https://github.com/joeydp/logwmf

The Role of Unknown Interactions in Implicit Matrix Factorization — A Probabilistic View
https://github.com/joeydp/logwmf

Last synced: 5 days ago
JSON representation

The Role of Unknown Interactions in Implicit Matrix Factorization — A Probabilistic View

• Host: GitHub
• URL: https://github.com/joeydp/logwmf
• Owner: JoeyDP
• Created: 2024-08-05T00:39:31.000Z (3 months ago)
• Default Branch: main
• Last Pushed: 2024-08-05T00:53:59.000Z (3 months ago)
• Last Synced: 2024-08-05T01:54:28.135Z (3 months ago)
• Language: C++
• Size: 21.5 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# The Role of Unknown Interactions in Implicit Matrix Factorization --- A Probabilistic View

Based on https://github.com/google-research/google-research/tree/master/ials.

## VAE benchmarks

This follows the evaluation protocol and uses the datasets from

[Liang et al., Variational Autoencoders for Collaborative Filtering, WWW '18](https://dl.acm.org/doi/10.1145/3178876.3186150).

### Instructions

1) Install packages `pip install -r requirements.txt`

2) Compile the code

- Download [Eigen](https://eigen.tuxfamily.org/):

```

wget https://gitlab.com/libeigen/eigen/-/archive/3.3.9/eigen-3.3.9.zip

unzip eigen-3.3.9.zip

```

- Create the subdirectories

```

mkdir lib

mkdir bin

```

- Compile the binaries

```

make all

```

3) Download and process the data

```

python generate_data.py --output_dir ./

```

This will generate two sub-directories `ml-20m` and `msd` corresponding respectively to the data sets [MovieLens 20M](https://grouplens.org/datasets/movielens/20m/) and the [Million Song Data](http://millionsongdataset.com/tasteprofile/).

Note: this code is adapted from https://github.com/dawenl/vae_cf/blob/master/VAE_ML20M_WWW2018.ipynb which requires a Python 3 runtime.

4) Run the training and evaluation code. Example usage:

**MovieLens 20M (ML20M)**

```

./bin/ialspp_main --train_data ml-20m/train.csv --test_train_data ml-20m/test_tr.csv \

  --test_test_data ml-20m/test_te.csv --embedding_dim 256 --stddev 0.1 \

  --regularization 0.003 --regularization_exp 1.0 --unobserved_weight 0.1 \

  --epochs 16 --block_size 128 --eval_during_training 0

```

**Million Song Data (MSD)**

```

./bin/ialspp_main --train_data msd/train.csv --test_train_data msd/test_tr.csv \

  --test_test_data msd/test_te.csv --embedding_dim 256 --stddev 0.1 \

  --regularization 0.002 --regularization_exp 1.0 --unobserved_weight 0.02 \

  --epochs 16 --block_size 128 --eval_during_training 0

```

Setting the flag `--eval_during_training` to 1 will run evaluation after each epoch.

## Reproducibility

Logs of runs can be found here: https://drive.google.com/drive/folders/1ElGp6pfxQxUYv2QWvBD-u0AUWIIuaKP_?usp=drive_link.