, Y., Diao, Y., and Li, H. Enhancing chemical reac- tion monitoring with a deep learning model for nmr spectra image matching to target compounds. Journal of Chemical Information and Modeling, 64(14):5624–5633, 2024

    2024

  66. [72]

    A., Karlsson, N

    Urban, J., Jin, C., Thomsson, K. A., Karlsson, N. G., Ives, C. M., Fadda, E., and Bojar, D. Predicting glycan structure from tandem mass spectrometry via deep learning. Nature Methods, 21(7):1206–1215, 2024

    2024

  67. [75]

    Unigen: A unified framework for textual dataset generation using large language models

    Wu, S., Huang, Y., Gao, C., Chen, D., Zhang, Q., Wan, Y., Zhou, T ., Zhang, X., Gao, J., Xiao, C., et al. Unigen: A unified framework for textual dataset generation using large language models. arXiv preprint arXiv:2406.18966, 2024

    2024

  68. [77]

    Conditional molec- ular generation net enables automated structure elucidation based on 13c nmr spectra and prior knowledge

    Yao, L., Yang, M., Song, J., Yang, Z., Sun, H., Shi, H., Liu, X., Ji, X., Deng, Y., and Wang, X. Conditional molec- ular generation net enables automated structure elucidation based on 13c nmr spectra and prior knowledge. Analytical chemistry, 95(12):5393–5401, 2023

    2023

  69. [78]

    D., Zhang, G., Mukamel, S., and Jiang, J

    Ye, S., Zhong, K., Zhang, J., Hu, W ., Hirst, J. D., Zhang, G., Mukamel, S., and Jiang, J. A machine learning protocol for predicting protein infrared spectra. Journal of the American Chemical Society , 142(45):19071– 19077, 2020

    2020

  70. [80]

    Massformer: Tandem mass spectrum prediction for small molecules using graph transformers

    Young, A., Wang, B., and Röst, H. Massformer: Tandem mass spectrum prediction for small molecules using graph transformers. arXiv preprint arXiv:2111.04824, 2021

    2021 arXiv

  71. [81]

    Cbmaff-net: An intelligent nmr-based nontargeted screening method for new psychoactive substances

    Zheng, X., Tang, B., Xu, P ., Wang, Y., Di, B., Hua, Z., Su, M., and Liao, J. Cbmaff-net: An intelligent nmr-based nontargeted screening method for new psychoactive substances. Analytical Chemistry, 96(47):18672–18680, 2024

    2024

  72. [82]

    V ., and Zhang, X

    Zhou, Y., Yang, J., Guo, K., Chen, P .-Y., Gao, T ., Geyer, W ., Moniz, N., Chawla, N. V ., and Zhang, X. Labsafety bench: Benchmarking llms on safety issues in scientific labs. arXiv preprint arXiv:2410.14182, 2024

    2024

  73. [83]

    Using graph neural networks for mass spectrometry prediction

    Zhu, H., Liu, L., and Hassoun, S. Using graph neural networks for mass spectrometry prediction. arXiv preprint arXiv:2010.04661, 2020

    2010 arXiv

  74. [84]

    Zhu, R. L. and Jonas, E. Rapid approximate subset-based spectra prediction for electron ionization–mass spectrometry. Analytical chemistry, 95(5):2653–2663, 2023. 13 Artificial Intelligence in Spectroscopy: Advancing Chemistry from Prediction to Generation and Beyond 6 Appendi...

    2023

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.