Multi-Task Learning as a Bargaining Game

Aviv Navon

Last update: Dec 26, 2022

Related tags

Deep Learning nash-mtl

Overview

Nash-MTL

Official implementation of "Multi-Task Learning as a Bargaining Game".

Setup environment

conda create -n nashmtl python=3.9.7
conda activate nashmtl
conda install pytorch==1.9.0 torchvision==0.10.0 torchaudio==0.9.0 cudatoolkit=10.2 -c pytorch
conda install pyg -c pyg -c conda-forge

Next -> cd to the unzipped folder i.e. cd RUN -> pip install -e .

QM9 Experiment

To run Nash-MTL:

cd experiments/quantum_chemistry
python trainer.py --method=nashmtl

To train using another MTL method simply replace 'nashmtl' with one of ['cagrad', 'pcgrad', 'mgda', 'ls', 'uw', 'scaleinvls'].

You might also like...

FocusFace: Multi-task Contrastive Learning for Masked Face Recognition

FocusFace This is the official repository of "FocusFace: Multi-task Contrastive Learning for Masked Face Recognition" accepted at IEEE International C

21 Nov 17, 2022

Code of U2Fusion: a unified unsupervised image fusion network for multiple image fusion tasks, including multi-modal, multi-exposure and multi-focus image fusion.

U2Fusion Code of U2Fusion: a unified unsupervised image fusion network for multiple image fusion tasks, including multi-modal (VIS-IR, medical), multi

129 Dec 11, 2022

Multi Task RL Baselines

MTRL Multi Task RL Algorithms Contents Introduction Setup Usage Documentation Contributing to MTRL Community Acknowledgements Introduction M

171 Jan 9, 2023

Multi-Task Temporal Shift Attention Networks for On-Device Contactless Vitals Measurement (NeurIPS 2020)

MTTS-CAN: Multi-Task Temporal Shift Attention Networks for On-Device Contactless Vitals Measurement Paper Xin Liu, Josh Fromm, Shwetak Patel, Daniel M

106 Dec 30, 2022

Implementation of PyTorch-based multi-task pre-trained models

mtdp Library containing implementation related to the research paper "Multi-task pre-training of deep neural networks for digital pathology" (Mormont

27 Oct 14, 2022

Multi Task Vision and Language

12-in-1: Multi-Task Vision and Language Representation Learning Please cite the following if you use this code. Code and pre-trained models for 12-in-

712 Dec 19, 2022

Exploring Relational Context for Multi-Task Dense Prediction [ICCV 2021]

Adaptive Task-Relational Context (ATRC) This repository provides source code for the ICCV 2021 paper Exploring Relational Context for Multi-Task Dense

35 Dec 5, 2022

This code uses generative adversarial networks to generate diverse task allocation plans for Multi-agent teams.

Mutli-agent task allocation This code uses generative adversarial networks to generate diverse task allocation plans for Multi-agent teams. To change

5 Oct 12, 2022

Face Detection and Alignment using Multi-task Cascaded Convolutional Networks (MTCNN)

Face-Detection-with-MTCNN Face detection is a computer vision problem that involves finding faces in photos. It is a trivial problem for humans to sol

3 Oct 7, 2022

Comments

Question about large weights

Thank you for the great work. When I tried the method, I found that the calculated weights are relatively large than other mtl method, like [47.24413258, 732.26542343] vs [0.4, 0.6]

Is it normal? Should I rescale the weights? Because I think the large weights of losses may influence the regularization. Thank you!

opened by yjdy 6
Logging

Hi @AvivNavon, thank you for your great work. Im trying to experiment with sth new, so could you please give me the log files for NYUv2 experiments so that I can benchmark my running?

opened by VietHoang1512 4
Questions about the property of convergence on theorem 5.4 and 5.5
Hi, experts Thanks sharing the excellent work about MTL.

||1/α(t)||≥σK((G(t))⊤G(t))||α(t)|| on the theorem 5.4, where σK((G(t))⊤G(t)) is the smallest singular value of (G(t))⊤G(t).

What the type of norm?

I think this is spectral norm, then ||(G(t))⊤G(t)|| is the maximum singular value of (G(t))⊤G(t) such that the Inequality holds. Is this right?

As σK((G(t))TG(t))→0 we have from continuity that σK(G⊤ ∗G∗)=0, where G∗is the matrix of gradients at θ∗.

Is this implication for arriving Pareto optimal point θ∗ we should check the smallest singular value of (G(t))⊤G(t) every iteration to observe whether to converge or not?
opened by BossunWang 1
Experiment Tracking with Weights & Biases
This pull request introduces experiment tracking with Weights & Biases with two additional parameters:

python trainer.py --method=nashmtl --wandb_project=<project-name> --wandb_entity=<entity-name>

A sample dashboard can be found here.
opened by soumik12345 0

Owner

Aviv Navon

Data Scientist @ Aiola // PhD Student @ BIU

GitHub

Multi-Object Tracking in Satellite Videos with Graph-Based Multi-Task Modeling

TGraM Multi-Object Tracking in Satellite Videos with Graph-Based Multi-Task Modeling, Qibin He, Xian Sun, Zhiyuan Yan, Beibei Li, Kun Fu Abstract Rece

6 Nov 25, 2022

Code and pre-trained models for MultiMAE: Multi-modal Multi-task Masked Autoencoders

MultiMAE: Multi-modal Multi-task Masked Autoencoders Roman Bachmann*, David Mizrahi*, Andrei Atanov, Amir Zamir Website | arXiv | BibTeX Official PyTo

Visual Intelligence & Learning Lab, Swiss Federal Institute of Technology (EPFL)

385 Jan 6, 2023

AdaShare: Learning What To Share For Efficient Deep Multi-Task Learning

AdaShare: Learning What To Share For Efficient Deep Multi-Task Learning (NeurIPS 2020) Introduction AdaShare is a novel and differentiable approach fo

94 Dec 22, 2022

Code for the ICML 2021 paper "Bridging Multi-Task Learning and Meta-Learning: Towards Efficient Training and Effective Adaptation", Haoxiang Wang, Han Zhao, Bo Li.

Bridging Multi-Task Learning and Meta-Learning Code for the ICML 2021 paper "Bridging Multi-Task Learning and Meta-Learning: Towards Efficient Trainin

57 Dec 15, 2022

Multi-Task Learning as a Bargaining Game

Related tags

Overview

Nash-MTL

Setup environment

QM9 Experiment

You might also like...

FocusFace: Multi-task Contrastive Learning for Masked Face Recognition

Code of U2Fusion: a unified unsupervised image fusion network for multiple image fusion tasks, including multi-modal, multi-exposure and multi-focus image fusion.

Multi Task RL Baselines

Multi-Task Temporal Shift Attention Networks for On-Device Contactless Vitals Measurement (NeurIPS 2020)

Implementation of PyTorch-based multi-task pre-trained models

Multi Task Vision and Language

Exploring Relational Context for Multi-Task Dense Prediction [ICCV 2021]

This code uses generative adversarial networks to generate diverse task allocation plans for Multi-agent teams.

Face Detection and Alignment using Multi-task Cascaded Convolutional Networks (MTCNN)

Comments

Question about large weights

Logging

Questions about the property of convergence on theorem 5.4 and 5.5

Experiment Tracking with Weights & Biases

Owner

Aviv Navon

Multi-Object Tracking in Satellite Videos with Graph-Based Multi-Task Modeling

Code and pre-trained models for MultiMAE: Multi-modal Multi-task Masked Autoencoders

AdaShare: Learning What To Share For Efficient Deep Multi-Task Learning

Code for the ICML 2021 paper "Bridging Multi-Task Learning and Meta-Learning: Towards Efficient Training and Effective Adaptation", Haoxiang Wang, Han Zhao, Bo Li.

Repository for "Improving evidential deep learning via multi-task learning," published in AAAI2022

A list of multi-task learning papers and projects.

A list of multi-task learning papers and projects.

RoboDesk A Multi-Task Reinforcement Learning Benchmark

Multi-task Learning of Order-Consistent Causal Graphs (NeuRIPs 2021)

MultiTaskLearning - Multi Task Learning for 3D segmentation