Uncertainty-Aware Adaptation of Large Language Models for Protein-Protein Interaction Analysis

Byung-Jun Yoon; Gilchan Park; Kriti Chopra; Nathan M. Urban; Nicholas Jeon; Sanket Jantre; Tianle Wang; Xiaoning Qian

REVIEW 4 major objections 5 minor 56 references

Uncertainty-Aware Adaptation of Large Language Models for Protein-Protein Interaction Analysis

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

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

Pith's one-line read Uncertainty-aware fine-tuning of large language models is shown to improve protein-protein interaction prediction while producing better-calibrated confidence estimates.

desk verdict A useful empirical comparison of two existing UQ methods on disease-specific PPI prediction, but the central accuracy and calibration claims are undermined by test-set-based rank selection and an unspecified train/test split; worth a major revision, not a desk reject. read the letter →

arxiv 2502.06173 v2 pith:WAASOCDY submitted 2025-02-10 cs.LG cs.AIcs.CLstat.APstat.ML

Sanket Jantre , Tianle Wang , Gilchan Park , Kriti Chopra , Nicholas Jeon , Xiaoning Qian , Nathan M. Urban , Byung-Jun Yoon This is my paper · ORCID

classification cs.LGcs.AIcs.CLstat.APstat.ML

keywords LargeLanguageModelsLow-RankAdaptationUncertaintyQuantificationBayesianInferenceDeepEnsembleProtein-ProteinInteractionConfidenceCalibrationBioMedGPT

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 sets out to show that the reliability of large language models in predicting protein-protein interactions can be improved by making the fine-tuning procedure uncertainty-aware. It compares a standard low-rank adapter (LoRA) against two uncertainty-aware variants—an ensemble of LoRA adapters and a Bayesian LoRA that approximates a posterior over the adapter weights—on balanced task-specific datasets for neurodegenerative, metabolic, and cancer-related interactions. Across most settings, the uncertainty-aware variants achieve higher accuracy and lower negative log-likelihood than the single LoRA baseline, and they produce confidence scores that are better calibrated to actual correctness. The reason this matters is that biomedical users need to know when a prediction is trustworthy, not just whether the model is often right.

What carries the argument

The central object is the low-rank adapter: with the backbone frozen, the forward pass becomes $h = (W_0 + BA)a$, so only the small matrices $B$ and $A$ are trained. Uncertainty is introduced in two ways. First, a LoRA ensemble trains $M$ such adapters independently and averages their output probabilities, capturing epistemic diversity at a fraction of the cost of full-model ensembles. Second, Bayesian LoRA places a zero-mean Gaussian prior on the adapter parameters, finds the MAP estimate, approximates the posterior as $N(\theta_{\mathrm{MAP}}, H^{-1})$ using a Laplace expansion with a Kronecker-factored (K-FAC) Fisher information matrix, and then linearizes the model at the MAP so that the predictive distribution over logits is Gaussian and can be sampled by Cholesky decomposition. This machinery is what lets the paper attach calibrated confidence to each yes/no prediction.

What would settle it

Re-run the three PPI experiments with an 80/20 split that holds out entire proteins, so that no protein appearing in the training set appears in the test set; if the accuracy and calibration advantages of the uncertainty-aware variants shrink or vanish, the original results partly reflect protein-level memorization rather than robust interaction prediction.

Watch

Extended reading notes

Core claim

The central claim, stated in the paper's own terms, is that 'incorporating UQ strategies not only enhances PPI prediction accuracy but also yields better-calibrated confidence measures.' Concretely, on the ND-PPI, M-PPI, and C-PPI datasets with LLaMA-3-8B and BioMedGPT-LM-7B backbones, the LoRA ensemble attains the highest accuracy, F1, MCC, and AUROC in most configurations, and Bayesian LoRA consistently delivers the lowest expected calibration error in the majority of settings (e.g., ECE as low as 0.027 on C-PPI with LLaMA-3). The paper presents this as evidence that principled uncertainty quantification can be integrated into parameter-efficient fine-tuning without sacrificing predictive performance.

Load-bearing premise

The load-bearing assumption is that the 80/20 train/test split separates interactions without leaking the same proteins into both sets; the paper never states whether the split is by protein or by interaction, and if proteins overlap, accuracy gains could come from memorizing protein identities rather than from learning to generalize.

Editorial extensions

If this is right

In most experimental configurations, replacing a single LoRA with either a LoRA ensemble or Bayesian LoRA improves or matches accuracy while reducing negative log-likelihood and calibration error, so the uncertainty-aware variants can serve as drop-in replacements in the same fine-tuning pipeline.
The calibration gains are consistent across two different LLM backbones (LLaMA-3-8B and BioMedGPT-LM-7B), indicating the benefit is tied to the uncertainty mechanism rather than to a specific pretrained model.
Reliability diagrams show a visibly reduced gap between confidence and observed accuracy for both uncertainty-aware methods relative to the single LoRA, supporting the claim that these methods mitigate the overconfidence problem.
Because the adapters are lightweight and the backbone is frozen, the added uncertainty estimation keeps computational overhead close to standard LoRA training, which matters for scaling to larger models.
The approach can be extended to other biomedical tasks that use language-like protein representations, including agentic AI workflows and prompt optimization, as the authors note.

Reading between the lines

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

The paper's train/test partition is described only as an 80/20 split with no statement of whether proteins are shared across the split; my inference is that a held-out-protein evaluation would be the decisive test of whether the accuracy gains reflect generalization to unseen proteins rather than memorization of protein identifiers.
The comparison baseline is a single deterministic LoRA; the reported advantage over that baseline does not establish superiority over other uncertainty-estimation strategies, so a broader benchmark would be needed before adopting this as the default.
The negative interaction pairs are pairs with no documented evidence of interaction, which makes the classification task easier than real discovery settings where absence of evidence is noisy; under noisier labels, the calibration gains may be smaller, which is a testable prediction.
If the calibrated confidence truly tracks correctness, then a practical use is to rank protein pairs by uncertainty and send only the low-confidence tail to experimental validation, which the paper suggests but does not demonstrate.

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.

Desk Editor's Note

read the letter

Colleague, here is my read on arXiv:2502.06173. The paper does something straightforward but useful: it applies two established uncertainty-quantification recipes (LoRA ensembles and Bayesian LoRA) to LLM-based protein–protein interaction prediction on three disease-focused datasets, using LLaMA-3 and BioMedGPT. The new content is the empirical comparison itself, not the methods. Credit where due: the experiments are reported fairly, with means and standard deviations over three seeds, and the authors explicitly flag which comparisons are not statistically significant (e.g., all BioMedGPT C-PPI metrics). The reliability diagrams are a nice addition.

The soft spot is real and load-bearing. In Appendix B the LoRA rank r is selected on the test set. Table I then reports the best performance among the three r values. That is test-set leakage: the headline numbers are hyperparameter-tuned on the evaluation set, not unbiased estimates of generalization. The stress-test example is on point: on M-PPI with LLaMA-3, r=16 is chosen, and Bayesian LoRA's accuracy is best at that rank and much worse at r=8, so the comparison cherry-picks favorable settings. The relative ranking across methods is probably not dramatically unfair since the same r is used for all three, but the absolute gains over single LoRA are inflated and the central claim that UQ "enhances accuracy and calibration" is not yet supported.

The second issue the reader flagged is also worth stating plainly: the paper never says whether the 80/20 split is by protein or by interaction. If proteins overlap between train and test, the model may memorize protein-specific patterns, and the reported accuracy does not measure generalization to unseen proteins. This is a missing experimental detail that needs fixing before the results can be trusted.

These are not fatal flaws in the sense that the methods are broken. They are fixable with a validation split and a clarified protocol. The paper is a reasonable recipe for practitioners who want to add calibrated confidence to LLM-based PPI screens. But as written, the evidence for the strongest claims is not trustworthy.

Who is this for? People in biomedical ML who want a worked example of LoRA ensembles and Bayesian LoRA on protein interaction tasks. It does not open new science, but it is a competent application. A serious referee should see it, but the authors need to re-run rank selection on a validation split and state the split granularity before the results can be accepted. That is a major-revision request, not a reject.

In short: engage with it, but require the fix.

Referee Report

4 major / 5 minor

Summary. The paper proposes uncertainty-aware parameter-efficient fine-tuning of large language models for protein-protein interaction (PPI) prediction. It applies LoRA ensembles and Bayesian LoRA (Laplace approximation with K-FAC) to LLaMA-3-8B and BioMedGPT-LM-7B on three disease-focused balanced PPI datasets (neurodegenerative, metabolic, cancer), comparing accuracy, negative log-likelihood, expected calibration error, and several secondary metrics against a single-LoRA baseline. The central claim is that uncertainty-aware LoRA adaptation improves both PPI prediction accuracy and confidence calibration.

Significance. If the empirical claim is supported, the paper would provide a practical and computationally efficient recipe for uncertainty quantification in biomedical LLM applications, with potential value for trustworthy PPI screening and downstream disease-mechanism studies. The authors report three independent runs, multiple complementary metrics, reliability diagrams, and use publicly available datasets and established UQ methods. However, the central empirical claim is currently undermined by a test-set-based hyperparameter selection protocol and by an unspecified train/test split granularity; these issues are load-bearing for the conclusion and must be fixed before the results can be interpreted as evidence of improved generalization.

major comments (4)

[§IV 'Implementation Details' and Appendix B] The LoRA rank r is selected by comparing test-set performance in the hyperparameter search shown in Table II, and the chosen r is then fixed across methods. This is test-set leakage: the Acc/NLL/ECE values in Table I are the best over three hyperparameter choices rather than unbiased estimates of generalization. The problem is visible in M-PPI with LLaMA-3, where Bayesian LoRA accuracy rises from 82.37 at r=8 to 83.41 at r=16 and the selected r=16 favors that method. The authors should either select r on a held-out validation split or report results averaged over all r values; without this, the reported accuracy and calibration gains are not trustworthy.
[§IV 'PPI Datasets'] The text states that each dataset is partitioned into an 80% training set and a 20% testing set, but it never specifies whether the split is at the protein level or at the interaction level. If the same proteins appear in both training and test pairs, the model can memorize protein-specific features and the reported accuracy does not measure generalization to unseen proteins. This distinction is essential for the reliability claim and should be stated explicitly; if the current split is interaction-level, a protein-level split should be added to validate the generalization claim.
[§IV 'Results'] The statistical evidence is weaker than the conclusion claims. The text acknowledges that LLaMA-3 ND-PPI accuracy, BioMedGPT M-PPI accuracy, and all BioMedGPT C-PPI metrics are not statistically significant (p>0.05), yet the abstract and conclusion state that uncertainty-aware adaptation 'enhances PPI prediction accuracy' and 'consistently improves' performance. The conclusion should be qualified to the settings where the gains are significant, and the actual p-values, together with a multiple-testing correction for the large number of metrics, should be reported.
[Appendix B] The selection rule is described as choosing the rank with 'overall best performance across LoRA-based models', but this criterion is not precisely defined: it is unclear which metric is optimized, whether accuracy or calibration is prioritized, and how ties are broken. Because the selected rank can systematically favor one method (as the M-PPI LLaMA-3 example shows), the comparison among Single LoRA, LoRA Ensemble, and Bayesian LoRA is not fully controlled. The selection rule should be specified precisely, ideally using a validation metric.

minor comments (5)

[§III-B] In the Bayesian LoRA derivation, 'AdamW with its weight deca' appears to be a typo for 'weight decay'.
[Appendix A] The text says 'we set |B_m| = 15', but the ECE formula uses |B_m| to denote bin size; if 15 is the number of bins, the notation should be M=15 or should be clarified.
[Table I] The standard deviations are printed as concatenated subscripts (e.g., 87.421.64), which is difficult to read; please use proper subscripts or separate columns for means and standard deviations.
[Fig. 2] Reliability diagrams are shown only for LLaMA-3, with BioMedGPT omitted for brevity; given that the conclusions cover both backbones, the BioMedGPT diagrams should be included in a supplement or their ECE values should be discussed in the text.
[General] No data or code availability statement is provided, which makes the exact preprocessing and split procedure impossible to reproduce from the paper alone.

Circularity Check

1 steps flagged · score 5.0 of 10

Test-set-selected LoRA rank makes reported UQ gains partly self-confirming, though the core UQ machinery is externally sourced and comparisons appear robust across ranks.

fitted input called prediction [Section IV, 'Implementation Details' and Appendix B, 'Effect of LoRA Rank' (Table II)]
"Implementation Details: 'We select the LoRA rank (r) via hyperparameter search (Appendix-B) and report results for each LLM backbone and PPI dataset combination in Table I using the finalized r.' Appendix B: 'In Table II, we present results of the hyperparameter search over the LoRA rank r ∈ {8, 16, 32} ... LoRA-based models achieve overall best performance on the ND-PPI dataset at r = 32 and on the C-PPI dataset at r = 16.'"

The test-set Acc/NLL/ECE that Table I reports as evidence are the same quantities used in Table II to choose the finalized LoRA rank r; no validation split is described. Selecting r by 'overall best performance' on the test set and then quoting the test metrics at that r makes the reported numbers the selected optima of the search, not unbiased predictions. The headline claim that UQ 'enhances PPI prediction accuracy' and yields 'better-calibrated confidence measures' is thus partly an artifact of fitting r to the test response: the metric used for selection is the metric presented as the result.

full rationale

The paper contains no self-referential equations: the LoRA ensemble average and Bayesian LoRA Laplace/K-FAC posterior are imported from external prior work ([44], [48], [49]) and applied as stated, so there is no derivation chain in which a result equals its input by construction. No load-bearing self-citation was found: citations to the authors' prior work ([37]-[39], [42], [56]) appear in related-work or future-work contexts and do not justify the central UQ claims. The one circular-adjacent step is the test-set hyperparameter selection for the LoRA rank r: Appendix B chooses r by the same test-set Acc/NLL/ECE that Table I later reports as evidence of UQ gains, with no validation split mentioned. This is a genuine fitted-input-called-prediction issue, but it is only partial: the same r is applied to all three methods, and in many cases the qualitative ordering (e.g., LoRA ensemble beating single LoRA in accuracy) holds across the tested r values, so the central claim retains independent content. The missing specification of whether the 80/20 split is by protein or by interaction is a related soundness concern about generalization, not circularity. Overall, one moderate fitted-input step, no definitional circularity: score 5.

Assumptions & free parameters 5 free parameters · 3 assumptions · 0 invented entities

The paper introduces no new entities, but it relies on fitted hyperparameters (rank, prior precision, ensemble size) and domain assumptions about dataset construction and splitting. The most consequential choice, rank selection on the test set, is a free parameter that affects all reported metrics.

free parameters (5)

LoRA rank r = 8, 16, or 32 (per dataset/backbone, selected from test results)
Chosen by hyperparameter search in Appendix B; reported metrics use the rank with best test performance, which risks overfitting.
Prior precision lambda = 0.1
Fixed by hand for the Bayesian LoRA prior.
Ensemble size M = 3
Number of LoRA ensemble members, chosen without sensitivity analysis.
LoRA alpha = 32
Scaling factor for LoRA updates.
Batch size = 4 (ND-PPI, M-PPI), 16 (C-PPI)
Inherited from ref 29.

assumptions (3)

domain assumption Balanced PPI datasets with negatives defined by absence of documented evidence are valid for training and evaluation
Section IV 'PPI Datasets'; negative pairs may include undiscovered true interactions, biasing the task.
domain assumption Random 80/20 split of interactions preserves i.i.d. test distribution
Section IV 'PPI Datasets'; no protein-disjoint split is described, so test results may not reflect generalization.
standard math K-FAC approximation of the Fisher information is accurate for the LoRA posterior
Section III-B, from ref 44; paper relies on this for Bayesian LoRA calibration.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Uncertainty-Aware Adaptation of Large Language Models for Protein-Protein Interaction Analysis." pith.science (2026). https://pith.science/paper/WAASOCDY

@misc{pith2026250206173,
  author       = {Pith},
  title        = {Pith review of: Uncertainty-Aware Adaptation of Large Language Models for Protein-Protein Interaction Analysis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WAASOCDY}},
  note         = {Machine review of arXiv:2502.06173}
}

read the original abstract

Identification of protein-protein interactions (PPIs) helps derive cellular mechanistic understanding, particularly in the context of complex conditions such as neurodegenerative disorders, metabolic syndromes, and cancer. Large Language Models (LLMs) have demonstrated remarkable potential in predicting protein structures and interactions via automated mining of vast biomedical literature; yet their inherent uncertainty remains a key challenge for deriving reproducible findings, critical for biomedical applications. In this study, we present an uncertainty-aware adaptation of LLMs for PPI analysis, leveraging fine-tuned LLaMA-3 and BioMedGPT models. To enhance prediction reliability, we integrate LoRA ensembles and Bayesian LoRA models for uncertainty quantification (UQ), ensuring confidence-calibrated insights into protein behavior. Our approach achieves competitive performance in PPI identification across diverse disease contexts while addressing model uncertainty, thereby enhancing trustworthiness and reproducibility in computational biology. These findings underscore the potential of uncertainty-aware LLM adaptation for advancing precision medicine and biomedical research.

Figures

Figures reproduced from arXiv: 2502.06173 by the authors.

**Figure 2.** Reliability Diagrams: A visual representation of model calibration [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

56 extracted references · 41 canonical work pages

[1]

Interactome: gateway into systems biology,

M. Cusick, N. Klitgord, M. Vidal, and D. Hill, “Interactome: gateway into systems biology,” Hum. Mol. Genet. , vol. 14, 2005

work page 2005
[2]

Chapter 4: Protein interactions and disease,

M. Gonzalez and M. Kann, “Chapter 4: Protein interactions and disease,” PLoS Comput. Biol. , vol. 8, no. 12, 2012

work page 2012
[3]

The STRING database in 2023: protein-protein association networks and functional enrichment analyses for any sequenced genome of interest,

D. Szklarczyk, R. Kirsch, M. Koutrouli, K. Nastou, et al., “The STRING database in 2023: protein-protein association networks and functional enrichment analyses for any sequenced genome of interest,” Nucleic Acids Res. , vol. 51, 2023

work page 2023
[4]

The BioGRID interaction database: 2019 update,

R. Oughtred, C. Stark, B.-J. Breitkreutz, J. Rust, et al., “The BioGRID interaction database: 2019 update,” Nucleic Acids Res. , vol. 47, 2019

work page 2019
[5]

The MIntAct project–intact as a common curation platform for 11 molecular interaction databases,

S. Orchard, M. Ammari, B. Aranda, L. Breuza, et al., “The MIntAct project–intact as a common curation platform for 11 molecular interaction databases,” Nucleic Acids Res. , vol. 42, 2014

work page 2014
[6]

Identification of direct residue contacts in protein–protein interaction by message passing,

M. Weigt, R. White, H. Szurmant, J. Hoch, et al., “Identification of direct residue contacts in protein–protein interaction by message passing,” PNAS, vol. 106, no. 1, 2009

work page 2009
[7]

BIPSPI: A method for the prediction of partner-specific protein-protein interfaces,

R. Sanchez-Garcia, C. Sorzano, J. Carazo, and J. Segura, “BIPSPI: A method for the prediction of partner-specific protein-protein interfaces,” Bioinform., vol. 35, no. 3, 2019

work page 2019
[8]

Cornea: A pipeline to decrypt the inter-protein interfaces from amino acid sequence information,

K. Chopra, B. Burdak, K. Sharma, A. Kembhavi, et al., “Cornea: A pipeline to decrypt the inter-protein interfaces from amino acid sequence information,” Biomol., vol. 10, no. 6, 2020

work page 2020

Show all 56 references

[9]

State-of-the-art computational methods to predict protein–protein interactions with high accuracy and coverage,

N. Kewalramani, A. Emili, and M. Crovella, “State-of-the-art computational methods to predict protein–protein interactions with high accuracy and coverage,” Proteomics, vol. 23, no. 1-2, 2023

2023
[10]

Revolutionizing protein–protein interaction prediction with deep learning,

J. Zhang, J. Durham, and Q. Cong, “Revolutionizing protein–protein interaction prediction with deep learning,” Curr. Opin. Struct. Biol. , vol. 85, 2024

2024
[11]

Protein- protein interaction prediction with deep learning: A comprehensive review,

F. Soleymani, E. Paquet, H. Viktor, W. Michalowski, et al., “Protein- protein interaction prediction with deep learning: A comprehensive review,” Comput. Struct. Biotechnol. J. , vol. 20, 2022

2022
[12]

A survey on uncertainty quantification of large language models: Taxonomy, open research challenges, and future directions,

O. Shorinwa, Z. Mei, J. Lidard, A. Z. Ren, et al., “A survey on uncertainty quantification of large language models: Taxonomy, open research challenges, and future directions,” arXiv:2412.05563, 2024

2024 arXiv
[13]

Large language models: A survey,

S. Minaee, T. Mikolov, N. Nikzad, M. Chenaghlu, et al., “Large language models: A survey,” arXiv:2402.06196, 2024

2024 arXiv
[14]

Language models are unsupervised multitask learners,

A. Radford, J. Wu, R. Child, D. Luan, et al., “Language models are unsupervised multitask learners,” OpenAI, 2019

2019
[15]

Language models are few-shot learners,

T. Brown, B. Mann, N. Ryder, M. Subbiah, et al., “Language models are few-shot learners,” in NeurIPS, 2020

2020
[16]

GPT-4 technical report,

OpenAI et al., “GPT-4 technical report,” arXiv:2303.08774, 2023

2023 arXiv
[17]

The llama 3 herd of models,

A. Dubey, A. Jauhri, A. Pandey, A. Kadian, et al., “The llama 3 herd of models,” arXiv:2407.21783, 2024

2024 arXiv
[18]

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

H. Cui, C. Wang, H. Maan, K. Pang, et al., “scgpt: toward building a foundation model for single-cell multi-omics using generative ai,” Nat. Methods, vol. 21, 2024

2024
[19]

BioGPT: Generative pre- trained transformer for biomedical text generation and mining,

R. Luo, L. Sun, Y . Xia, T. Qin, et al., “BioGPT: Generative pre- trained transformer for biomedical text generation and mining,” Brief. Bioinform., vol. 23, no. 6, 2022

2022
[20]

Large language models encode clinical knowledge,

K. Singhal, S. Azizi, T. Tu, S. Mahdavi, et al., “Large language models encode clinical knowledge,” Nature, vol. 620, no. 7972, 2023

2023
[21]

Gatortron: A large clinical language model to unlock patient information from unstructured electronic health records,

X. Yang, A. Chen, N. PourNejatian, H. Shin, et al., “Gatortron: A large clinical language model to unlock patient information from unstructured electronic health records,” arXiv:2203.03540, 2022

2022 arXiv
[22]

Chem- Crow: Augmenting large-language models with chemistry tools,

A. Bran, S. Cox, O. Schilter, C. Baldassari, et al., “Chem- Crow: Augmenting large-language models with chemistry tools,” arXiv:2304.05376, 2023

2023 arXiv
[23]

Galactica: A large language model for science,

R. Taylor, M. Kardas, G. Cucurull, T. Scialom, et al., “Galactica: A large language model for science,” arXiv:2211.09085, 2022

2022 arXiv
[24]

BloombergGPT: A large language model for finance,

S. Wu, O. Irsoy, S. Lu, V . Dabravolski, et al., “BloombergGPT: A large language model for finance,” arXiv:2303.17564, 2023

2023 arXiv
[25]

ClimateGPT: Towards ai synthesizing interdisciplinary research on climate change,

D. Thulke, Y . Gao, P. Pelser, R. Brune, et al., “ClimateGPT: Towards ai synthesizing interdisciplinary research on climate change,” arXiv:2401.09646, 2024

2024 arXiv
[26]

ProteinBERT: a universal deep-learning model of protein sequence and function,

N. Brandes, D. Ofer, Y . Peleg, N. Rappoport, et al., “ProteinBERT: a universal deep-learning model of protein sequence and function,” Bioinform., vol. 38, no. 8, 2022

2022
[27]

Simulating 500 million years of evolution with a language model,

T. Hayes, R. Rao, H. Akin, et al., “Simulating 500 million years of evolution with a language model,” Science, vol. 387, 2025

2025
[28]

BioMedGPT: Open multimodal generative pre-trained transformer for biomedicine,

Y . Luo, J. Zhang, S. Fan, K. Yang, et al., “BioMedGPT: Open multimodal generative pre-trained transformer for biomedicine,” arXiv:2308.09442, 2023

2023 arXiv
[29]

Evaluating large language models for predicting protein behavior under radiation exposure and disease conditions,

R. Engel and G. Park, “Evaluating large language models for predicting protein behavior under radiation exposure and disease conditions,” in BioNLP, 2024

2024
[30]

Think twice before trusting: Self-detection for large language models through comprehensive answer reflection,

M. Li, W. Wang, F. Feng, F. Zhu, et al., “Think twice before trusting: Self-detection for large language models through comprehensive answer reflection,” in EMNLP, 2024

2024
[31]

Can LLMs express their uncertainty? an empirical evaluation of confidence elicitation in LLMs,

M. Xiong, Z. Hu, X. Lu, Y . Li, et al., “Can LLMs express their uncertainty? an empirical evaluation of confidence elicitation in LLMs,” arXiv:2306.13063, 2023

2023 arXiv
[32]

Taming overconfidence in LLMs: Reward calibration in RLHF,

J. Leng, C. Huang, B. Zhu, and J. Huang, “Taming overconfidence in LLMs: Reward calibration in RLHF,” arXiv:2410.09724, 2024

2024 arXiv
[33]

Preserving pre-trained features helps calibrate fine-tuned language models,

G. He, J. Chen, and J. Zhu, “Preserving pre-trained features helps calibrate fine-tuned language models,” arXiv:2305.19249, 2023

2023 arXiv
[34]

R. M. Neal, Bayesian Learning for Neural Networks , New York: Springer Verlag, 1996

1996
[35]

What are Bayesian neural network posteriors really like?,

P. Izmailov, S. Vikram, M. Hoffman, and A. Wilson, “What are Bayesian neural network posteriors really like?,” in ICML, 2021

2021
[36]

Weight uncertainty in neural network,

C. Blundell, J. Cornebise, K. Kavukcuoglu, and D. Wierstra, “Weight uncertainty in neural network,” in ICML, 2015

2015
[37]

Layer adaptive node selection in Bayesian neural networks: Statistical guarantees and implementation details,

S. Jantre, S. Bhattacharya, and T. Maiti, “Layer adaptive node selection in Bayesian neural networks: Statistical guarantees and implementation details,” Neural Networks, vol. 167, pp. 309–330, 2023

2023
[38]

Spike-and-slab shrinkage priors for structurally sparse Bayesian neural networks,

S. Jantre, S. Bhattacharya, and T. Maiti, “Spike-and-slab shrinkage priors for structurally sparse Bayesian neural networks,” IEEE Trans. Neural Netw. Learn. Syst. , 2024

2024
[39]

Learning active subspaces for effective and scalable uncertainty quantification in deep neural networks,

S. Jantre, N. M. Urban, X. Qian, and B.-J. Yoon, “Learning active subspaces for effective and scalable uncertainty quantification in deep neural networks,” in IEEE ICASSP, 2024

2024
[40]

Simple and scalable predictive uncertainty estimation using deep ensembles,

B. Lakshminarayanan, A. Pritzel, and C. Blundell, “Simple and scalable predictive uncertainty estimation using deep ensembles,” in NIPS, 2017

2017
[41]

A simple baseline for Bayesian uncertainty in deep learning,

W. Maddox, P. Izmailov, T. Garipov, D. Vetrov, et al., “A simple baseline for Bayesian uncertainty in deep learning,” in NeurIPS, 2019

2019
[42]

Sequential Bayesian neural subnetwork ensembles,

S. Jantre, S. Bhattacharya, N. M. Urban, B.-J. Yoon, et al., “Sequential Bayesian neural subnetwork ensembles,” arXiv:2206.00794, 2022

2022 arXiv
[43]

LoRA: Low-rank adaptation of large language models,

E. Hu, yelong shen, P. Wallis, Z. Allen-Zhu, et al., “LoRA: Low-rank adaptation of large language models,” in ICLR, 2022

2022
[44]

Bayesian low-rank adaptation for large language models,

A. Yang, M. Robeyns, X. Wang, and L. Aitchison, “Bayesian low-rank adaptation for large language models,” in ICLR, 2024

2024
[45]

BLoB: Bayesian low-rank adaptation by backpropagation for large language models,

Y . Wang, H. Shi, L. Han, et al., “BLoB: Bayesian low-rank adaptation by backpropagation for large language models,” in NeurIPS, 2024

2024
[46]

Bayesian-LoRA: LoRA based parameter efficient fine-tuning using optimal quantization levels and rank values trough differentiable Bayesian gates,

C. Meo, K. Sycheva, A. Goyal, and J. Dauwels, “Bayesian-LoRA: LoRA based parameter efficient fine-tuning using optimal quantization levels and rank values trough differentiable Bayesian gates,” in 2nd Workshop on Advancing Neural Network Training: Computational Efficiency, Sca...

2024
[47]

Gaussian stochastic weight averaging for Bayesian low-rank adaptation of large language models,

E. Onal, K. Fl ¨oge, E. Caldwell, A. Sheverdin, et al., “Gaussian stochastic weight averaging for Bayesian low-rank adaptation of large language models,” in 6th Symposium on Advances in Approximate Bayesian Inference - Non Archival Track , 2024

2024
[48]

LoRA ensembles for large language model fine-tuning,

X. Wang, L. Aitchison, and M. Rudolph, “LoRA ensembles for large language model fine-tuning,” arXiv:2310.00035, 2023

2023 arXiv
[49]

Uncertainty quantification in fine-tuned LLMs using LoRA ensembles,

O. Balabanov and H. Linander, “Uncertainty quantification in fine-tuned LLMs using LoRA ensembles,” arXiv:2402.12264, 2024

2024 arXiv
[50]

Laplace redux-effortless Bayesian deep learning,

E. Daxberger, A. Kristiadi, A. Immer, R. Eschenhagen, et al., “Laplace redux-effortless Bayesian deep learning,” in NeurIPS, 2021

2021
[51]

Adapting the linearised laplace model evidence for modern deep learning,

J. Antor ´an, D. Janz, J. Allingham, et al., “Adapting the linearised laplace model evidence for modern deep learning,” in ICML, 2022

2022
[52]

Peft: State-of- the-art parameter-efficient fine-tuning methods,

S. Mangrulkar, S. Gugger, L. Debut, Y . Belkada, et al., “Peft: State-of- the-art parameter-efficient fine-tuning methods,” URL: https://github. com/huggingface/peft, 2022

2022
[53]

Predicting protein–protein interactions using symmetric logistic matrix factorization,

F. Pei, Q. Shi, H. Zhang, and I. Bahar, “Predicting protein–protein interactions using symmetric logistic matrix factorization,” J. Chem. Inf. Model., vol. 61, no. 4, 2021

2021
[54]

Network-based protein- protein interaction prediction method maps perturbations of cancer interactome,

J. Qiu, K. Chen, C. Zhong, S. Zhu, et al., “Network-based protein- protein interaction prediction method maps perturbations of cancer interactome,” PLoS Genet., vol. 17, no. 11, 2021

2021
[55]

Delving deep into rectifiers: Surpassing human-level performance on imagenet classification,

K. He, X. Zhang, S. Ren, et al., “Delving deep into rectifiers: Surpassing human-level performance on imagenet classification,” in ICCV, 2015

2015
[56]

Pareto prompt optimization,

G. Zhao, B.-J. Yoon, G. Park, S. Jha, et al., “Pareto prompt optimization,” in ICLR, 2025

2025

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.

[1] [1]

Interactome: gateway into systems biology,

M. Cusick, N. Klitgord, M. Vidal, and D. Hill, “Interactome: gateway into systems biology,” Hum. Mol. Genet. , vol. 14, 2005

work page 2005

[2] [2]

Chapter 4: Protein interactions and disease,

M. Gonzalez and M. Kann, “Chapter 4: Protein interactions and disease,” PLoS Comput. Biol. , vol. 8, no. 12, 2012

work page 2012

[3] [3]

The STRING database in 2023: protein-protein association networks and functional enrichment analyses for any sequenced genome of interest,

D. Szklarczyk, R. Kirsch, M. Koutrouli, K. Nastou, et al., “The STRING database in 2023: protein-protein association networks and functional enrichment analyses for any sequenced genome of interest,” Nucleic Acids Res. , vol. 51, 2023

work page 2023

[4] [4]

The BioGRID interaction database: 2019 update,

R. Oughtred, C. Stark, B.-J. Breitkreutz, J. Rust, et al., “The BioGRID interaction database: 2019 update,” Nucleic Acids Res. , vol. 47, 2019

work page 2019

[5] [5]

The MIntAct project–intact as a common curation platform for 11 molecular interaction databases,

S. Orchard, M. Ammari, B. Aranda, L. Breuza, et al., “The MIntAct project–intact as a common curation platform for 11 molecular interaction databases,” Nucleic Acids Res. , vol. 42, 2014

work page 2014

[6] [6]

Identification of direct residue contacts in protein–protein interaction by message passing,

M. Weigt, R. White, H. Szurmant, J. Hoch, et al., “Identification of direct residue contacts in protein–protein interaction by message passing,” PNAS, vol. 106, no. 1, 2009

work page 2009

[7] [7]

BIPSPI: A method for the prediction of partner-specific protein-protein interfaces,

R. Sanchez-Garcia, C. Sorzano, J. Carazo, and J. Segura, “BIPSPI: A method for the prediction of partner-specific protein-protein interfaces,” Bioinform., vol. 35, no. 3, 2019

work page 2019

[8] [8]

Cornea: A pipeline to decrypt the inter-protein interfaces from amino acid sequence information,

K. Chopra, B. Burdak, K. Sharma, A. Kembhavi, et al., “Cornea: A pipeline to decrypt the inter-protein interfaces from amino acid sequence information,” Biomol., vol. 10, no. 6, 2020

work page 2020