Pith. sign in

REVIEW 5 major objections 5 minor 2 cited by

Navigating Chemical-Linguistic Sharing Space with Heterogeneous Molecular Encoding

T0 review · 5 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read A unified molecular encoder that fuses fragment, 2D, and 3D views lets chemical language models both design molecules from text and describe molecules from structure.

desk verdict A useful, ambitious framework and a big new dataset, but the central generation claim needs a same-backbone SMILES-only control before it convinces. read the letter →

arxiv 2412.20888 v2 pith:52WIVFQK submitted 2024-12-30 cs.CE

Liuzhenghao Lv , Hao Li , Yu Wang , Zhiyuan Yan , Zijun Chen , Zongying Lin , Li Yuan , Yonghong Tian This is my paper · ORCID
classification cs.CE
keywords chemicallanguagemodelsmoleculardesignheterogeneousencodingfragment-baseddrugchain-of-thoughtmulti-conditionalgenerationcaptioningzero-shot
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 paper claims that chemical language models fail to explore the shared chemical-linguistic space because each molecular representation—SMILES strings, 2D graphs, 3D coordinates—carries its own bias, and no single view captures enough information to reliably translate between molecules and natural language. To close that gap, the authors propose Heterogeneous Molecular Encoding (HME), which compresses all three views plus mined molecular fragment sequences into a single set of tokens aligned with an LLM decoder. They also build a new million-molecule dataset, MCMoD, with property, fragment, and description conditions. On this foundation, HME reports large gains in both directions: molecule-to-text tasks such as captioning and QA, and text-to-molecule tasks such as multi-conditional and zero-shot molecular design. A sympathetic reader would take the paper's claim to be that a multi-view, fragment-aware encoding is what lets CLMs navigate chemical-linguistic space in both directions at once.

What carries the argument

The central object is the HME encoder, which combines four tokenizers: the LLM's own SMILES embedding, a pretrained graph encoder for 2D topology, a pretrained 3D coordinate encoder, and a fragment tokenizer built by mining maximum-frequency subgraphs into a fixed-size vocabulary that is appended to the LLM's native vocabulary. Q-learning compresses the variable-length 2D and 3D features into a fixed number of query tokens via cross-attention, a self-attention fusion module merges those into a unified geometry feature, and a projection module aligns all views with the textual space. The fragment tokens serve a dual role: they act as explicit control conditions for fragment-targeted design and as an intermediate chain-of-thought sequence that the model generates before producing the final SMILES string.

What would settle it

Retrain HME with the fragment chains randomly shuffled for each molecule while keeping the fragment set identical; if the reported captioning and design scores do not drop, the specific ordering carries no information and the chain-of-thought mechanism is not what drives the gains.

Watch

Extended reading notes

Core claim

The central claim is that a heterogeneous molecular encoder, HME, enables a single chemical language model to perform both molecular comprehension and molecular generation by feeding the LLM four complementary views—SMILES tokens, a 2D graph encoding, a 3D coordinate encoding, and fragment-sequence tokens—compressed into a fixed number of query tokens and fused before autoregressive decoding. Using this encoder and the newly constructed MCMoD dataset, the authors report state-of-the-art results on molecular captioning, general molecular QA, and property QA, clearly beating baselines that use only one or two views. On the generation side, HME improves description-based molecular generation by +8.9% in FCD and molecule-to-text generation by +11.6% in BLEU, and it achieves high success in multi-objective inverse design with property and fragment controls, including a 79.4% success rate on a zero-shot dual-property condition and 92.64% complete fragment inclusion when a single fragment is specified. The paper further argues that fragment sequences work as a chain-of-thought, letting the model first reason about which building blocks are needed and then assemble the final molecule.

Load-bearing premise

The method assumes that the lexicographic ordering of fragment sequences used in MCMoD carries chemically meaningful step-by-step reasoning; if that ordering is arbitrary, the reported chain-of-thought benefit may be only an effect of adding many new tokens to the vocabulary.

Editorial extensions

If this is right

  • One model can handle both comprehension and generation tasks, including captioning, property QA, description-to-molecule design, fragment-constrained design, and docking-aware ligand generation, all within a single autoregressive framework.
  • Text and fragment conditions compose: users can jointly specify properties such as LogP, QED, SAS, or docking score together with one or more molecular fragments, and the model reportedly follows those controls.
  • Fragment tokens make molecular design more interpretable: the model first outputs a fragment chain and then the final molecule, and the paper reports high consistency between the chain and the final structure (precision 0.769, recall 0.774).
  • The model generalizes zero-shot: it handles four-fragment conditions despite training on at most three, and it handles dual-property conditions despite training on single-property controls.
  • The MCMoD dataset provides a large-scale benchmark with over one million molecules and diverse conditions, enabling systematic evaluation of conditional molecular generation for future methods.

Reading between the lines

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

  • If the fragment chain-of-thought is genuinely doing reasoning work, the same vocabulary-expansion idea could transfer to other structured scientific outputs—such as reaction sequences, retrosynthetic paths, or polymer assembly steps—where an explicit intermediate token stream might guide generation.
  • The lexicographic ordering of fragment sequences is a confound; a controlled comparison with randomly shuffled fragment order, or with an equal number of added non-chemical tokens, would separate vocabulary-capacity effects from true chain-of-thought reasoning.
  • The paper's bias analysis, showing very low correlation between 1D, 2D, and 3D similarity matrices, implies that any single-view molecular LLM leaves substantial information unused; this motivates multi-view encoders even for tasks that appear text-only.
  • The zero-shot results suggest that natural-language condition prompts can be composed at inference time beyond what was seen in training, which points toward a practical recipe for extending chemical LLMs to new constraint combinations without retraining.
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

5 major / 5 minor

Summary. The paper proposes Heterogeneous Molecular Encoding (HME), a framework that augments a chemical language model decoder (Llama-3-8B) with 1D SMILES, 2D graph, 3D coordinate, and molecular fragment encoders, and introduces a large multi-conditional molecular design dataset (MCMoD). The authors report improvements over several baselines in molecular captioning, general/property QA, description-based molecular generation, multi-objective inverse design, and zero-shot design settings, with abstract-level claims of +8.9% FCD for chemical-space exploration and +11.6% BLEU for text generation. The paper also claims that fragment sequences act as a Chain of Thought that improves interpretability and generation.

Significance. The paper makes a substantial empirical contribution: it releases MCMoD, a large public dataset for multi-conditional molecular generation; it provides code; and it unifies comprehension and generation tasks in a single framework. The zero-shot tests (four-fragment control, dual-property control) are a strength, as they appear to be held out from training. The property-QA results in Supplementary Table 4 are notably strong, with large improvements in MAE and validity over the baselines. The authors also explicitly acknowledge limitations in dataset scale and the absence of molecular dynamics, which is appropriate. However, the central attribution of the generation-task gains to the heterogeneous encoders is currently under-supported because no same-backbone, same-data ablation is provided for the Table 2 tasks.

major comments (5)
  1. [Results, Table 2] The central +8.9% FCD claim (abstract and description-based molecular generation) is not attributable to HME because no same-backbone, same-data control is provided. HME is a Llama-3-8B decoder fine-tuned on MCMoD with fragment, graph, and 3D streams plus fragment CoT, while the baselines are MolT5 (T5-scale), MolXPT (GPT-2-scale), and GPT-3.5. There is no Llama-3-8B model trained on MCMoD with SMILES and text only (or with SMILES, text, and the same LoRA protocol but without the extra encoder streams). The only component ablations (Supplementary Table 7) are performed on HME-Small and only on captioning/QA, not on the generation tasks in Table 2. The reported gains could therefore be caused by the larger backbone and the MCMoD fine-tuning corpus rather than by HME's heterogeneous encoding. Please add a same-backbone, same-data ablation that removes the fragment/graph/3D streams and the fragment CoT objective.
  2. [Fig. 4, Multi-Objective Molecule Reverse Design] The multi-objective inverse-design results are reported only for HME itself, with no comparison to prior conditional generators. The text reports success rates such as δ=60.2%, 69.4%, and 79.4%, and fragment inclusion rates of 92.64%, 78.84%, and 62.59%, but no baseline model is evaluated on the same MCMoD test prompts. Without a comparison to a property-conditioned or fragment-conditioned baseline (e.g., an RNN or transformer trained with the same property/fragment prompts, or an existing conditional generator), the claim of 'significant improvements' for multi-objective design is unsupported. Please add at least one strong baseline evaluated under identical prompt and evaluation conditions.
  3. [Supplementary Section 2; Description-Based Molecular Generation] The claim that fragment sequences serve as a Chain of Thought with interpretability rests on the assumption that lexicographically sorted fragment sequences are a meaningful intermediate reasoning representation. The MCMoD construction note states that fragments are 'sorted lexicographically to mitigate the adverse effects of sequence order,' but no experiment compares lexicographic order with random order or with a chemically motivated order. In addition, no experiment separates the benefit of the fragment vocabulary itself from the benefit of the CoT objective: a model with the same augmented vocabulary but trained to generate the molecule directly (without first emitting fragment tokens) is not tested. The reported CoT precision/recall (0.769/0.774) and the generation gains in Table 2 could thus be artifacts of vocabulary expansion rather than of chemical reasoning. Please provide a control that isolates the CoT ordering and the CoT objective.
  4. [Results, zero-shot subsection] The zero-shot claims (four-fragment control and dual-property control) require that those exact conditions do not appear in the training set, but the paper never describes the MCMoD train/test split or the procedure for excluding overlapping molecules or conditions. The only leakage-prevention statement is in Supplementary Section 2, where molecules overlapping the ChEBI test set are filtered out for PubChem; no analogous deduplication or condition-exclusion is described for the ZINC, COCONUT, and DTP subsets. Please specify the split and confirm by construction that no training sample contains four fragments or dual-property conditions.
  5. [Throughout Results; Abstract] The word 'significant' is used repeatedly (e.g., '+8.9% FCD' in the abstract, 'significant improvements' in Results) but no variance estimates, confidence intervals, or significance tests are reported anywhere. Tables 1 and 2 and Fig. 4 contain single-point estimates only; no number of independent runs, seeds, or bootstrap intervals is given. The statistical significance of the claimed improvements is therefore unsupported. Please add replication runs or bootstrap confidence intervals for the main claims, or replace 'significant' with 'substantial' and avoid statistical language.
minor comments (5)
  1. [Methods, Eq. (9)] The cross-attention compression using query tokens is called 'Q-learning,' which is easily confused with the reinforcement-learning algorithm of the same name; consider renaming it 'query-token compression' or adding a sentence to disambiguate.
  2. [Figure 2] The subfigure labels for RMSE values (e.g., 'RMSE = 21.36' vs. 'RMSE = 0.60') are not all clearly associated with the corresponding models or distributions; a legend or clearer axis annotations would improve readability.
  3. [Supplementary Figure 5] Several fragment strings in Supplementary Figure 5 contain ellipses (e.g., 'NC(=O)C...C=CC=C1'), which appear to be display artifacts rather than valid SMILES; please correct the figure or replace them with placeholder notation that cannot be mistaken for actual token content.
  4. [Table 2 caption] The phrase 'the fragment highlight area' is vague; please specify which color or annotation in the example corresponds to the fragment tokens and how the highlighted text relates to the fragment chain.
  5. [Abstract] The abbreviation 'Meteor' is used inconsistently with the standard spelling 'METEOR' in Figure 2 and throughout the text; standardize the spelling.

Circularity Check

0 steps flagged · score 2.0 of 10

No load-bearing circularity: held-out evaluations anchor the central claims; only peripheral self-citations.

full rationale

The paper's central quantitative claims are held-out evaluations: comprehension metrics (Table 1) on the external 3D-MoIT benchmark, generation metrics (Table 2) on MCMoD test prompts with standard BLEU/FCD/fingerprint metrics, and zero-shot property/fragment controls whose success is scored by RDKit-computed values after generation. None of these reduces to a fitted parameter renamed as a prediction; the MCMoD construction, while self-built, is used to compare methods rather than to define the outcome. The fragment CoT consistency check (precision/recall 0.769/0.774) uses the authors' fragment generator both to build training chains and to score generated molecules, so it is a weak, learned-consistency check; however, it is not the origin of the headline FCD/BLEU gains and is not algebraically forced at inference. Self-citations (e.g., refs 43, 71) name a cross-attention module and fragment-based-design context; they are not load-bearing theorems, and no uniqueness result is imported to forbid alternatives. The lack of a same-backbone, same-data SMILES-only control for Table 2 is a genuine attribution/correctness limitation (scale and MCMoD fine-tuning could contribute to the gap), but it is not circularity. Accordingly, no derivation step is equivalent to its input by construction; score 2 only for the peripheral self-citations.

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

The main claims rest on learned encoders and standard autoregressive training, but the paper introduces hand-chosen hyperparameters (fragment vocabulary size, query token count, property quantization, sampled condition fragments) and relies on RDKit and docking-score approximators as ground truth. No new physical entities are introduced.

free parameters (4)
  • Fragment vocabulary size = 800 entries per vocabulary
    Hand-chosen vocabulary size for the fragment tokenizer; determines the CoT control space and is used in all fragment-based tasks (Supplementary Section 7).
  • Number of query tokens Nq = not reported
    Query-based compression of 2D/3D features in Eq. (9) uses a hand-set sequence length; the paper never states Nq, yet it controls information loss and model performance.
  • Property discretization step = LogP: 1.0, SAS: 1.0, QED: 0.1
    Property values are quantized for conditional prompts; this choice defines the granularity of property control and the zero-shot delta success ranges.
  • Condition fragment count sampling = 1-3 fragments per training sample
    Randomly selected one to three fragments are used as fragment conditions in MCMoD; this sets the distribution that makes 4-fragment inference 'zero-shot'.
assumptions (4)
  • domain assumption RDKit-calculated LogP, QED, and SAS values are treated as ground truth for training and evaluation.
    Property QA and multi-objective design use these computed values as labels; any systematic RDKit error propagates into reported controls (Methods and MCMoD construction).
  • domain assumption QuickVina-2 and SMINA docking scores are treated as ground-truth binding affinity.
    Affinity-based ligand generation and high-affinity rates depend on these approximators, with no error bars or experimental validation.
  • domain assumption Morgan fingerprint similarity is the reference for substructure information in the bias analysis.
    The claim that HME is the least biased encoder relies on correlation with Morgan fingerprints on a 2,000-molecule sample (Supplementary Section 1).
  • ad hoc to paper Fragment sequences, sorted lexicographically, retain chemically meaningful order when used as Chain of Thought.
    MCMoD sorts fragment chains lexicographically to reduce order effects; the paper gives no chemical justification for this ordering, yet it underpins the fragment CoT and coherence claims.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Navigating Chemical-Linguistic Sharing Space with Heterogeneous Molecular Encoding." pith.science (2026). https://pith.science/paper/52WIVFQK

@misc{pith2026241220888,
  author       = {Pith},
  title        = {Pith review of: Navigating Chemical-Linguistic Sharing Space with Heterogeneous Molecular Encoding},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/52WIVFQK}},
  note         = {Machine review of arXiv:2412.20888}
}
read the original abstract

Chemical language models (CLMs) are prominent for their effectiveness in exploring chemical space and enabling molecular engineering. However, while exploring chemical-linguistic space, CLMs suffer from the gap between natural language and molecular representations. This challenge is primarily due to the inherent modeling differences between molecules and texts: molecules operate unified modeling to learn chemical space, while natural language sequentially models the semantic space. Additionally, the limited availability of high-quality text-to-molecule datasets further exacerbates this challenge. To address the problem, we first verified the information bias in molecular representations from different perspectives. We then developed the Heterogeneous Molecular Encoding (HME) framework, a unified molecular encoder compressing the molecular features from fragment sequence, topology, and conformation with Q-learning. To better model chemical-linguistic space, we further constructed the MCMoD dataset, which contains over one million molecules with various conditions, including properties, fragments, and descriptions. Experimentally, HME promotes CLMs to achieve chemical-linguistic sharing space exploration: (1) chemical space exploration with linguistic guidance, where HME achieves significant improvements (+8.9\% FCD) for molecular design in multiple constraints, even in zero-shot scenarios; (2) linguistic space exploration with molecular guidance, where HME generates textual descriptions with high qualities (+11.6\% BLEU) for molecules. These results highlight the precision of HME in handling multi-objective and cross-domain tasks, as well as its remarkable generalization capability on unseen task combinations. HME offers a new perspective on navigating chemical-linguistic sharing space, advancing the potential of CLMs in both fundamental research and practical applications in chemistry.

Figures

Figures reproduced from arXiv: 2412.20888 by the authors.

Figure 2
Figure 2. Experimental Analysis for Molecular Comprehension. a Captioning Task and General QA Task: The violin plot of the similarity scores between generated texts and reference texts from HME and baselines. The area size is positively correlated with the proportion of high-quality texts. B-2 means the metric BLEU-2, R-1 means the metric ROUGE-1, and M means the metric Meteor. µ means the average score of all texts from HME … view at source ↗
Figure 3
Figure 3. Visualization for Conditional Molecular Generation of HME. a We visualize eight synthetic molecules designed by HME under the joint control of property values and specific fragments. α denotes the target property value while β denotes the actual value calculated by RDKit. To the left of the dotted line are the specified fragments, which are also highlighted in the generated molecules. b Similarly, we visualize eight… view at source ↗
Figure 4
Figure 4. Experimental Analysis for Conditional Molecular Generation. a The distribution of the actual property values calculated by RDKit of the generated molecules under the control of the target property. Our model can effectively follow property controls. b Nine examples to demonstrate the effectiveness of property control. With the Pyridine as the anchor fragment condition, we use different target property types and valu… view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Hypothesis-and-Refinement Learning of Organic Structures from Multimodal Spectroscopic Data

    physics.chem-ph 2026-07 conditional novelty 6.0 of 10

    A two-stage AI pipeline — spectral hypothesis generation followed by mass-constrained molecular refinement — reconstructs organic structures from multimodal spectra, with 93.8% top-1 accuracy on simulated QM9 data and...

  2. LatentChem: From Textual CoT to Latent Thinking in Chemical Reasoning

    physics.chem-ph 2026-02 conditional novelty 6.0 of 10

    LatentChem reasons in continuous latent space for chemistry, achieving a 59.88% non-tie win rate over explicit CoT on ChemCoTBench with a 10.84x average reduction in reasoning overhead.

Reference graph

Works this paper leans on

78 extracted references · 56 canonical work pages · cited by 2 Pith papers

  1. [1]

    Chemical space: limits, evolution and modelling of an object bigger than our universal library

    Restrepo, G. Chemical space: limits, evolution and modelling of an object bigger than our universal library. Digit. Discov. 1, 568–585 (2022)

    work page 2022

  2. [2]

    S., Henson, A

    Gromski, P. S., Henson, A. B., Granda, J. M. & Cronin, L. How to explore chemical space using algorithms and automation. Nat. Rev. Chem. 3, 119–128 (2019)

    work page 2019

  3. [3]

    keywords

    Wo´ zniak, M.et al. Linguistic measures of chemical diversity and the “keywords” of molecular collections. Sci. reports 8, 7598 (2018)

    work page 2018

  4. [4]

    & Huang, M

    Zhou, J. & Huang, M. Navigating the landscape of enzyme design: from molecular simulations to machine learning. Chem. Soc. Rev. (2024)

    work page 2024

  5. [5]

    Smiles, a chemical language and information system

    Weininger, D. Smiles, a chemical language and information system. 1. introduction to methodology and encoding rules. J. chemical information computer sciences 28, 31–36 (1988)

    work page 1988

  6. [6]

    & Aspuru-Guzik, A

    Krenn, M., Häse, F., Nigam, A., Friederich, P. & Aspuru-Guzik, A. Self-referencing embedded strings (selfies): A 100% robust molecular string representation. Mach. Learn. Sci. Technol. 1, 045024 (2020)

    work page 2020

  7. [7]

    Edwards, C. et al. Translation between molecules and natural language. In 2022 Conference on Empirical Methods in Natural Language Processing, EMNLP 2022 (2022)

    work page 2022

  8. [8]

    Pei, Q. et al. Biot5: Enriching cross-modal integration in biology with chemical knowledge and natural language associations. arXiv preprint arXiv:2310.07276 (2023)

    arXiv 2023

Show all 78 references
  1. [9]

    Mol-instructions: A large-scale biomolecular instruction dataset for large language models

    Fang, Y .et al. Mol-instructions: A large-scale biomolecular instruction dataset for large language models. arXiv preprint arXiv:2306.08018 (2023)

    2023 arXiv

  2. [10]

    C., Boukouvalas, Z., Fuge, M

    Elton, D. C., Boukouvalas, Z., Fuge, M. D. & Chung, P. W. Deep learning for molecular design—a review of the state of the art. Mol. Syst. Des. & Eng. 4, 828–849 (2019)

    2019

  3. [11]

    Brown, N., Fiscato, M., Segler, M. H. & Vaucher, A. C. Guacamol: benchmarking models for de novo molecular design. J. chemical information modeling 59, 1096–1108 (2019)

    2019

  4. [12]

    Li, J. et al. Empowering molecule discovery for molecule-caption translation with large language models: A chatgpt perspective. IEEE Transactions on Knowl. Data Eng. (2024)

    2024

  5. [13]

    T., Nguyen, N

    Tran, D., Pham, N. T., Nguyen, N. & Manavalan, B. Mol2lang-vlm: Vision-and text-guided generative pre-trained language models for advancing molecule captioning through multimodal fusion. In Proceedings of the 1st Workshop on Language+ Molecules (L+ M 2024), 97–102 (2024)

    2024

  6. [14]

    Qureshi, R. et al. Ai in drug discovery and its clinical relevance. Heliyon 9 (2023)

    2023

  7. [15]

    Zongying, L. et al. Taxdiff: Taxonomic-guided diffusion model for protein sequence generation. arXiv preprint arXiv:2402.17156 (2024)

    2024 arXiv

  8. [16]

    Pei, Q. et al. Leveraging biomolecule and natural language through multi-modal learning: A survey. arXiv preprint arXiv:2403.01528 (2024)

    2024

  9. [17]

    Wang, H. et al. Chemical-reaction-aware molecule representation learning. In International Conference on Learning Representations (2022)

    2022

  10. [18]

    Zhou, G. et al. Uni-mol: A universal 3d molecular representation learning framework. In The Eleventh International Conference on Learning Representations (2023)

    2023

  11. [19]

    Zhang, Q. et al. Scientific large language models: A survey on biological & chemical domains. arXiv preprint arXiv:2401.14656 (2024)

    2024 arXiv

  12. [20]

    A comprehensive survey of scientific large language models and their applications in scientific discovery

    Zhang, Y .et al. A comprehensive survey of scientific large language models and their applications in scientific discovery. arXiv preprint arXiv:2406.10833 (2024)

    2024 arXiv

  13. [21]

    Liu, S. et al. Multi-modal molecule structure–text model for text-based retrieval and editing. Nat. Mach. Intell. 5, 1447–1457 (2023)

    2023

  14. [22]

    Cao, H., Liu, Z., Lu, X., Yao, Y . & Li, Y . Instructmol: Multi-modal integration for building a versatile and reliable molecular assistant in drug discovery. arXiv preprint arXiv:2311.16208 (2023)

    2023 arXiv

  15. [23]

    Li, S. et al. Towards 3d molecule-text interpretation in language models. arXiv preprint arXiv:2401.13923 (2024)

    2024 arXiv

  16. [24]

    Skinnider, M. A. Invalid smiles are beneficial rather than detrimental to chemical language models. Nat. Mach. Intell. 6, 437–448 (2024). 14/26

    2024

  17. [25]

    Chen, D. et al. Measuring and relieving the over-smoothing problem for graph neural networks from the topological view. In Proceedings of the AAAI conference on artificial intelligence, vol. 34, 3438–3445 (2020)

    2020

  18. [26]

    & Shi, C

    Xing, Y ., Wang, X., Li, Y ., Huang, H. & Shi, C. Less is more: on the over-globalizing problem in graph transformers. In Forty-first International Conference on Machine Learning(2024)

    2024

  19. [27]

    M., Havel, T

    Crippen, G. M., Havel, T. F. et al. Distance geometry and molecular conformation , vol. 74 (Research Studies Press Taunton, 1988)

    1988

  20. [28]

    Liu, S. et al. Pre-training molecular graph representation with 3d geometry. In International Conference on Learning Representations (2022)

    2022

  21. [29]

    Multimodal molecular pretraining via modality blending

    Yu, Q.et al. Multimodal molecular pretraining via modality blending. In The Twelfth International Conference on Learning Representations (2024)

    2024

  22. [30]

    Feng, S. et al. Unicorn: A unified contrastive learning approach for multi-view molecular representation learning. arXiv preprint arXiv:2405.10343 (2024)

    2024 arXiv

  23. [31]

    Feng, B. et al. A bioactivity foundation model using pairwise meta-learning. Nat. Mach. Intell. 6, 962–974 (2024)

    2024

  24. [32]

    A., Malde, A

    Khedkar, S. A., Malde, A. K., Coutinho, E. C. & Srivastava, S. Pharmacophore modeling in drug discovery and development: an overview. Medicinal Chem. 3, 187–197 (2007)

    2007

  25. [33]

    Influence of molecular vibrations on substitution coordinates

    Van Eijck, B. Influence of molecular vibrations on substitution coordinates. J. Mol. Spectrosc. 91, 348–362 (1982)

    1982

  26. [34]

    Tang, Z. et al. Cycle3d: High-quality and consistent image-to-3d generation via generation-reconstruction cycle. arXiv preprint arXiv:2407.19548 (2024)

    2024 arXiv

  27. [35]

    On the perception of molecules from 3d atomic coordinates

    Labute, P. On the perception of molecules from 3d atomic coordinates. J. chemical information modeling 45, 215–221 (2005)

    2005

  28. [36]

    Wei, J. et al. Chain-of-thought prompting elicits reasoning in large language models. Adv. neural information processing systems 35, 24824–24837 (2022)

    2022

  29. [37]

    Chu, Z. et al. A survey of chain of thought reasoning: Advances, frontiers and future. arXiv preprint arXiv:2309.15402 (2023)

    2023 arXiv

  30. [38]

    Embodiedgpt: Vision-language pre-training via embodied chain of thought

    Mu, Y .et al. Embodiedgpt: Vision-language pre-training via embodied chain of thought. Adv. Neural Inf. Process. Syst. 36 (2024)

    2024

  31. [39]

    The curse of dimensionality

    Köppen, M. The curse of dimensionality. In 5th online world conference on soft computing in industrial applications (WSC5), vol. 1, 4–8 (2000)

    2000

  32. [40]

    & Ueda, H

    Honda, S., Shi, S. & Ueda, H. R. Smiles transformer: Pre-trained molecular fingerprint for low data drug discovery. arXiv preprint arXiv:1911.04738 (2019)

    2019 arXiv

  33. [41]

    Morgan, H. L. The generation of a unique machine description for chemical structures-a technique developed at chemical abstracts service. J. chemical documentation 5, 107–113 (1965)

    1965

  34. [42]

    & Héberger, K

    Bajusz, D., Rácz, A. & Héberger, K. Why is tanimoto index an appropriate choice for fingerprint-based similarity calculations? J. cheminformatics 7, 1–13 (2015)

    2015

  35. [43]

    Li, H. et al. Decoupled peak property learning for efficient and interpretable ecd spectra prediction. arXiv preprint arXiv:2401.03403 (2024)

    2024 arXiv

  36. [44]

    Wu, Z. et al. Moleculenet: a benchmark for molecular machine learning. Chem. science 9, 513–530 (2018)

    2018

  37. [45]

    Luo, Y . et al. Biomedgpt: Open multimodal generative pre-trained transformer for biomedicine. arXiv preprint arXiv:2308.09442 (2023)

    2023 arXiv

  38. [46]

    Edwards, C., Zhai, C. & Ji, H. Text2mol: Cross-modal molecule retrieval with natural language queries. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, 595–607 (2021)

    2021

  39. [47]

    & Cho, K

    Kang, S. & Cho, K. Conditional molecular design with deep generative models. J. chemical information modeling 59, 43–52 (2018)

    2018

  40. [48]

    Kim, S. et al. Pubchem substance and compound databases. Nucleic acids research 44, D1202–D1213 (2016)

    2016

  41. [49]

    Irwin, J. J. & Shoichet, B. K. Zinc- a free database of commercially available compounds for virtual screening. J. chemical information modeling 45, 177–182 (2005)

    2005

  42. [50]

    Degtyarenko, K. et al. Chebi: a database and ontology for chemical entities of biological interest. Nucleic acids research 36, D344–D350 (2007). 15/26

    2007

  43. [51]

    Sorokina, M., Merseburger, P., Rajan, K., Yirik, M. A. & Steinbeck, C. Coconut online: collection of open natural products database. J. Cheminformatics 13, 2 (2021)

    2021

  44. [52]

    & Sausville, E

    Monga, M. & Sausville, E. A. Developmental therapeutics program at the nci: molecular target and drug discovery process. Leukemia 16, 520–526 (2002)

    2002

  45. [53]

    D., Mu, Y

    Alhossary, A., Handoko, S. D., Mu, Y . & Kwoh, C.-K. Fast, accurate, and reliable molecular docking with quickvina 2. Bioinformatics 31, 2214–2216 (2015)

    2015

  46. [54]

    R., Baumgartner, M

    Koes, D. R., Baumgartner, M. P. & Camacho, C. J. Lessons learned in empirical scoring with smina from the csar 2011 benchmarking exercise. J. chemical information modeling 53, 1893–1904 (2013)

    2013

  47. [55]

    & Sun, M

    Zeng, Z., Yao, Y ., Liu, Z. & Sun, M. A deep-learning system bridging molecule structure and biomedical text with comprehension comparable to human professionals. Nat. communications 13, 862 (2022)

    2022

  48. [56]

    Su, B. et al. A molecular multimodal foundation model associating molecule graphs with natural language. arXiv preprint arXiv:2209.05481 (2022)

    2022 arXiv

  49. [57]

    Touvron, H. et al. Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288 (2023)

    2023 arXiv

  50. [58]

    Dubey, A. et al. The llama 3 herd of models. arXiv preprint arXiv:2407.21783 (2024)

    2024 arXiv

  51. [59]

    Guo, T. et al. What can large language models do in chemistry? a comprehensive benchmark on eight tasks. Adv. Neural Inf. Process. Syst. 36, 59662–59688 (2023)

    2023

  52. [60]

    Ouyang, L. et al. Training language models to follow instructions with human feedback.Adv. neural information processing systems 35, 27730–27744 (2022)

    2022

  53. [61]

    Fundamentals of recurrent neural network (rnn) and long short-term memory (lstm) network

    Sherstinsky, A. Fundamentals of recurrent neural network (rnn) and long short-term memory (lstm) network. Phys. D: Nonlinear Phenom. 404, 132306 (2020)

    2020

  54. [62]

    Attention is all you need

    Vaswani, A. Attention is all you need. Adv. Neural Inf. Process. Syst. (2017)

    2017

  55. [63]

    Liu, Z. et al. Molxpt: Wrapping molecules with text for generative pre-training. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), 1606–1616 (2023)

    2023

  56. [64]

    & Yang, G

    Yang, R., Zhou, H., Wang, F. & Yang, G. Digfrag as a digital fragmentation method used for artificial intelligence-based drug design. Commun. Chem. 7, 258 (2024)

    2024

  57. [65]

    & Klambauer, G

    Preuer, K., Renz, P., Unterthiner, T., Hochreiter, S. & Klambauer, G. Fréchet chemnet distance: a metric for generative models for molecules in drug discovery. J. chemical information modeling 58, 1736–1741 (2018)

    2018

  58. [66]

    Hu, E. J. et al. Lora: Low-rank adaptation of large language models. arXiv preprint arXiv:2106.09685 (2021)

    2021 arXiv

  59. [67]

    & Kingma, D

    Salimans, T. & Kingma, D. P. Weight normalization: A simple reparameterization to accelerate training of deep neural networks. Adv. neural information processing systems 29 (2016)

    2016

  60. [68]

    P., Salimans, T

    Kingma, D. P., Salimans, T. & Welling, M. Variational dropout and the local reparameterization trick. Adv. neural information processing systems 28 (2015)

    2015

  61. [69]

    & Zhang, H

    Li, Y ., Ma, T. & Zhang, H. Algorithmic regularization in over-parameterized matrix sensing and neural networks with quadratic activations. In Conference On Learning Theory, 2–47 (PMLR, 2018)

    2018

  62. [70]

    Liu, S.-y. et al. Dora: Weight-decomposed low-rank adaptation. In Forty-first International Conference on Machine Learning (2024)

    2024

  63. [71]

    Chen, Z. et al. Multi-granularity score-based generative framework enables efficient inverse design of complex organics. arXiv preprint arXiv:2409.07912 (2024)

    2024 arXiv

  64. [72]

    & Rarey, M

    Degen, J., Wegscheid-Gerlach, C., Zaliani, A. & Rarey, M. On the art of compiling and using’drug-like’chemical fragment spaces. ChemMedChem 3, 1503 (2008)

    2008

  65. [73]

    Q., Judd, D

    Lewell, X. Q., Judd, D. B., Watson, S. P. & Hann, M. M. Recap retrosynthetic combinatorial analysis procedure: a powerful new technique for identifying privileged molecular fragments with useful applications in combinatorial chemistry. J. chemical information computer sciences...

    1998

  66. [74]

    & Liu, Y

    Kong, X., Huang, W., Tan, Z. & Liu, Y . Molecule generation by principal subgraph mining and assembling.Adv. Neural Inf. Process. Syst. 35, 2550–2563 (2022)

    2022

  67. [75]

    Geng, Z. et al. De novo molecular generation via connection-aware motif mining. InThe Eleventh International Conference on Learning Representations (2023). 16/26

    2023

  68. [76]

    Wu, J.-N. et al. t-smiles: a fragment-based molecular representation framework for de novo ligand design. Nat. Commun. 15, 4993 (2024)

    2024

  69. [77]

    Ramesh, A. et al. Zero-shot text-to-image generation. In International conference on machine learning, 8821–8831 (Pmlr, 2021)

    2021

  70. [78]

    What are some of the physical properties of 2-Phenylethylamine?

    Van Den Oord, A., Vinyals, O.et al. Neural discrete representation learning. Adv. neural information processing systems 30 (2017). 17/26 Supplementary Materials Supplementary Section 1: Evaluation of Molecular Encoding Bias We used 2,000 molecules from the test dataset of the ...

    2017

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.