Pith. sign in

REVIEW 5 major objections 5 minor 32 references

BoKDiff: Best-of-K Diffusion Alignment for Target-Specific 3D Molecule Generation

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 ↗ Review JSON

Pith's one-line read A best-of-K fine-tuning loop over a base diffusion model's own samples improves 3D molecule generation for drug targets, reaching -8.58 average Vina docking score and 26% success rate.

desk verdict Useful engineering recipe for aligning DecompDiff, but the evaluation protocol undermines the SOTA claim; the headline numbers are partly selected on the test set. read the letter →

arxiv 2501.15631 v1 pith:EUUYB4Q6 submitted 2025-01-26 q-bio.BM cs.LG

Ali Khodabandeh Yalabadi , Mehdi Yazdani-Jahromi , Ozlem Ozmen Garibay This is my paper · ORCID
classification q-bio.BMcs.LG
keywords 3DdrugdesignBest-of-Nbest-of-Kalignmentdiffusionmodelsstructure-basedmoleculargenerationdockingscorereward-rankedfine-tuning
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper proposes BoKDiff, a framework that takes an existing target-specific 3D molecule diffusion model (DecompDiff) and aligns it to drug-design objectives through reward-ranked fine-tuning. The authors generate several candidate ligands per protein pocket, rank them with a weighted reward of drug-likeness (QED), synthesizability (SA), and a Vina docking score, and fine-tune the base model on the best candidate. Because generated ligands often sit far from the protein in 3D space, which breaks the substructure extraction the model relies on, the paper introduces a data-preparation step that relocates each winning ligand's center of mass to that of its Vina docking pose before re-extracting sub-pockets and substructures. On the CrossDocked2020 benchmark, BoKDiff reports a -8.58 average Vina docking score and a 26% success rate, outperforming the base DecompDiff model (-8.39 and 24.5%). A separate Best-of-N variant, which skips fine-tuning and simply selects the best of N generated samples, reports QED above 0.6, SA above 0.75, and success rates above 35%.

What carries the argument

The central object is the DecompDiff generator, a diffusion model that produces atoms and bonds conditioned on data-dependent priors built from protein sub-pockets and ligand fragments (arms and a scaffold). Around it, BoKDiff wraps a Best-of-K reward-ranked fine-tuning loop (RAFT): sample K candidate ligands per pocket, score them with a weighted reward of QED, SA, and Vina dock, keep the top one, relocate its center of mass to the docking pose's center of mass, re-extract sub-pockets and substructures with adjusted Alphaspace2 parameters, and fine-tune the generator on the curated sample. The relocation is the paper's key adaptation; without it, generated ligands are often too far from the protein for substructure extraction to succeed.

What would settle it

A reader could test the mechanism by taking a fixed batch of generated ligands and comparing substructure-extraction success with and without the center-of-mass relocation; if extraction succeeds at similar rates without relocation, or fails often even with it, the alignment step is not doing the work attributed to it. The paper does not report extraction success rates, only the final aggregate metrics.

Watch

Extended reading notes

Core claim

The central claim is that a reward-ranked fine-tuning loop, adapted from RAFT, can push a strong diffusion generator beyond its original performance on binding and drug-likeness metrics. Concretely, BoKDiff samples K=10 ligands per training pocket, scores them with a weighted reward (with the best configuration found to be QED weight 1, SA weight 0, Vina weight 0), keeps the top-scoring ligand, relocates its center of mass to the center of mass of its Vina docking pose, and fine-tunes DecompDiff for 1,000 iterations at a learning rate of 1e-6 on the resulting curated set. The paper reports that this procedure raises mean Vina Dock from -8.39 to -8.58, success rate from 24.5% to 26.0%, and mean QED from 0.45 to 0.48 relative to DecompDiff, and it argues that the center-of-mass relocation is what makes the fine-tuning signal usable by enabling reliable sub-pocket and substructure extraction. It also claims that a no-fine-tuning Best-of-N sampling strategy, selecting the best of 20 generated samples by the same reward, achieves higher QED, SA, and success rate than the fine-tuned model, at greater inference cost.

Load-bearing premise

The data-preparation step assumes that translating a generated ligand so its center of mass coincides with that of its docking pose preserves the ligand's structural integrity and lets the modified Alphaspace2 extraction reliably identify sub-pockets and substructures; if that alignment corrupts or misassigns components, the fine-tuning signal is degraded and the gains would not generalize.

Editorial extensions

If this is right

  • A base diffusion model for structure-based drug design can be improved on docking and success-rate metrics by fine-tuning on its own best-ranked generated samples, with no change to the model architecture and a fine-tuning cost of roughly 10 minutes on one GPU.
  • Because the optimal reward configuration was QED-only, the paper implies that optimizing drug-likeness alone can transfer to improved Vina metrics and success rate for this model.
  • The Best-of-N results indicate that simply drawing N=20 samples and picking the highest-reward one yields success rates above 35% and QED above 0.65, outperforming the fine-tuned model on those metrics at higher inference cost.
  • The success of the pipeline depends on having a docking pose for each generated sample; every improvement in the paper is therefore conditional on the availability and quality of those poses.

Reading between the lines

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

  • An untested extension is to relocate to multiple docking poses per ligand and aggregate the reward, which the paper lists as future work but does not run; if the relocation mechanism is what matters, pose diversity should sharpen the reward signal.
  • The success-rate threshold comes from Jin et al. and is permissive (QED > 0.25, SA > 0.59, Vina Dock < -8.18); adopting a more stringent threshold could reorder the relative standing of BoKDiff and Best-of-N.
  • The paper's comparison of CoM relocation to the docking pose versus the reference ligand suggests that docking-pose alignment is the more useful signal; a natural next probe is whether partial rotations toward the pose add further gains without breaking substructure extraction.
  • The finding that a QED-only reward also lifts Vina metrics may be specific to DecompDiff's inductive bias (it favors simpler arms), so transferring BoKDiff to a different generator may not reproduce the same correlation.
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 BoKDiff, a fine-tuning framework built on DecompDiff for target-specific 3D molecule generation. BoKDiff applies the RAFT reward-ranked fine-tuning recipe: for each protein in a batch, it generates K ligand samples, ranks them with a weighted sum of QED, SA, and Vina docking score, relocates the best sample's center of mass to that of its docking pose, extracts sub-pockets and substructures from the relocated sample, and fine-tunes the base diffusion model on the curated high-reward set. The paper also evaluates a Best-of-N sampling strategy that skips fine-tuning and directly selects the highest-reward candidate from N samples. The headline claims are a -8.58 average Vina docking score and a 26% success rate on CrossDocked2020, presented as state-of-the-art results, together with a BoN success rate above 35%. The main methodological novelty is the application of best-of-K alignment and BoN sampling to structure-based drug design, plus a center-of-mass relocation step intended to fix ligand-protein misalignment before substructure extraction.

Significance. If the evaluation were sound, this would be a useful and clearly described contribution: it transplants a successful LLM alignment strategy (RAFT) into structure-based drug design, provides a concrete fix for the misalignment problem in decomposed diffusion models, and includes a public code repository, which is valuable for reproducibility. The center-of-mass relocation idea is simple and potentially transferable to other SBDD pipelines. However, the current manuscript does not establish the state-of-the-art claim because the central evaluation protocol selects hyperparameters on the same test set used for final reporting, reports no uncertainty estimates, and relies on small aggregate differences over a single run. The core alignment mechanism may well be sound, but the evidence as presented is not yet sufficient for the claimed strength of the results.

major comments (5)
  1. [Weight Configurations, Tables 4-5] The optimal weight configuration [QED:1, SA:0, Vina:0] is selected by comparing five configurations on 30 samples taken from the test set (Tables 4 and 5), and the same configuration is then used to produce the headline results in Table 1. This is test-set selection: the hyperparameter that drives the -8.58 Vina Dock and 26.0% success rate is chosen by looking at the evaluation proteins. Moreover, the selection signal in Table 4 is very weak, with accumulated mean scores of 1.880 for [1,0,0] versus 1.867 for [1,1,1], a difference of 0.013, and no confidence intervals or significance tests are given. Please re-run the weight selection on a validation split that is disjoint from the reported test set, and report the variability of the selection criterion.
  2. [Ablation Studies, Tables 4-6] The ablation studies are described as 'conducted on 30 test set samples,' and the table captions state that the results use '30 data samples selected from the test set.' Since the BoKDiff pipeline fine-tunes on the collected samples, this means the model is fine-tuned and evaluated on the same test proteins. This contamination undermines the conclusions about weight configurations, batch size, iteration number, and the docking-pose versus reference-ligand relocation comparison. Please use a separate validation set for all ablations and report results on a truly held-out test set.
  3. [Main Results, Table 1] Table 1 reports a single mean and median per metric with no error bars, no per-protein paired analysis, and no statement about the number of random seeds or independent runs. The differences that support the state-of-the-art claim are small: BoKDiff versus DecompDiff gives Vina Dock -8.58 versus -8.39 and success rate 26.0% versus 24.5%. With roughly 1000 generated molecules, these differences may be within sampling noise, especially if docking seeds and generation randomness are not controlled. Please report repeated runs, per-protein paired comparisons, and confidence intervals, and clarify whether the baseline numbers in Table 1 were re-run under identical conditions or taken from prior publications.
  4. [Data Preparation] The center-of-mass relocation step is central to the method, but the paper does not directly validate its key assumption: that translating a generated ligand so that its center of mass coincides with the center of mass of its Vina docking pose preserves the ligand's structural integrity and allows Alphaspace2 to extract the correct sub-pockets and substructures. The initial attempt with raw docking poses is said to alter ligand structure, and the proposed fix avoids that alteration, yet no quantitative evidence is given that the relocated ligand still corresponds to a valid, extractable decomposition. Table 6 compares end metrics for two relocation sources, but it does not show extraction success rates, substructure assignment fidelity, or validity before and after relocation. Please add direct validation of the extraction step.
  5. [Best-of-N, Table 3] The BoN results in Table 3 are also affected by selection on the evaluation data: the text states that the weight set [1.1,1,0.9] 'was identified through a grid search centered around the baseline [1,1,1],' and then Table 3 reports performance for that selected configuration. The claim that BoN surpasses 35% success rate is therefore based on the best configuration found on the same data used to report the result. In addition, the sentence 'other combinations emphasizing SA or Vina Dock are omitted due to their similarity to [1,0,0]' is inconsistent with the table, which includes [1,1,1] and [1.1,1,0.9]. Please clarify the selection procedure and report BoN results on a held-out set.
minor comments (5)
  1. [Title page] The manuscript still contains placeholder header text ('Journal Title Here', 'FOR PUBLISHER ONLY', '© The Author 2022') and several typographical issues such as 'F or permissions' and 'Y ear'; these should be cleaned before submission.
  2. [Data Ranking section] The description of Vina score normalization is underspecified: min-max scaling across the K=10 samples per input depends on the particular set of generated samples and can be unstable when K is small; please state the exact formula and how ties are handled.
  3. [Equations (1)-(3)] The notation for the DecompDiff equations is not fully self-contained: the variables H, Σ, µ, and the tilde notation for prior-centered positions are introduced only briefly. Readers not already familiar with DecompDiff would benefit from explicit definitions.
  4. [Best-of-N section] The abstract and Section 5 both claim QED values exceeding 0.6, SA above 0.75, and success above 35% for BoN, but these numbers only hold for the larger N values and for selected weight sets; the text should make clear that these are properties of a particular, data-dependent configuration rather than of BoN sampling per se.
  5. [Experimental Setup] The implementation details state that fine-tuning is run for '1,000 epochs' but earlier text says '1,000 iterations' and the original model is trained for 500,000 iterations; please use consistent terminology and specify what one epoch or iteration means.

Circularity Check

2 steps flagged · score 6.0 of 10

BoN's headline property results are its own selection objective, and the BoKDiff reward weights were tuned on a subset of the same test set used for the reported SOTA numbers.

  1. self definitional [Best-of-N section and Table 3; Data Ranking (Practical Implementation)]
    "Instead of curating a high-reward dataset by selecting the best-of-K samples and fine-tuning the model based on this dataset, an alternative approach skips the fine-tuning step entirely by directly selecting the best sample from N generated samples. ... the selection of the best sample relies entirely on its reward value. [Table 3 caption:] Comparison of Best-of-N performance on the reference model for various values of N . Results are provided for the weight combination prioritizing QED, while other combinations emphasizing SA or Vina Dock are omitted due to their similarity to [1, 0, 0]."

    The reward used for ranking is explicitly a weighted sum of QED, SA, and Vina Dock. Best-of-N is defined as choosing the candidate that maximizes this reward, and Table 3 then reports QED, SA, and success rate (a threshold over QED/SA/Vina Dock) for the chosen candidate. The improvement of these metrics with N is an order-statistic consequence of maximizing the very objective being reported; no generative model is changed and no independent benchmark is used. The 'exceptional results' in the abstract are therefore the selection criterion by construction, not an external validation of the generative method.

  2. fitted input called prediction [Ablation Studies, Weight Configurations (Tables 4-5) and Main Results, Table 1]
    "These results correspond to the first iteration, using 30 data samples selected from the test set. (Accumulated columns represent the simple summation (with equal weights) of three metrics, where the Vina Dock scores are normalized by dividing them by -10 to ensure they fall within the [0, 1] range.) ... Notably, the best-performing weight configuration was (QED:1, SA:0, Vina Dock:0). [Table 1:] This experiment was conducted using our optimal weight configuration (QED: 1, SA: 0, Vina: 0)."

    The reward-weight hyperparameter controls which generated molecules are selected into the fine-tuning set. It was chosen by comparing accumulated QED/SA/Vina scores on 30 proteins explicitly drawn from the test set, and the headline Table 1 results are then reported on the same test set under that 'optimal' configuration. The configuration is thus a fitted parameter selected from the evaluation data, and the reported state-of-the-art numbers are not an independent out-of-sample prediction; they are the outcome of choosing the hyperparameter that looked best on a subset of the benchmark. The remaining 70 test proteins are not reported separately, so the claimed gain is partly in-sample.

full rationale

The paper's core engineering contribution — CoM relocation plus RAFT-style best-of-K fine-tuning of DecompDiff — is not circular: DecompDiff, RAFT, and BoN are external methods, and the fine-tuned model is assessed with standard affinity and property metrics. There is no self-citation chain and no imported uniqueness theorem. However, two load-bearing numerical claims reduce to their own inputs. First, the abstract's BoN 'exceptional results' (QED>0.6, SA>0.75, success>35%) are reported on the candidate that BoN selects by maximizing a reward that is literally a weighted sum of QED, SA, and Vina Dock; as N increases these numbers improve by the order statistics of the selection objective, so the prediction is the selection criterion by construction. Second, the weight configuration used for the Table 1 BoKDiff state-of-the-art claim was chosen as 'best-performing' from ablations run on 30 proteins drawn from the same test set; the final results are therefore a hyperparameter fit to a subset of the benchmark rather than a held-out evaluation. Neither issue involves self-citation, but both affect the paper's headline quantitative claims. The CoM-relocation and fine-tuning pipeline itself is non-circular and could be validated by a properly held-out protocol.

Assumptions & free parameters 7 free parameters · 4 assumptions · 0 invented entities

The central claim rests on four domain assumptions inherited from the base model, reward proxies, and the RAFT formulation, plus a set of hand-chosen hyperparameters. The most consequential free parameter is the reward weight vector, which is tuned on the evaluation data and directly steers the reported success rate. No new physical or conceptual entities are introduced beyond the CoM relocation preprocessing step, which is a method rather than an entity.

free parameters (7)
  • Reward weights [QED, SA, Vina Dock] = [1, 0, 0] for BoKDiff; [1.1, 1, 0.9] for BoN, selected by grid search
    The weight configuration is called 'optimal' and was identified using evaluation data, directly affecting the reported success rates.
  • K (number of candidates per protein in data collection) = 10
    Chosen by hand; determines which molecule is the 'best' for fine-tuning each iteration.
  • Batch size B for high-reward dataset = 128
    Chosen by hand; ablation compares 128 vs 256 with inconsistent results.
  • Fine-tuning steps = 1000 (called both iterations and epochs)
    Limited to avoid overfitting while preserving base distribution.
  • Learning rate = 1e-6
    Small value chosen to minimize distribution shift.
  • N for Best-of-N sampling = 20
    Selected because performance stabilizes near N=20-30 in Table 3.
  • Cutoff and Radius for Alphaspace2 extraction = Minimum values determined dynamically, exact values not reported
    Dynamic adaptation is described qualitatively; values would be needed to reproduce.
assumptions (4)
  • domain assumption DecompDiff generates valid, diverse ligand candidates from decomposed structural priors.
    The method inherits the pretrained DecompDiff checkpoint and assumes its generated molecules are a suitable starting distribution for reward-based alignment (Practical Implementation, Data Collection).
  • domain assumption QED, SA, and Vina docking score are reliable computational proxies for drug-likeness, synthesizability, and binding affinity.
    The reward function is built from these three metrics (Data Ranking section); the paper cites prior studies for this practice.
  • domain assumption The generative model can achieve the maximum reward for each input protein, so per-input best-of-K selection followed by SFT maximizes expected reward.
    State explicitly in the RAFT section (Equation 5); the convergence of BoKDiff relies on this simplification being approximately true.
  • domain assumption Vina docking poses provide a dependable reference for relocating generated ligands near the binding site.
    The CoM relocation uses the docking pose as the target geometry (Data Preparation), even though the paper notes Vina often alters the ligand structure.

how reviews work

0 comments
Cite this review

Pith. "Pith review of BoKDiff: Best-of-K Diffusion Alignment for Target-Specific 3D Molecule Generation." pith.science (2026). https://pith.science/paper/EUUYB4Q6

@misc{pith2026250115631,
  author       = {Pith},
  title        = {Pith review of: BoKDiff: Best-of-K Diffusion Alignment for Target-Specific 3D Molecule Generation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EUUYB4Q6}},
  note         = {Machine review of arXiv:2501.15631}
}
read the original abstract

Structure-based drug design (SBDD) leverages the 3D structure of biomolecular targets to guide the creation of new therapeutic agents. Recent advances in generative models, including diffusion models and geometric deep learning, have demonstrated promise in optimizing ligand generation. However, the scarcity of high-quality protein-ligand complex data and the inherent challenges in aligning generated ligands with target proteins limit the effectiveness of these methods. We propose BoKDiff, a novel framework that enhances ligand generation by combining multi-objective optimization and Best-of-K alignment methodologies. Built upon the DecompDiff model, BoKDiff generates diverse candidates and ranks them using a weighted evaluation of molecular properties such as QED, SA, and docking scores. To address alignment challenges, we introduce a method that relocates the center of mass of generated ligands to their docking poses, enabling accurate sub-component extraction. Additionally, we integrate a Best-of-N (BoN) sampling approach, which selects the optimal ligand from multiple generated candidates without requiring fine-tuning. BoN achieves exceptional results, with QED values exceeding 0.6, SA scores above 0.75, and a success rate surpassing 35%, demonstrating its efficiency and practicality. BoKDiff achieves state-of-the-art results on the CrossDocked2020 dataset, including a -8.58 average Vina docking score and a 26% success rate in molecule generation. This study is the first to apply Best-of-K alignment and Best-of-N sampling to SBDD, highlighting their potential to bridge generative modeling with practical drug discovery requirements. The code is provided at https://github.com/khodabandeh-ali/BoKDiff.git.

Figures

Figures reproduced from arXiv: 2501.15631 by the authors.

Figure 1
Figure 1. The BoKDiff framework: From left to right, (1) Inputs: Randomly select a batch of the desired size from the training set. (2) Data Collection: Generate K samples for each input pair. (3) Data Ranking: Compute the desired metrics—approximations of QED, SA, and the Vina Docking score—for the generated samples. Rank them using a weighted sum approach; this example emphasizes QED and SA for the final ranking. (4) Data P… view at source ↗
Figure 2
Figure 2. compares the performance of the reference model and aligned models, illustrating that while BoN sampling offers [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 17 canonical work pages

  1. [1]

    The process of structure-based drug design

    Amy C Anderson. The process of structure-based drug design. Chemistry & biology , 10(9):787–797, 2003

    work page 2003

  2. [2]

    Quantifying the chemical beauty of drugs

    G Richard Bickerton, Gaia V Paolini, J´ er´ emy Besnard, Sorel Muresan, and Andrew L Hopkins. Quantifying the chemical beauty of drugs. Nature chemistry, 4(2):90–98, 2012

    work page 2012

  3. [3]

    T raining diffusion models with reinforcement BoKDiff 9 learning

    Kevin Black, Michael Janner, Yilun Du, Ilya Kostrikov, and Sergey Levine. T raining diffusion models with reinforcement BoKDiff 9 learning. arXiv preprint arXiv:2305.13301 , 2023

    arXiv 2023

  4. [4]

    Deep reinforcement learning from human preferences

    Paul F Christiano, Jan Leike, Tom Brown, Miljan Martic, Shane Legg, and Dario Amodei. Deep reinforcement learning from human preferences. Advances in neural information processing systems, 30, 2017

    2017

  5. [5]

    Raft: Reward ranked finetuning for generative foundation model alignment

    Hanze Dong, Wei Xiong, Deepanshu Goyal, Yihan Zhang, Winnie Chow, Rui Pan, Shizhe Diao, Jipeng Zhang, Kashun Shum, and Tong Zhang. Raft: Reward ranked finetuning for generative foundation model alignment. arXiv preprint arXiv:2304.06767, 2023

    arXiv 2023

  6. [6]

    Autodock vina 1.2

    Jerome Eberhardt, Diogo Santos-Martins, Andreas F Tillack, and Stefano Forli. Autodock vina 1.2. 0: New docking methods, expanded force field, and python bindings. Journal of chemical information and modeling , 61(8):3891–3898, 2021

    work page 2021

  7. [7]

    Estimation of synthetic accessibility score of drug-like molecules based on molecular complexity and fragment contributions

    Peter Ertl and Ansgar Schuffenhauer. Estimation of synthetic accessibility score of drug-like molecules based on molecular complexity and fragment contributions. Journal of cheminformatics , 1:1–11, 2009

    work page 2009

  8. [8]

    Three-dimensional convolutional neural networks and a cross-docked data set for structure-based drug design

    Paul G Francoeur, Tomohide Masuda, Jocelyn Sunseri, Andrew Jia, Richard B Iovanisci, Ian Snyder, and David R Koes. Three-dimensional convolutional neural networks and a cross-docked data set for structure-based drug design. Journal of chemical information and modeling , 60(9):4200–4215, 2020

    work page 2020

Show all 32 references
  1. [9]

    Reinforced genetic algorithm for structure-based drug design

    Tianfan Fu, Wenhao Gao, Connor Coley, and Jimeng Sun. Reinforced genetic algorithm for structure-based drug design. Advances in Neural Information Processing Systems, 35:12325–12338, 2022

    2022

  2. [10]

    3d equivariant diffusion for target-aware molecule generation and affinity prediction

    Jiaqi Guan, Wesley Wei Qian, Xingang Peng, Yufeng Su, Jian Peng, and Jianzhu Ma. 3d equivariant diffusion for target-aware molecule generation and affinity prediction. arXiv preprint arXiv:2303.03543 , 2023

    2023 arXiv

  3. [11]

    Decompdiff: diffusion models with decomposed priors for structure-based drug design

    Jiaqi Guan, Xiangxin Zhou, Yuwei Yang, Yu Bao, Jian Peng, Jianzhu Ma, Qiang Liu, Liang Wang, and Quanquan Gu. Decompdiff: diffusion models with decomposed priors for structure-based drug design. arXiv preprint arXiv:2403.07902, 2024

    2024 arXiv

  4. [12]

    Bonbon alignment for large language models and the sweetness of best-of-n sampling

    Lin Gui, Cristina Gˆ arbacea, and Victor Veitch. Bonbon alignment for large language models and the sweetness of best-of-n sampling. arXiv preprint arXiv:2406.00832 , 2024

    2024 arXiv

  5. [13]

    Multi-objective molecule generation using interpretable substructures

    Wengong Jin, Regina Barzilay, and Tommi Jaakkola. Multi-objective molecule generation using interpretable substructures. In International conference on machine learning, pages 4849–4859. PMLR, 2020

    2020

  6. [14]

    Diffbp: Generative diffusion of 3d molecules for target protein binding

    Haitao Lin, Yufei Huang, Meng Liu, Xuanjing Li, Shuiwang Ji, and Stan Z Li. Diffbp: Generative diffusion of 3d molecules for target protein binding. arXiv preprint arXiv:2211.11214, 2022

    2022 arXiv

  7. [15]

    Generating 3d molecules for target protein binding

    Meng Liu, Youzhi Luo, Kanji Uchino, Koji Maruhashi, and Shuiwang Ji. Generating 3d molecules for target protein binding. arXiv preprint arXiv:2204.09410 , 2022

    2022 arXiv

  8. [16]

    A 3d generative model for structure-based drug design

    Shitong Luo, Jiaqi Guan, Jianzhu Ma, and Jian Peng. A 3d generative model for structure-based drug design. Advances in Neural Information Processing Systems , 34:6229–6239, 2021

    2021

  9. [17]

    Training language models to follow instructions with human feedback

    Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, et al. Training language models to follow instructions with human feedback. Advances in neural information processing systems, 35:2773...

    2022

  10. [18]

    Pocket2mol: Efficient molecular sampling based on 3d protein pockets

    Xingang Peng, Shitong Luo, Jiaqi Guan, Qi Xie, Jian Peng, and Jianzhu Ma. Pocket2mol: Efficient molecular sampling based on 3d protein pockets. In International Conference on Machine Learning , pages 17644–17655. PMLR, 2022

    2022

  11. [19]

    Geometric deep learning for structure-based ligand design

    Alexander S Powers, Helen H Yu, Patricia Suriana, Rohan V Koodli, Tianyu Lu, Joseph M Paggi, and Ron O Dror. Geometric deep learning for structure-based ligand design. ACS Central Science, 9(12):2257–2267, 2023

    2023

  12. [20]

    Direct preference optimization: Your language model is secretly a reward model

    Rafael Rafailov, Archit Sharma, Eric Mitchell, Christopher D Manning, Stefano Ermon, and Chelsea Finn. Direct preference optimization: Your language model is secretly a reward model. Advances in Neural Information Processing Systems, 36, 2024

    2024

  13. [21]

    Generating 3d molecules conditional on receptor binding sites with deep generative models

    Matthew Ragoza, Tomohide Masuda, and David Ryan Koes. Generating 3d molecules conditional on receptor binding sites with deep generative models. Chemical science, 13(9):2701–2713, 2022

    2022

  14. [22]

    Structure-based drug design with equivariant diffusion models

    Arne Schneuing, Yuanqi Du, Charles Harris, Arian Jamasb, Ilia Igashov, Weitao Du, Tom Blundell, Pietro Li´ o, Carla Gomes, Max Welling, et al. Structure-based drug design with equivariant diffusion models. arXiv preprint arXiv:2210.13695, 2022

    arXiv

  15. [23]

    Proximal policy optimization algorithms

    John Schulman, Filip Wolski, Prafulla Dhariwal, Alec Radford, and Oleg Klimov. Proximal policy optimization algorithms. arXiv preprint arXiv:1707.06347 , 2017

    2017 arXiv

  16. [24]

    Autogrow4: an open- source genetic algorithm for de novo drug design and lead optimization

    Jacob O Spiegel and Jacob D Durrant. Autogrow4: an open- source genetic algorithm for de novo drug design and lead optimization. Journal of cheminformatics , 12:1–16, 2020

    2020

  17. [25]

    Applications of machine learning in drug discovery and development

    Jessica Vamathevan, Dominic Clark, Paul Czodrowski, Ian Dunham, Edgardo Ferran, George Lee, Bin Li, Anant Madabhushi, Parantu Shah, Michaela Spitzer, et al. Applications of machine learning in drug discovery and development. Nature reviews Drug discovery , 18(6):463– 477, 2019

    2019

  18. [26]

    Diffusion model alignment using direct preference optimization

    Bram Wallace, Meihua Dang, Rafael Rafailov, Linqi Zhou, Aaron Lou, Senthil Purushwalkam, Stefano Ermon, Caiming Xiong, Shafiq Joty, and Nikhil Naik. Diffusion model alignment using direct preference optimization. In Proceedings of the IEEE/CVF Conference on Computer Vision and...

    2024

  19. [27]

    Large-scale reinforcement learning for diffusion models

    Yinan Zhang, Eric Tzeng, Yilun Du, and Dmitry Kislyuk. Large-scale reinforcement learning for diffusion models. arXiv preprint arXiv:2401.12244 , 4, 2024

    2024 arXiv

  20. [28]

    Learning subpocket prototypes for generalizable structure-based drug design

    Zaixi Zhang and Qi Liu. Learning subpocket prototypes for generalizable structure-based drug design. In International Conference on Machine Learning , pages 41382–41398. PMLR, 2023

    2023

  21. [29]

    Molecule generation for target protein binding with structural motifs

    Zaixi Zhang, Yaosen Min, Shuxin Zheng, and Qi Liu. Molecule generation for target protein binding with structural motifs. In The Eleventh International Conference on Learning Representations, 2023

    2023

  22. [30]

    Decompopt: Controllable and decomposed diffusion models for structure-based molecular optimization

    Xiangxin Zhou, Xiwei Cheng, Yuwei Yang, Yu Bao, Liang Wang, and Quanquan Gu. Decompopt: Controllable and decomposed diffusion models for structure-based molecular optimization. arXiv preprint arXiv:2403.13829 , 2024

    2024 arXiv

  23. [31]

    Antigen-specific antibody design via direct energy-based preference optimization

    Xiangxin Zhou, Dongyu Xue, Ruizhe Chen, Zaixiang Zheng, Liang Wang, and Quanquan Gu. Antigen-specific antibody design via direct energy-based preference optimization. arXiv preprint arXiv:2403.16576 , 2024

    2024 arXiv

  24. [32]

    Fine-tuning language models from human preferences

    Daniel M Ziegler, Nisan Stiennon, Jeffrey Wu, Tom B Brown, Alec Radford, Dario Amodei, Paul Christiano, and Geoffrey Irving. Fine-tuning language models from human preferences. arXiv preprint arXiv:1909.08593 , 2019

    1909 arXiv

Pith tools

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