Robust Asymmetric Bayesian Adaptive Matrix Factorization
Authors: Xin Guo, Boyuan Pan, Deng Cai, Xiaofei He
IJCAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We compare ALAMF with other state-of-the-art matrix factorization methods both on data sets ranging from synthetic and real-world application. The experimental results demonstrate the effectiveness of our proposed approach. and 5 Experiments In this section, we empirically compare the proposed ALAMF model with seven state-of-the-art methods. |
| Researcher Affiliation | Academia | Xin Guo Boyuan Pan Deng Cai Xiaofei He State Key Lab of CAD&CG, College of Computer Science, Zhejiang University, China guoxinzju@gmail.com panby@zju.edu.cn {dengcai, xiaofeihe}@cad.zju.edu.cn |
| Pseudocode | Yes | Algorithm 1 Variational Inference for ALAMF |
| Open Source Code | No | No explicit statement or link to open-source code for the methodology was found in the paper. |
| Open Datasets | Yes | Similar to [Meng et al., 2013], we study a real application using face images captured under varying illumination. We generate some relatively large datasets and some relatively small datasets in the experiments. Firstly, a larger dataset was built by using the first and fifth subsets of Extended Yale B datasets(Georghiades, Belhumeur and Kriegman 2001;Basri and Jacobs 2003). |
| Dataset Splits | No | The paper describes data corruption and missingness levels, but does not provide explicit training, validation, or test dataset splits in terms of percentages or sample counts for model training/evaluation in the typical supervised learning sense. For example, '20% of entries in Ygt as missing data' refers to input data conditions, not a dataset split. |
| Hardware Specification | No | No specific details about the hardware used for running experiments (e.g., CPU/GPU models, memory) are provided in the paper. |
| Software Dependencies | No | No specific software dependencies with version numbers (e.g., Python 3.8, PyTorch 1.9) are mentioned in the paper. |
| Experiment Setup | Yes | For all the experiments we have conducted, the hyperparameters of ALAMF are fixed without further tuning: a0 = b0 = 10 4, a1 = b1 = 0.1, α = 1. For all the methods, we set the rank of the lowrank component to 8 and apply the random initialization strategy to U and V. |