Contrastive Loss Gradient Attack (CLGA)

Official implementation of Unsupervised Graph Poisoning Attack via Contrastive Loss Back-propagation, WWW22

Built based on GCA and DeepRobust.

Requirements

Tested on pytorch 1.7.1 and torch_geometric 1.6.3.

Usage

1.To produce poisoned graphs with CLGA

python CLGA.py --dataset Cora --num_epochs 3000 --device cuda:0

It will automatically save three poisoned adjacency matrices in ./poisoned_adj which have 1%/5%/10% edges perturbed respectively. You may reduce the number of epochs for a faster training.

2.To produce poisoned graphs with baseline attack methods

python baseline_attacks.py --dataset Cora --method dice --rate 0.10 --device cuda:0

It will save one poisoned adjacency matrix in ./poisoned_adj.

3.To train the graph contrastive model with the poisoned graph

python train_GCA.py --dataset Cora --perturb --attack_method CLGA --attack_rate 0.10 --device cuda:0

It will load and train on the corresponding poisoned adjacency matrix specified by dataset, attack_method and attack_rate.

Contrastive Loss Gradient Attack (CLGA)

Related tags

Overview

Contrastive Loss Gradient Attack (CLGA)

Requirements

Usage

Owner

Simple Text-Generator with OpenAI gpt-2 Pytorch Implementation

LF-YOLO (Lighter and Faster YOLO) is used to detect defect of X-ray weld image.

Moer Grounded Image Captioning by Distilling Image-Text Matching Model

A new version of the CIDACS-RL linkage tool suitable to a cluster computing environment.

This repository contains the source code of our work on designing efficient CNNs for computer vision

TensorFlow implementation of Deep Reinforcement Learning papers

Negative Sample Matters: A Renaissance of Metric Learning for Temporal Grounding

Character-Input - Create a program that asks the user to enter their name and their age

📖 Deep Attentional Guided Image Filtering

PyTorch-LIT is the Lite Inference Toolkit (LIT) for PyTorch which focuses on easy and fast inference of large models on end-devices.

An attempt at the implementation of Glom, Geoffrey Hinton's new idea that integrates neural fields, predictive coding, top-down-bottom-up, and attention (consensus between columns)

CSAW-M: An Ordinal Classification Dataset for Benchmarking Mammographic Masking of Cancer

PyTorch implementation of Federated Learning with Non-IID Data, and federated learning algorithms, including FedAvg, FedProx.

Robotics environments

In the case of your data having only 1 channel while want to use timm models

Dataset for the Research2Clinics @ NeurIPS 2021 Paper: What Do You See in this Patient? Behavioral Testing of Clinical NLP Models

StudioGAN is a Pytorch library providing implementations of representative Generative Adversarial Networks (GANs) for conditional/unconditional image generation.

Back to Basics: Efficient Network Compression via IMP

Library of deep learning models and datasets designed to make deep learning more accessible and accelerate ML research.

Train robotic agents to learn pick and place with deep learning for vision-based manipulation in PyBullet.