LLMRG: Improving Recommendations through Large Language Model Reasoning Graphs

Authors: Yan Wang, Zhixuan Chu, Xin Ouyang, Simeng Wang, Hongyan Hao, Yue Shen, Jinjie Gu, Siqiao Xue, James Zhang, Qing Cui, Longfei Li, Jun Zhou, Sheng Li

AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental To evaluate our proposed method, we conduct experiments on three benchmark datasets: the Amazon Beauty, Amazon-Clothing, and Movie Lens-1M (ML-1M) datasets (Mc Auley et al. 2015; Harper and Konstan 2015).
Researcher Affiliation Collaboration 1Ant Group 2University of Virginia
Pseudocode No The paper includes architectural diagrams and descriptions of its modules but does not present any structured pseudocode or algorithm blocks.
Open Source Code No The paper does not provide any explicit statement about releasing source code, nor does it include a link to a code repository.
Open Datasets Yes To evaluate our proposed method, we conduct experiments on three benchmark datasets: the Amazon Beauty, Amazon-Clothing, and Movie Lens-1M (ML-1M) datasets (Mc Auley et al. 2015; Harper and Konstan 2015).
Dataset Splits No To evaluate the performance of our recommendation system, we utilize a leave-one-out strategy where we repeatedly hold out one item from each user s sequence of interactions.
Hardware Specification No The paper discusses the use of LLMs like GPT3.5 and GPT4 but does not specify any hardware (e.g., GPU models, CPU types, memory) used for running the experiments or training the LLMRG model.
Software Dependencies No The paper mentions using GPT3.5 or GPT4 and base models like SR-GNN, but it does not provide specific version numbers for any software dependencies, such as programming languages, libraries, or frameworks.
Experiment Setup No The 'Settings' section details the datasets and evaluation metrics (HR@n, NDCG@n) but does not provide specific experimental setup details such as hyperparameter values (learning rate, batch size, epochs), optimizer settings, or other system-level training configurations.