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..
Graph-Theoretic Insights into Bayesian Personalized Ranking for Recommendation
Authors: Kai Zheng, Jianxin Wang, Jinhui Xu
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We validate BPR+ through extensive empirical testing across five real-world datasets and demonstrate its efficacy in refining graph self-supervised learning frameworks. Additionally, we explore the application of BPR+ in drug repositioning, highlighting its potential to support pharmaceutical research and development. Our findings not only illuminate the success factors of previous methodologies but also offer new theoretical insights into this learning paradigm. |
| Researcher Affiliation | Academia | 1School of Computer Science and Engineering, Central South University Changsha 410083, China 2The Hunan Provincial Key Lab of Bioinformatics, Central South University Changsha 410083, China 3The College of Computer Science and Technology,China University of Petroleum Qingdao 266580, China 4School of Information Science and Technology, University of Science and Technology of China He Fei, 230026, China 5Institute of Artificial Intelligence, Hefei Comprehensive National Science Center He Fei, 230026, China EMAIL, EMAIL |
| Pseudocode | No | The paper describes methods and formulations in regular text and mathematical equations, but does not include any clearly labeled pseudocode or algorithm blocks with structured steps. |
| Open Source Code | Yes | We have uploaded the code in the additional supplementary material. |
| Open Datasets | Yes | In this section, we evaluate our model and several baselines on five real-world datasets: Amazon, Gowalla, Yelp, Last FM, and Beer. The Amazon dataset includes implicit feedback from users on books from the Amazon platform... We downloaded the datasets from prior work, including its predefined training, validation, and test sets. The Amazon, Gowalla, and Yelp datasets were downloaded from SSLRec [18] and have been predivided into training, validation, and test sets . The Last FM and Beer datasets were obtained from the corresponding code and data provided by ada GCL [5], and have also been pre-partitioned. |
| Dataset Splits | Yes | We downloaded the datasets from prior work, including its predefined training, validation, and test sets. The Amazon, Gowalla, and Yelp datasets were downloaded from SSLRec [18] and have been predivided into training, validation, and test sets . The Last FM and Beer datasets were obtained from the corresponding code and data provided by ada GCL [5], and have also been pre-partitioned. |
| Hardware Specification | Yes | Experiments were conducted on a high-performance hardware platform comprising an Intel Xeon Platinum 8352V processor, an NVIDIA RTX 4090 with 24 GB of memory, and a system running Ubuntu 20.04. |
| Software Dependencies | Yes | The software environment included Py Torch version 1.11.0, Python 3.8, and CUDA 11.3. |
| Experiment Setup | Yes | The code for all models originates from SSLRec, with each model s embedding size fixed at 32 and the batch size set to 4096. To ensure a fair comparison, we employed a grid search method to determine the optimal parameter combination for each model, with λ search range set to {1e-3, 1e-4, 1e-5, 1e-6, 1e-7}. |