Boosted CVaR Classification (NeurIPS 2021)

Overview

Boosted CVaR Classification

Runtian Zhai, Chen Dan, Arun Sai Suggala, Zico Kolter, Pradeep Ravikumar
NeurIPS 2021

Table of Contents

Quick Start

Before running the code, please install all the required packages in requirements.txt by running:

pip install -r requirements.txt

In the code, we solve linear programs with the MOSEK solver, which requires a license. You can acquire a free academic license from https://www.mosek.com/products/academic-licenses/. Please make sure that the license file is placed in the correct folder so that the solver could work.

Train

To train a set of base models with boosting, run the following shell command:

python train.py --dataset [DATASET] --data_root /path/to/dataset 
                --alg [ALGORITHM] --epochs [EPOCHS] --iters_per_epoch [ITERS]
                --scheduler [SCHEDULER] --warmup [WARMUP_EPOCHS] --seed [SEED]

Use the --download option to download the dataset if you are running for the first time. Use the --save_file option to save your training results into a .mat file. Set the training hyperparameters with --alpha, --beta and --eta.

For example, to train a set of base models on Cifar-10 with AdaLPBoost, use the following shell command:

python train.py --dataset cifar10 --data_root data --alg adalpboost 
                --eta 1.0 --epochs 100 --iters_per_epoch 5000
                --scheduler 2000,4000 --warmup 20 --seed 2021
                --save_file cifar10.mat

Evaluation

To evaluate the models trained with the above command, run:

python test.py --file cifar10.mat

Introduction

In this work, we study the CVaR classification problem, which requires a classifier to have low α-CVaR loss, i.e. low average loss over the worst α fraction of the samples in the dataset. While previous work showed that no deterministic model learning algorithm can achieve a lower α-CVaR loss than ERM, we address this issue by learning randomized models. Specifically we propose the Boosted CVaR Classification framework that learns ensemble models via Boosting. Our motivation comes from the direct relationship between the CVaR loss and the LPBoost objective. We implement two algorithms based on the framework: one uses LPBoost, and the other named AdaLPBoost uses AdaBoost to pick the sample weights and LPBoost to pick the model weights.

Algorithms

We implement three algorithms in algs.py:

Name Description
uniform All sample weight vectors are uniform distributions.
lpboost Regularized LPBoost (set --beta for regularization).
adalpboost α-AdaLPBoost.

train.py only trains the base models. After the base models are trained, use test.py to select the model weights by solving the dual LPBoost problem.

Parameters

All default training parameters can be found in config.py. For Regularized LPBoost we use β = 100 for all α. For AdaLPBoost we use η = 1.0.

Citation and Contact

To cite this work, please use the following BibTex entry:

@inproceedings{zhai2021boosted,
  author = {Zhai, Runtian and Dan, Chen and Suggala, Arun Sai and Kolter, Zico and Ravikumar, Pradeep},
  booktitle = {Advances in Neural Information Processing Systems},
  title = {Boosted CVaR Classification},
  volume = {34},
  year = {2021}
}

To contact us, please email to the following address: Runtian Zhai <[email protected]>

You might also like...
Code to reproduce the experiments from our NeurIPS 2021 paper " The Limitations of Large Width in Neural Networks: A Deep Gaussian Process Perspective"

Code To run: python runner.py new --save SAVE_NAME --data PATH_TO_DATA_DIR --dataset DATASET --model model_name [options] --n 1000 - train - t

Companion code for the paper "An Infinite-Feature Extension for Bayesian ReLU Nets That Fixes Their Asymptotic Overconfidence" (NeurIPS 2021)

ReLU-GP Residual (RGPR) This repository contains code for reproducing the following NeurIPS 2021 paper: @inproceedings{kristiadi2021infinite, title=

This repo includes our code for evaluating and improving transferability in domain generalization (NeurIPS 2021)

Transferability for domain generalization This repo is for evaluating and improving transferability in domain generalization (NeurIPS 2021), based on

Code for MarioNette: Self-Supervised Sprite Learning, in NeurIPS 2021
Code for MarioNette: Self-Supervised Sprite Learning, in NeurIPS 2021

MarioNette | Webpage | Paper | Video MarioNette: Self-Supervised Sprite Learning Dmitriy Smirnov, Michaël Gharbi, Matthew Fisher, Vitor Guizilini, Ale

[NeurIPS 2021] “Improving Contrastive Learning on Imbalanced Data via Open-World Sampling”,
[NeurIPS 2021] “Improving Contrastive Learning on Imbalanced Data via Open-World Sampling”,

Improving Contrastive Learning on Imbalanced Data via Open-World Sampling Introduction Contrastive learning approaches have achieved great success in

Code for Parameter Prediction for Unseen Deep Architectures (NeurIPS 2021)
Code for Parameter Prediction for Unseen Deep Architectures (NeurIPS 2021)

Parameter Prediction for Unseen Deep Architectures (NeurIPS 2021) authors: Boris Knyazev, Michal Drozdzal, Graham Taylor, Adriana Romero-Soriano Overv

[NeurIPS 2021 Spotlight] Aligning Pretraining for Detection via Object-Level Contrastive Learning
[NeurIPS 2021 Spotlight] Aligning Pretraining for Detection via Object-Level Contrastive Learning

SoCo [NeurIPS 2021 Spotlight] Aligning Pretraining for Detection via Object-Level Contrastive Learning By Fangyun Wei*, Yue Gao*, Zhirong Wu, Han Hu,

Official implementation of NeurIPS 2021 paper
Official implementation of NeurIPS 2021 paper "Contextual Similarity Aggregation with Self-attention for Visual Re-ranking"

CSA: Contextual Similarity Aggregation with Self-attention for Visual Re-ranking PyTorch training code for CSA (Contextual Similarity Aggregation). We

Code for our NeurIPS 2021 paper 'Exploiting the Intrinsic Neighborhood Structure for Source-free Domain Adaptation'

Exploiting the Intrinsic Neighborhood Structure for Source-free Domain Adaptation (NeurIPS 2021) Code for our NeurIPS 2021 paper 'Exploiting the Intri

Owner
Runtian Zhai
2nd year PhD at CMU CSD.
Runtian Zhai
Multivariate Boosted TRee

Multivariate Boosted TRee What is MBTR MBTR is a python package for multivariate boosted tree regressors trained in parameter space. The package can h

SUPSI-DACD-ISAAC 61 Dec 19, 2022
Official Pytorch implementation of "Unbiased Classification Through Bias-Contrastive and Bias-Balanced Learning (NeurIPS 2021)

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

Youngkyu 17 Jan 1, 2023
Implement face detection, and age and gender classification, and emotion classification.

YOLO Keras Face Detection Implement Face detection, and Age and Gender Classification, and Emotion Classification. (image from wider face dataset) Ove

Chloe 10 Nov 14, 2022
Simple-Image-Classification - Simple Image Classification Code (PyTorch)

Simple-Image-Classification Simple Image Classification Code (PyTorch) Yechan Kim This repository contains: Python3 / Pytorch code for multi-class ima

Yechan Kim 8 Oct 29, 2022
Image Classification - A research on image classification and auto insurance claim prediction, a systematic experiments on modeling techniques and approaches

A research on image classification and auto insurance claim prediction, a systematic experiments on modeling techniques and approaches

null 0 Jan 23, 2022
Hl classification bc - A Network-Based High-Level Data Classification Algorithm Using Betweenness Centrality

A Network-Based High-Level Data Classification Algorithm Using Betweenness Centr

Esteban Vilca 3 Dec 1, 2022
Official implementation of NeurIPS 2021 paper "One Loss for All: Deep Hashing with a Single Cosine Similarity based Learning Objective"

Official implementation of NeurIPS 2021 paper "One Loss for All: Deep Hashing with a Single Cosine Similarity based Learning Objective"

Ng Kam Woh 71 Dec 22, 2022
[NeurIPS 2021] Galerkin Transformer: a linear attention without softmax

[NeurIPS 2021] Galerkin Transformer: linear attention without softmax Summary A non-numerical analyst oriented explanation on Toward Data Science abou

Shuhao Cao 159 Dec 20, 2022
Code for our NeurIPS 2021 paper Mining the Benefits of Two-stage and One-stage HOI Detection

CDN Code for our NeurIPS 2021 paper "Mining the Benefits of Two-stage and One-stage HOI Detection". Contributed by Aixi Zhang*, Yue Liao*, Si Liu, Mia

null 71 Dec 14, 2022
Reducing Information Bottleneck for Weakly Supervised Semantic Segmentation (NeurIPS 2021)

Reducing Information Bottleneck for Weakly Supervised Semantic Segmentation (NeurIPS 2021) The implementation of Reducing Infromation Bottleneck for W

Jungbeom Lee 81 Dec 16, 2022