Fast Learning of MNL Model From General Partial Rankings with Application to Network Formation Modeling

Last update: Aug 19, 2022

Related tags

Overview

Fast-Partial-Ranking-MNL

This repo provides a PyTorch implementation for the CopulaGNN models as described in the following paper:

Fast Learning of MNL Model From General Partial Rankings with Application to Network Formation Modeling.

Jiaqi Ma*, Xingjian Zhang*, and Qiaozhu Mei. WSDM 2022.

(*: Equal contribution.)

Requirements

The code requires the following packages.

more_itertools==8.10.0
networkx==2.5.1
numpy==1.19.5
pandas==1.1.5
pyclustering==0.10.1.2
torch==1.9.0
tqdm==4.62.3

Example Commands to Run the Experiments

Learning single MNL from partial rankings on synthetic data

python3 dag_synthetic.py --num_classes 100 --num_samples 5000  # single MNL

Learning mixture of MNL from partial rankings on synthetic data

python3 dag_synthetic.py --num_classes 60 --num_samples 5000 --alphas [1,1,1]  --init_by_cluster # 3 MNLs with clustering based init

Network formation modeling of synthetic network data

python3 network_synthetic.py -r 0.5 -p 0.5 --fof --ua --pa --loss topk  # run full model with 4 components on a mixed (r,p)-graph

Network formation modeling of Flickr & Microsoft Academic Graph

cd source
wget -4 http://socialnetworks.mpi-sws.mpg.de/data/flickr-growth.txt.gz ../data/
python3 flickr_process.py # process flickr-growth.txt.gz, which is downloaded from http://socialnetworks.mpi-sws.mpg.de/data/flickr-growth.txt.gz
python3 flickr_train.py

# download mag_cli.csv by google drive
python3 mag_process.py  # process mag_cli.csv, which is downloaded from https://drive.google.com/file/d/17bgLs1iR96JW3Rd0mex3IK8qyU-qRElB/view?usp=sharing
python3 mag_train.py

Cite

@article{ma2022fast,
  title={Fast Learning of MNL Model From General Partial Rankings with Application to Network Formation Modeling},
  author={Ma, Jiaqi and Zhang, Xingjian and Mei, Qiaozhu},
  journal={Proceedings of the 15th ACM International Conference on Web Search and Data Mining},
  year={2022}
}

Fast Learning of MNL Model From General Partial Rankings with Application to Network Formation Modeling

Related tags

Overview

Fast-Partial-Ranking-MNL

Requirements

Example Commands to Run the Experiments

Cite

Owner

Xingjian Zhang

Implementation of ICCV 2021 oral paper -- A Novel Self-Supervised Learning for Gaussian Mixture Model

The official implementation of the paper, "SubTab: Subsetting Features of Tabular Data for Self-Supervised Representation Learning"

Repo for "Event-Stream Representation for Human Gaits Identification Using Deep Neural Networks"

Source code for the Paper: CombOptNet: Fit the Right NP-Hard Problem by Learning Integer Programming Constraints}

Various operations like path tracking, counting, etc by using yolov5

Code for Emergent Translation in Multi-Agent Communication

A map update dataset and benchmark

Application of K-means algorithm on a music dataset after a dimensionality reduction with PCA

This repository contains pre-trained models and some evaluation code for our paper Towards Unsupervised Dense Information Retrieval with Contrastive Learning

A setup script to generate ITK Python Wheels

Official Pytorch implementation of "Unbiased Classification Through Bias-Contrastive and Bias-Balanced Learning (NeurIPS 2021)

Source code for "Pack Together: Entity and Relation Extraction with Levitated Marker"

Segmentation in Style: Unsupervised Semantic Image Segmentation with Stylegan and CLIP

Pytorch implementation for the paper: Contrastive Learning for Cold-start Recommendation

Single Image Deraining Using Bilateral Recurrent Network (TIP 2020)

A collection of Jupyter notebooks to play with NVIDIA's StyleGAN3 and OpenAI's CLIP for a text-based guided image generation.

Code to reproduce the results for Compositional Attention

[ICCV 2021] Amplitude-Phase Recombination: Rethinking Robustness of Convolutional Neural Networks in Frequency Domain

Fully Automatic Page Turning on Real Scores

Code for "Share With Thy Neighbors: Single-View Reconstruction by Cross-Instance Consistency" paper