Pith. sign in

REVIEW 4 major objections 7 minor 1 cited by

Efficient Fine-Tuning of Single-Cell Foundation Models Enables Zero-Shot Molecular Perturbation Prediction

T0 review · 4 major / 7 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Frozen single-cell foundation models can predict drug responses in unseen cell lines with a drug-conditional adapter.

desk verdict Useful adapter method, but the DEG selection is underspecified and every headline R2 depends on it; the zero-shot claim is also thin at three cell lines. read the letter →

arxiv 2412.13478 v2 pith:KSPESTMP submitted 2024-12-18 cs.LG q-bio.QM

Sepideh Maleki , Jan-Christian Huetter , Kangway V. Chuang , David Richmond , Gabriele Scalia , Tommaso Biancalani This is my paper · ORCID
classification cs.LGq-bio.QM MSC 68T0792C40
keywords single-cellfoundationmodelsdrug-conditionaladapterperturbationpredictionparameter-efficientfine-tuningzero-shotgeneralizationscGPTChemBERTasci-Plex3
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 aims to show that a single-cell foundation model can be adapted to predict transcriptional responses to chemical perturbations by freezing the foundation model and training only small drug-conditioned adapter layers. The proposed method, scDCA, injects a molecule embedding into every transformer layer as biases for down- and up-projections, so the model stays in the gene-expression input space seen during pre-training. The authors report state-of-the-art R2 on sci-Plex3 across four generalization tasks, with the largest gains on unseen-cell-line zero- and few-shot settings, while training less than 1% of the original model parameters. If correct, this gives a practical way to extend pre-trained cell atlases to virtual screening and new biological contexts with very little paired perturbation data.

What carries the argument

The drug-conditional adapter. In each of the 12 transformer layers of scGPT, a two-layer network $f^m_l$ maps a frozen ChemBERTa molecule embedding $\mathrm{emb}_m(d)$ to a bias vector $b_l$. The adapter transforms the layer hidden state as $h^{\mathrm{down}}_l = W^{\mathrm{down}}_l h_l + \Pi^{\mathrm{down}} b_l$, passes it through a residual network, and outputs $h^{\mathrm{up}}_l = W^{\mathrm{up}}_l h^{\mathrm{res\_net}}_l + b_l$; the bias is what carries the drug signal into every layer. Original scGPT weights, gene tokens, expression embedding, and molecule encoder are frozen, so the input distribution stays close to pre-training and only the adapter parameters are learned, which the paper reports as less than 1% of the original model.

What would settle it

Recompute R2 on a fixed gene set or on a DEG set chosen a priori from control versus treated data without using test labels, and check whether scDCA still beats the baselines on zero-shot cell lines; the current reported numbers cannot discriminate because the DEG-selection step is unspecified.

Watch

Extended reading notes

Core claim

The central claim is that molecular conditioning can be added to a frozen single-cell transformer through trainable adapter layers whose biases are generated from a molecular embedding, and that this preserves the pre-trained biological representation well enough to predict perturbation outcomes for drugs and cell lines never seen during fine-tuning. On the sci-Plex3 dataset (A549, MCF7, K562; 188 drugs), scDCA reports R2 of 0.81 for unseen drugs, 0.83 for unseen drug-cell-line combinations, 0.88 for few-shot unseen cell lines, and 0.82 for zero-shot unseen cell lines, outperforming full fine-tuning of scGPT and the ChemCPA, BioLORD, and SAMS-VAE baselines, especially in zero-shot cell-line generalization. The authors frame the result as evidence that careful parameter-efficient fine-tuning, not naive fine-tuning or feature extraction, is what unlocks foundation-model knowledge for molecular perturbation prediction.

Load-bearing premise

All reported R2 values are computed on the top 20 differentially expressed genes, but the paper never specifies how those genes are selected; if the selection uses the test perturbation labels, the reported performance advantage could be an artifact of the chosen gene subset rather than a property of the model.

Editorial extensions

If this is right

  • A frozen single-cell foundation model plus small adapters can generalize to entirely unseen cell lines in zero-shot mode, provided control gene expression for the new cell line is available.
  • Because the adapter accepts any molecular embedding, the same recipe extends to other molecule encoders without retraining the single-cell foundation model.
  • The performance gap over baselines grows as the task becomes harder, suggesting the pre-trained representation is the main carrier of generalization to new biological contexts.
  • The approach is limited to transformer-based foundation models and requires control expression data, so it cannot be applied directly to datasets lacking controls or to non-transformer architectures.

Reading between the lines

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

  • If DEG selection is made label-free and fixed in advance, the zero-shot advantage may shrink or vanish; the paper's load-bearing comparison should be re-run under such a protocol.
  • The same bias-from-another-modality adapter could condition single-cell foundation models on other inputs, such as genetic perturbations or growth conditions, since the machinery does not depend on chemistry-specific details.
  • With only three cell lines, leave-one-out results are a weak test of unseen-cell-line generalization; validation on more diverse cell lines would show whether the effect holds beyond adenocarcinoma and leukemia splits.
  • The authors' own appendix shows performance varies by held-out line, so practical deployment would need a way to estimate which new cell lines are likely to be predictable.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 7 minor

Summary. The paper introduces scDCA, a parameter-efficient fine-tuning strategy that couples a frozen single-cell foundation model (scGPT) with drug-conditional adapter layers. The molecule embedding from ChemBERTa is injected into each transformer layer via learned biases, while the original scGPT weights remain frozen, so that less than 1% of the backbone parameters are trainable. The method is evaluated on sci-Plex 3 in four settings: unseen drugs, unseen drug-cell-line combinations, few-shot unseen cell lines, and zero-shot unseen cell lines, with R2 computed on the top 20 differentially expressed genes. The authors report that scDCA outperforms ChemCPA, BioLORD, and SAMS-VAE across all tasks, with the largest gains in zero-shot and few-shot cell-line generalization, and support this with ablations, target-robustness analyses, and paired t-tests.

Significance. If the reported results are credible, the paper makes a useful contribution by showing that a cross-modal adapter, rather than full fine-tuning, can adapt a single-cell foundation model to chemical perturbation prediction and can generalize to cell lines not seen during fine-tuning. The study is reasonably broad: it compares several recent baselines, includes ablations of the adapter design, examines robustness across drug target clusters, and reports per-cell-line zero-shot results. The main barrier to accepting the central claim is that the evaluation protocol for selecting the genes on which all R2 metrics are computed is underspecified, and the zero-shot claim rests on only three cell lines. These issues are fixable with additional clarification and experiments, so the work warrants a major revision rather than rejection.

major comments (4)
  1. [Section 4.1, Evaluation Metrics] The protocol for selecting the top 20 differentially expressed genes (DEGs) is never specified. The text states only that the analysis 'targets those genes that exhibit noticeable changes in their expression levels compared to their initial (control) expression profiles' and that 'we consider the top 20 DEGs,' but it does not say whether the DEGs are chosen per drug, per cell line, or globally, nor whether the selection uses training-set labels only, the full dataset, or the test perturbation outcomes. Because every reported R2 value in Figure 3, Tables 1-3, and the paired t-tests in Appendix A.8 is computed on this subset, an outcome-dependent DEG choice could inflate and re-rank all models. The authors must specify the exact DEG-selection algorithm and report results on the full 2,000-gene set or on a pre-registered, training-only DEG criterion, so that the state-of-the-art claim can be verified.
  2. [Appendix A.9 and Section 4.1] The zero-shot cell-line evaluation uses leave-one-out over only three cell lines (A549, K562, MCF7), but the paper reports means and standard errors over five runs with 'different random splits.' With three cell lines there are only three possible leave-one-out splits, so it is unclear what the five runs represent and what the degrees of freedom are for the paired t-tests in Tables 4-7. The authors should clarify the run structure and report per-split results for every method, including the fine-tuning and ablation baselines, so that the statistical significance claims are interpretable.
  3. [Section 3.1 and Section 4.2, unseen cell line tasks] The claim of zero-shot generalization to 'unseen cell lines' is weakened by the fact that the scGPT pretraining corpus contains many human cell types and likely includes the A549, MCF7, and K562 cell lines. The model has therefore seen these biological contexts during pretraining, even if they are absent from the fine-tuning data. The manuscript should explicitly acknowledge this, quantify the overlap with pretraining data if possible, and ideally validate the method on a cell type or tissue that is genuinely absent from the pretraining corpus, to support the stronger novelty claim.
  4. [Section 4.2, Figure 4] The claim that 'scDCA predictions are within single-cell measurement uncertainty' is supported only by two example molecules (Quisinostat and Dacinostat) in Figure 4, with no quantitative definition of the uncertainty range and no aggregate statistics across the test set. The authors should define the measurement-noise threshold explicitly and report, for example, the fraction of test drugs and genes for which the prediction error lies within one or two standard deviations of the single-cell distribution, rather than relying on two selected examples.
minor comments (7)
  1. [Section 2, Related Work] In the first paragraph, the reference 'e.g., (Roohani et al., 2024)' appears mid-sentence without a closing period, and the quotation mark after '10^60' is unbalanced; these should be corrected.
  2. [Section 3.1, Preliminaries] The phrase 'as input for for all models considered here' contains a duplicated 'for'; please fix the typo.
  3. [Section 4.2, Table 1] The text says scDCA 'consistently outperforms the naive finetuning approach across all tasks,' but the unseen-drug row shows identical means (0.81 for both methods). Please qualify the claim or explain why the difference is not reflected in the reported means.
  4. [Appendix A.2, Preprocessing] The notation X^(0)(d) for the control expression vector should likely be X^(0)(c), since the control profile depends on the cell line, not the drug; please correct the notation.
  5. [Appendix A.6, Additional discussion] The sentence 'ensuring that all annotated compounds are reserved for the test set' suggests that no annotated compound appears in the training set, but the target-cluster evaluation then has no in-training examples for those targets. Please clarify how the training set is formed for this experiment.
  6. [Appendix A.7, Figure 8] The DEG-size sensitivity analysis in Figure 8 does not state whether the same unspecified DEG-selection criterion is used for each size; the figure is difficult to interpret without this information.
  7. [Appendix A.8, Paired t-tests] The summary sentence says scDCA improves over 'all baselines' for all tasks, but the zero-shot table (Table 7) contains no ChemCPA row because ChemCPA cannot run that task; please rephrase to reflect the actual comparisons made.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: scDCA is trained and evaluated on held-out perturbation data; no equation defines the prediction from its target.

full rationale

The derivation chain is self-contained. scDCA's forward pass (Eqs. 6-11) maps control expression X^(0)(c), gene tokens, and a frozen ChemBERTa drug embedding to a predicted perturbed expression, and Eq. 12 trains it against X^(d)(c) with MSE on train splits. The test targets are held out by the four split protocols in Sec. 4.1 and App. A.1, so the reported R2 values are empirical generalizations, not identities. The 'zero-shot unseen cell line' setting uses only the control expression of the held-out cell line as input, which the paper explicitly states in Sec. 3.1 and Figure 2d; using control measurements as features is not equivalent to using the perturbation outcome as an input. Baselines (ChemCPA, BioLORD, SAMS-VAE) are external published methods, and the comparisons are not underwritten by any self-citation: the paper cites scGPT and ChemBERTa as external frozen components, and neither is tuned on the perturbation targets in a way that encodes the claimed result. The main unresolved concerns---the top-20 DEG evaluation set is not fully specified (Sec. 4.1), only three cell lines are available for zero-shot claims (Sec. 4.2 and App. A.9), and scGPT's pretraining corpus may overlap with the test cell lines---are validity and robustness risks, not circularity: none of them makes the predicted quantity equal to an input or to a fitted parameter by construction. The conclusion and limitations (Sec. 5) are candid about scope but do not smuggle in the target. Accordingly, no step satisfies the standard for a circularity finding, and the score is 0.

Assumptions & free parameters 5 free parameters · 4 assumptions · 1 invented entities

The central claim is an empirical performance claim. The model relies on pre-trained scGPT and ChemBERTa as frozen priors. The evaluation relies on the sciplex3 dataset and a hand-selected DEG subset. No new physical entities are introduced beyond the adapter module.

free parameters (5)
  • Number of DEGs = 20
    The reported R2 is computed only on the top 20 differentially expressed genes. The selection procedure is not specified, and Appendix A.7 shows performance varies with this number, so it directly modulates the headline results.
  • Learning rate = 1e-4
    Chosen by hand for training the adapters (Appendix A.2).
  • Batch size = 16
    Chosen by hand (Appendix A.2).
  • Number of epochs = 20
    Chosen by hand (Appendix A.2).
  • Number of runs = 5
    All results are averages over 5 random splits; the small number of runs limits the statistical power of the paired t-tests.
assumptions (4)
  • domain assumption scGPT pre-training on tens of millions of cells provides a useful biological prior for perturbation prediction.
    The method's success depends on the frozen scGPT representations being informative for the perturbation task (Section 3.1).
  • domain assumption ChemBERTa molecular embeddings capture structure-response relationships relevant to transcriptional outcomes.
    Drugs are represented exclusively through frozen ChemBERTa embeddings (Section 3.1).
  • domain assumption The sciplex3 dataset with three cell lines is sufficient to evaluate generalization to unseen cell lines.
    Zero-shot and few-shot cell line conclusions are drawn from leave-one-cell-line-out experiments (Appendix A.9).
  • domain assumption The top 20 DEGs are selected without using test labels.
    The paper does not describe the DEG selection procedure, so it is assumed to be label-free. If violated, the reported R2 values are inflated.
invented entities (1)
  • Drug-Conditional Adapter (scDCA) independent evidence
    purpose: Injects molecular embeddings into frozen scGPT transformer layers via bias vectors from a small projection network.
    The adapter is evaluated on held-out drugs and cell lines in sciplex3, providing external evidence for its utility.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Efficient Fine-Tuning of Single-Cell Foundation Models Enables Zero-Shot Molecular Perturbation Prediction." pith.science (2026). https://pith.science/paper/KSPESTMP

@misc{pith2026241213478,
  author       = {Pith},
  title        = {Pith review of: Efficient Fine-Tuning of Single-Cell Foundation Models Enables Zero-Shot Molecular Perturbation Prediction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KSPESTMP}},
  note         = {Machine review of arXiv:2412.13478}
}
read the original abstract

Predicting transcriptional responses to novel drugs provides a unique opportunity to accelerate biomedical research and advance drug discovery efforts. However, the inherent complexity and high dimensionality of cellular responses, combined with the extremely limited available experimental data, makes the task challenging. In this study, we leverage single-cell foundation models (FMs) pre-trained on tens of millions of single cells, encompassing multiple cell types, states, and disease annotations, to address molecular perturbation prediction. We introduce a drug-conditional adapter that allows efficient fine-tuning by training less than 1% of the original foundation model, thus enabling molecular conditioning while preserving the rich biological representation learned during pre-training. The proposed strategy allows not only the prediction of cellular responses to novel drugs, but also the zero-shot generalization to unseen cell lines. We establish a robust evaluation framework to assess model performance across different generalization tasks, demonstrating state-of-the-art results across all settings, with significant improvements in the few-shot and zero-shot generalization to new cell lines compared to existing baselines.

Figures

Figures reproduced from arXiv: 2412.13478 by the authors.

Figure 1
Figure 1. scDCA architecture. On the upper left, we show the input embedding to scGPT, which consists of gene tokens and unperturbed gene expressions. Input is passed through scGPT, which consists of stacked transformer blocks, where each layer incorporates a drug-conditional adapter module. Primary goal of this adapter is to introduce parameter-efficient fine-tuning by leveraging molecule embeddings to dynamically adjust bia… view at source ↗
Figure 2
Figure 2. Problem formulation. Rows represents different cell lines and columns different drugs. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Comparison with different baselines. scDCA is our proposed method. X-axis represents [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Examples of predicted gene expression across 20 most differentially expressed genes for [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Clustering drugs based on their targets. X-axis represents targets. [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 7
Figure 7. Figure 7: 16 [PITH_FULL_IMAGE:figures/full_fig_p016_7.png]
Figure 6
Figure 6. Figure 6: Examples of predicted gene expression across 20 most differentially expressed genes for [PITH_FULL_IMAGE:figures/full_fig_p017_6.png]
Figure 7
Figure 7. Figure 7: Clustering drugs based on their targets. X-axis represents targets. [PITH_FULL_IMAGE:figures/full_fig_p017_7.png]
Figure 8
Figure 8. Figure 8: R2 values for different tasks on different numbers of DEGs . A.8 PAIRED T-TESTS RESULTS OF SCDCA VERSUS THE OTHER MODELS We performed paired t-tests to compare the performance of scDCA against ChemCPA, BioLORD, and SAMS_VAE across multiple tasks. The results are summar…
Figure 9
Figure 9. Figure 9: Similarity of unseen drugs. Histogram of maximum Tanimoto similarity to the training set [PITH_FULL_IMAGE:figures/full_fig_p019_9.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Multi-modal single-cell foundation models via dynamic token adaptation

    q-bio.GN 2025-04 conditional novelty 7.0 of 10

    The authors introduce dynamic token adaptation to combine DNA language models with single-cell foundation models, and show that mutating GATA4's promoter in silico shifts predicted target gene embeddings in fetal card...

Reference graph

Works this paper leans on

48 extracted references · 33 canonical work pages · cited by 1 Pith paper

  1. [1]

    Modelling cellular perturbations with the sparse additive mechanism shift variational autoencoder

    Michael Bereket and Theofanis Karaletsos. Modelling cellular perturbations with the sparse additive mechanism shift variational autoencoder. Advances in Neural Information Processing Systems, 36, 2024

    2024

  2. [2]

    The art and practice of structure-based drug design: a molecular modeling perspective

    Regine S Bohacek, Colin McMartin, and Wayne C Guida. The art and practice of structure-based drug design: a molecular modeling perspective. Medicinal research reviews, 16 0 (1): 0 3--50, 1996

    work page 1996

  3. [3]

    Hudson, Ehsan Adeli, Russ Altman, Simran Arora, Sydney von Arx, Michael S

    Rishi Bommasani, Drew A. Hudson, Ehsan Adeli, Russ Altman, Simran Arora, Sydney von Arx, Michael S. Bernstein, Jeannette Bohg, Antoine Bosselut, Emma Brunskill, Erik Brynjolfsson, Shyamal Buch, Dallas Card, Rodrigo Castellon, Niladri Chatterji, Annie Chen, Kathleen Creel, Jared Quincy Davis, Dora Demszky, Chris Donahue, Moussa Doumbouya, Esin Durmus, Stef...

    arXiv 2022

  4. [4]

    How to build the virtual cell with artificial intelligence: Priorities and opportunities

    Charlotte Bunne, Yusuf Roohani, Yanay Rosen, Ankit Gupta, Xikun Zhang, Marcel Roed, Theo Alexandrov, Mohammed AlQuraishi, Patricia Brennan, Daniel B Burkhardt, et al. How to build the virtual cell with artificial intelligence: Priorities and opportunities. arXiv preprint arXiv:2409.11654, 2024

    arXiv 2024

  5. [5]

    Chemberta: Large-scale self-supervised pretraining for molecular property prediction

    Seyone Chithrananda, Gabriel Grand, and Bharath Ramsundar. Chemberta: Large-scale self-supervised pretraining for molecular property prediction. arXiv, 2020

    work page 2020

  6. [6]

    scgpt: toward building a foundation model for single-cell multi-omics using generative ai

    Haoran Cui, Chen Wang, Haroon Maan, et al. scgpt: toward building a foundation model for single-cell multi-omics using generative ai. Nature Methods, 21 0 (12): 0 1470--1480, 2024

    work page 2024

  7. [7]

    The chemfp project

    Andrew Dalke. The chemfp project. Journal of cheminformatics, 11: 0 1--21, 2019

    work page 2019

  8. [8]

    Bert: Pre-training of deep bidirectional transformers for language understanding

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. Bert: Pre-training of deep bidirectional transformers for language understanding. In North American Chapter of the Association for Computational Linguistics, 2019

    2019

Show all 48 references
  1. [9]

    Chembl: a large-scale bioactivity database for drug discovery

    Anna Gaulton, Louisa J Bellis, A Patr \' cia Bento, Jon Chambers, Mark Davies, Anne Hersey, Yvonne Light, Sean McGlinchey, David Michalovich, Bissan Al-Lazikani, and John P Overington. Chembl: a large-scale bioactivity database for drug discovery. Nucleic acids research, 40, 2012

    2012

  2. [10]

    Construction of a human cell landscape at single-cell level

    Xiaoping Han, Ziming Zhou, Lijiang Fei, Huiyu Sun, Renying Wang, Yao Chen, Haide Chen, Jingjing Wang, Huanna Tang, Wenhao Ge, et al. Construction of a human cell landscape at single-cell level. Nature, 581 0 (7808): 0 303--309, 2020

    2020

  3. [11]

    Parameter-efficient fine-tuning for large models: A comprehensive survey

    Zeyu Han, Chao Gao, Jinyang Liu, Sai Qian Zhang, et al. Parameter-efficient fine-tuning for large models: A comprehensive survey. Nature Machine Intelligence, 5: 0 220--235, 2023

    2023

  4. [12]

    Large-scale foundation model on single-cell transcriptomics

    Minsheng Hao, Jing Gong, Xin Zeng, Chiming Liu, Yucheng Guo, Xingyi Cheng, Taifeng Wang, Jianzhu Ma, Xuegong Zhang, and Le Song. Large-scale foundation model on single-cell transcriptomics. Nature Methods, pp.\ 1--11, 2024

    2024

  5. [13]

    Towards a unified view of parameter-efficient transfer learning

    Junxian He, Chunting Zhou, Xuezhe Ma, Taylor Berg-Kirkpatrick, and Graham Neubig. Towards a unified view of parameter-efficient transfer learning. In International Conference on Learning Representations, 2022

    2022

  6. [14]

    DePianto, et al

    Gabriel Heimberg, Ta-Chun Kuo, Dennis J. DePianto, et al. A cell atlas foundation model for scalable search of similar human cells. Nature, 2024

    2024

  7. [15]

    Predicting cellular responses to novel drug perturbations at a single-cell resolution

    Leon Hetzel, Simon Böhm, Niki Kilbertus, Stephan Günnemann, Mohammad Lotfollahi, and Fabian J Theis. Predicting cellular responses to novel drug perturbations at a single-cell resolution. In NeurIPS 2022, 2022

    2022

  8. [16]

    Parameter-efficient transfer learning for nlp

    Neil Houlsby, Andrei Giurgiu, Stanislaw Jastrzebski, Bruna Morrone, Quentin de Laroussilhe, Andrea Gesmundo, Mona Attariyan, and Sylvain Gelly. Parameter-efficient transfer learning for nlp. In Proceedings of the 36th International Conference on Machine Learning, 2019

    2019

  9. [17]

    Fine-tuned language models for text classification

    Jeremy Howard and Sebastian Ruder. Fine-tuned language models for text classification. arXiv preprint arXiv:1801.06146, 2018

    2018 arXiv

  10. [18]

    Lo RA : Low-rank adaptation of large language models

    Edward J Hu, yelong shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. Lo RA : Low-rank adaptation of large language models. In International Conference on Learning Representations, 2022

    2022

  11. [19]

    Few-shot molecular property prediction via hierarchically structured learning on relation graphs

    Wei Ju, Zequn Liu, Yifang Qin, Bin Feng, Chen Wang, Zhihui Guo, Xiao Luo, and Ming Zhang. Few-shot molecular property prediction via hierarchically structured learning on relation graphs. Neural Networks, 163: 0 122--131, 2023

    2023

  12. [20]

    Parameter-efficient multi-task fine-tuning for transformers via shared hypernetworks

    Rabeeh Karimi Mahabadi, Sebastian Ruder, Mostafa Dehghani, and James Henderson. Parameter-efficient multi-task fine-tuning for transformers via shared hypernetworks. In Chengqing Zong, Fei Xia, Wenjie Li, and Roberto Navigli (eds.), Proceedings of the 59th Annual Meeting of th...

    2021

  13. [21]

    Pubchem 2023 update

    Sunghwan Kim, Jie Chen, Tiejun Cheng, Asta Gindulyte, Jia He, Siqian He, Qingliang Li, Benjamin A Shoemaker, Paul A Thiessen, Bo Yu, et al. Pubchem 2023 update. Nucleic acids research, 51 0 (D1): 0 D1373--D1380, 2023

    2023

  14. [22]

    Disentanglement via mechanism sparsity regularization: A new principle for nonlinear ica

    Sébastien Lachapelle, Pau Rodríguez López, Yash Sharma, Katie Everett, Rémi Le Priol, Alexandre Lacoste, and Simon Lacoste-Julien. Disentanglement via mechanism sparsity regularization: A new principle for nonlinear ica. Causal Learning and Reasoning, 2022

    2022

  15. [23]

    Neural architecture search for parameter-efficient fine-tuning of large pre-trained language models

    Neal Lawton, Anoop Kumar, Govind Thattai, Aram Galstyan, and Greg Ver Steeg. Neural architecture search for parameter-efficient fine-tuning of large pre-trained language models. In Anna Rogers, Jordan Boyd-Graber, and Naoaki Okazaki (eds.), Findings of the Association for Comp...

    2023

  16. [24]

    Conditional adapters: Parameter-efficient transfer learning with fast inference

    Tao Lei, Junwen Bai, Siddhartha Brahma, Joshua Ainslie, Kenton Lee, Yanqi Zhou, Nan Du, Vincent Zhao, Yuexin Wu, Bo Li, et al. Conditional adapters: Parameter-efficient transfer learning with fast inference. Advances in Neural Information Processing Systems, 36: 0 8152--8172, 2023

    2023

  17. [25]

    Prefix-tuning: Optimizing continuous prompts for generation

    Xiang Lisa Li and Percy Liang. Prefix-tuning: Optimizing continuous prompts for generation. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Pap...

    2021

  18. [26]

    Learning causal representations of single cells via sparse mechanism shift modeling

    Romain Lopez, Natasa Tagasovska, Stephen Ra, Kyunghyun Cho, Jonathan Pritchard, and Aviv Regev. Learning causal representations of single cells via sparse mechanism shift modeling. In Conference on Causal Learning and Reasoning, pp.\ 662--691. PMLR, 2023

    2023

  19. [27]

    Predicting cellular responses to complex perturbations in high-throughput screens

    Mohammad Lotfollahi, Anna Klimovskaia Susmelj, Carlo De Donno, Leon Hetzel, Yuge Ji, Ignacio L Ibarra, Sanjay R Srivatsan, Mohsen Naghipourfar, Riza M Daza, Beth Martin, et al. Predicting cellular responses to complex perturbations in high-throughput screens. Molecular systems...

    2023

  20. [28]

    Harnessing the deep learning power of foundation models in single-cell omics

    Qin Ma, Yi Jiang, Hao Cheng, and Dong Xu. Harnessing the deep learning power of foundation models in single-cell omics. Nature Reviews Molecular Cell Biology, 25 0 (8): 0 593--594, 2024

    2024

  21. [29]

    Multi-seq: sample multiplexing for single-cell rna sequencing using lipid-tagged indices

    Christopher S McGinnis, David M Patterson, Juliane Winkler, Daniel N Conrad, Marco Y Hein, Vasudha Srivastava, Jennifer L Hu, Lyndsay M Murrow, Jonathan S Weissman, Zena Werb, et al. Multi-seq: sample multiplexing for single-cell rna sequencing using lipid-tagged indices. Natu...

    2019

  22. [30]

    M. Moor, O. Banerjee, Z. S. H. Abad, et al. Foundation models for generalist medical artificial intelligence. Nature, 616: 0 259--265, 2023

    2023

  23. [31]

    Disentanglement of single-cell data with biolord

    Zoe Piran, Niv Cohen, Yedid Hoshen, and Mor Nitzan. Disentanglement of single-cell data with biolord. Nature Biotechnology, pp.\ 1--6, 2024

    2024

  24. [32]

    Toward a foundation model of causal cell and tissue biology with a perturbation cell and tissue atlas

    Jennifer E Rood, Anna Hupalowska, and Aviv Regev. Toward a foundation model of causal cell and tissue biology with a perturbation cell and tissue atlas. Cell, 187 0 (17): 0 4520--4545, 2024

    2024

  25. [33]

    Predicting transcriptional outcomes of novel multigene perturbations with gears

    Yashar Roohani, Kai Huang, and Jure Leskovec. Predicting transcriptional outcomes of novel multigene perturbations with gears. Nature Biotechnology, 42 0 (8): 0 927--935, 2024

    2024

  26. [34]

    How much can clip benefit vision-and-language tasks? International Conference on Learning Representations, 2022

    Sheng Shen, Liunian Harold Li, Hao Tan, Mohit Bansal, Anna Rohrbach, Kai-Wei Chang, Zhewei Yao, and Kurt Keutzer. How much can clip benefit vision-and-language tasks? International Conference on Learning Representations, 2022

    2022

  27. [35]

    S. R. Srivatsan, J. L. McFaline-Figueroa, V. Ramani, L. Saunders, J. Cao, J. Packer, H. A. Pliner, D. L. Jackson, R. M. Daza, L. Christiansen, F. Zhang, F. Steemers, J. Shendure, and C. Trapnell. Massively multiplex chemical transcriptomics at single-cell resolution. Science, ...

    2020

  28. [36]

    Scaling single-cell genomics from phenomenology to mechanism

    Amos Tanay and Aviv Regev. Scaling single-cell genomics from phenomenology to mechanism. Nature, 541 0 (7637): 0 331--338, 2017

    2017

  29. [37]

    Transfer learning enables predictions in network biology

    Christina V Theodoris, Ling Xiao, Anant Chopra, Mark D Chaffin, Zeina R Al Sayed, Matthew C Hill, Helene Mantineo, Elizabeth M Brydon, Zexian Zeng, X Shirley Liu, et al. Transfer learning enables predictions in network biology. Nature, 618 0 (7965): 0 616--624, 2023

    2023

  30. [38]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, ukasz Kaiser, and Illia Polosukhin. Attention is all you need. In I. Guyon, U. Von Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett (eds.), Advances in Neural Info...

    2017

  31. [39]

    Evaluating self-supervised learning for molecular graph embeddings

    Hanchen Wang, Jean Kaddour, Shengchao Liu, Jian Tang, Joan Lasenby, and Qi Liu. Evaluating self-supervised learning for molecular graph embeddings. Advances in Neural Information Processing Systems, 36, 2024

    2024

  32. [40]

    Molecular contrastive learning of representations via graph neural networks

    Yuyang Wang, Jianren Wang, Zhonglin Cao, and Amir Barati Farimani. Molecular contrastive learning of representations via graph neural networks. Nature Machine Intelligence, 4 0 (3): 0 279–287, March 2022. ISSN 2522-5839

    2022

  33. [41]

    Dai, and Quoc V Le

    Jason Wei, Maarten Bosma, Vincent Zhao, Kelvin Guu, Adams Wei Yu, Brian Lester, Nan Du, Andrew M. Dai, and Quoc V Le. Finetuned language models are zero-shot learners. In International Conference on Learning Representations, 2022

    2022

  34. [42]

    Parameter-efficient fine-tuning for pre-trained vision models: A survey

    Yi Xin, Siqi Luo, Haodi Zhou, Junlong Du, Xiaohong Liu, Yue Fan, Qing Li, and Yuntao Du. Parameter-efficient fine-tuning for pre-trained vision models: A survey. arXiv preprint arXiv:2402.02242, 2024

    2024

  35. [43]

    F. Yang, W. Wang, F. Wang, et al. scbert as a large-scale pretrained deep language model for cell type annotation of single-cell rna-seq data. Nat Mach Intell, 4: 0 852--866, 2022

    2022

  36. [44]

    Llama-adapter: Efficient fine-tuning of language models with zero-init attention

    Renrui Zhang, Jiaming Han, Chris Liu, Peng Gao, Aojun Zhou, Xiangfei Hu, Shilin Yan, Pan Lu, Hongsheng Li, and Yu Qiao. Llama-adapter: Efficient fine-tuning of language models with zero-init attention. In International Conference on Learning Representations, 2024

    2024

  37. [45]

    write newline

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

  38. [46]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should ...

  39. [47]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@firs...

  40. [48]

    NÆ ݻ <ԩS KF _ =MF ҤIz.rJEFFK.rrru9V?K &|HR ս w5oҲgQ|| /^lR!x_ڽ =Ç+00PժUSppf͚ ժUK< ,X;w΂ I]vѦ ^3^ <X7oNӷk׮Ç-گ_J*_ i ܹ3S5] ] Odӄ OշzKW ?C4ijժaÆ 6loߞL

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibset...

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.