Interpretable Recommendation via Attraction Modeling: Learning Multilevel Attractiveness over Multimodal Movie Contents

Authors: Liang Hu, Songlei Jian, Longbing Cao, Qingkui Chen

IJCAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The experimental results show the superiority of MLAM over the state-of-the-art methods.
Researcher Affiliation Academia 1Advanced Analytics Institute, University of Technology, Sydney 2Institute of Network Computing & Io T, University of Shanghai for Science and Technology 3College of Computer, National University of Defense Technology, China rainmilk@gmail.com, jiansonglei@163.com, longbing.cao@uts.edu.au, chenqingkui@usst.edu.cn
Pseudocode Yes Algorithm 1 The learning procedure for a mini-batch
Open Source Code Yes The code for more detail is available at: https://github.com/rainmilk/ijcai18-mlma.
Open Datasets Yes The experiments are conducted on the real-world movie watch dataset Movie Lens 1M [Harper and Konstan, 2016].
Dataset Splits No The paper mentions 'we randomly held out 20% user watch records as the testing set, and the remainder were served as the training set.' It does not explicitly define a separate validation set.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory) used for running the experiments.
Software Dependencies No Our model is implemented using Keras [Chollet, 2015] with TensorFlow as the backend. However, specific version numbers for Keras and TensorFlow are not provided.
Experiment Setup Yes where the parameter margin needs to be tuned over data. ... we find that α = 4 performs good through our experiments. ... where α = 2 is set throughh experiments. ... where α = 1 is set throughh experiments.