Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
PRESCRIBE: Predicting Single-Cell Responses with Bayesian Estimation
Authors: Jiabei Cheng, Changxi Chi, Jingbo Zhou, Hongyi Xin, Jun Xia
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | This section details PRESCRIBE s empirical evaluation. The experiments were designed to: (i) assess the quality of its uncertainty ( 4.34.6)). (ii) demonstrate its utility in prediction accuracy ( 4.7), and (iii) explore contributions of its core components and initialization strategies ( 4.84.9). |
| Researcher Affiliation | Academia | 1Shanghai Jiao Tong University, 2Westlake University, 3AIMS Lab, The Hong Kong University of Science and Technology (Guangzhou), 4The Hong Kong University of Science and Technology EMAIL EMAIL |
| Pseudocode | Yes | For reference, detailed notation and hyperparameters are provided in Appx. Tabs. 5 and 6, and the core algorithm in Appx. Algo. 1. |
| Open Source Code | Yes | Code is available at https://github.com/Bunnybeibei/PRESCRIBE. |
| Open Datasets | Yes | Datasets. We evaluated PRESCRIBE on three widely recognized benchmark datasets: Norman [4], Replogle2022_Rep1, and Replogle2022_K562 [5]. (Details are provided in Appx. G.) |
| Dataset Splits | Yes | Spilt Dataset. The filtered perturbations were divided into training, testing, and validation sets according to the default operation in GEARS [1]. The training set was used to fit the models, the validation set was used to select the best model, and the testing set was used to evaluate the final model performance. And for more details, see Tab.7. |
| Hardware Specification | Yes | The PRESCRIBE framework was trained on a single NVIDIA RTX 4090 GPU. |
| Software Dependencies | Yes | The implementation leverages PYTORCH LIGHTNING [25] 2.4.0 s modular architecture. |
| Experiment Setup | Yes | B Default Settings table including Learning rate 1e-4, Total epochs 50, Weight decay 1e-5, B 4096, Gradient Accumulation 4, λ1 1e-7, λ2 0.1, λ3 1e-5. |