Adaptive Loss

Adaptive loss functions dynamically adjust the weighting of different loss components during training, aiming to improve model performance and robustness by addressing issues like class imbalance, noisy labels, and uncertainty in predictions. Current research focuses on developing novel weighting schemes, often incorporating uncertainty estimation or task-specific considerations, and applying these to various architectures including convolutional neural networks, transformers, and multi-task learning frameworks. This research is significant because it enhances the accuracy and reliability of machine learning models across diverse applications, from medical image analysis and autonomous driving to natural language processing and recommendation systems.

Papers

March 7, 2023

Guiding Pseudo-labels with Uncertainty Estimation for Source-free Unsupervised Domain Adaptation
Mattia Litrico, Alessio Del Bue, Pietro Morerio
Pseudo Label Uncertainty Estimation Source Free Unsupervised Domain Adaptation Adaptive Loss Multi Target Domain Adaptation

February 28, 2023

Meta-Learning with Adaptive Weighted Loss for Imbalanced Cold-Start Recommendation
Minchang Kim, Yongjin Yang, Jung Hyun Ryu, Taesup Kim
Sequential Recommendation Adaptive Loss Gradient Based Meta Learning Cold Start Recommendation

November 28, 2022

Learning deep abdominal CT registration through adaptive loss weighting and synthetic data generation
Javier Pérez de Frutos, André Pedersen, Egidijus Pelanis, David Bouget, Shanmugapriya Survarachakan, Thomas Langø, Ole-Jakob Elle, Frank Lindseth
Synthetic Data Generation Image Registration Image to Image Adaptive Loss Abdominal CT Registration

November 22, 2022

Mitigating Negative Transfer in Multi-Task Learning with Exponential Moving Average Loss Weighting Strategies
Anish Lakkapragada, Essam Sleiman, Saimourya Surabhi, Dennis P. Wall
Deep Learning Multi Task Learning Single Task Adaptive Loss Balanced Loss Negative Transfer

September 23, 2022

Grouped Adaptive Loss Weighting for Person Search
Yanling Tian, Di Chen, Yunan Liu, Shanshan Zhang, Jian Yang
Multi Task Learning Person Re Identification Person Search Adaptive Loss

September 19, 2022

Provably Uncertainty-Guided Universal Domain Adaptation
Yifan Wang, Lin Zhang, Ran Song, Paul L. Rosin, Yibin Li, Wei Zhang
Unlabeled Target Domain Adaptive Loss Universal Domain Adaptation Label Set

September 15, 2022

Self-Supervised Attention Networks and Uncertainty Loss Weighting for Multi-Task Emotion Recognition on Vocal Bursts
Vincent Karas, Andreas Triantafyllopoulos, Meishu Song, Björn W. Schuller
Self Supervised Speech Emotion Recognition Adaptive Loss Emotional Impact Vocal Burst

July 22, 2022

Complementing Semi-Supervised Learning with Uncertainty Quantification
Ehsan Kazemi
Semi Supervised Learning Semi Supervised Uncertainty Quantification Pseudo Label Adaptive Loss Epistemic Uncertainty Quantification

June 22, 2022

Dynamic Restrained Uncertainty Weighting Loss for Multitask Learning of Vocal Expression
Meishu Song, Zijiang Yang, Andreas Triantafyllopoulos, Xin Jing, Vincent Karas, Xie Jiangjian, Zixing Zhang, Yamamoto Yoshiharu, Bjoern W. Schuller
Multitask Learning Adaptive Loss Lightweight CNN Loss Design Vocal Expression Weight Selection

April 20, 2022

Epistemic Uncertainty-Weighted Loss for Visual Bias Mitigation
Rebecca S Stone, Nishant Ravikumar, Andrew J Bulpitt, David C Hogg
Adaptive Loss Potential Bias Bias Benchmark Visual Bias High Epistemic Uncertainty

April 8, 2022

Global Update Guided Federated Learning
Qilong Wu, Lin Liu, Shibei Xue
Federated Learning Model Convergence Adaptive Loss

April 3, 2022

AdaFace: Quality Adaptive Margin for Face Recognition
Minchul Kim, Anil K. Jain, Xiaoming Liu
Face Recognition Adaptive Loss Face Recognition Performance Adaptive Margin

February 27, 2022

Arrhythmia Classifier Using Convolutional Neural Network with Adaptive Loss-aware Multi-bit Networks Quantization
Hanshi Sun, Ao Wang, Ninghao Pu, Zhiqing Li, Junguang Huang, Hao Liu, Zhi Qi
Deep Learning Convolutional Neural Network Wearable Device Adaptive Loss Arrhythmia Detection Network Quantization Arrhythmia Classification

December 15, 2021

Robust Depth Completion with Uncertainty-Driven Loss Functions
Yufan Zhu, Weisheng Dong, Leida Li, Jinjian Wu, Xin Li, Guangming Shi
Depth Completion Adaptive Loss Sparse Depth

Adaptive Loss

Papers

Guiding Pseudo-labels with Uncertainty Estimation for Source-free Unsupervised Domain Adaptation

Meta-Learning with Adaptive Weighted Loss for Imbalanced Cold-Start Recommendation

Learning deep abdominal CT registration through adaptive loss weighting and synthetic data generation

Mitigating Negative Transfer in Multi-Task Learning with Exponential Moving Average Loss Weighting Strategies

Grouped Adaptive Loss Weighting for Person Search

Provably Uncertainty-Guided Universal Domain Adaptation

Self-Supervised Attention Networks and Uncertainty Loss Weighting for Multi-Task Emotion Recognition on Vocal Bursts

Complementing Semi-Supervised Learning with Uncertainty Quantification

Dynamic Restrained Uncertainty Weighting Loss for Multitask Learning of Vocal Expression

Epistemic Uncertainty-Weighted Loss for Visual Bias Mitigation

Global Update Guided Federated Learning

AdaFace: Quality Adaptive Margin for Face Recognition

Arrhythmia Classifier Using Convolutional Neural Network with Adaptive Loss-aware Multi-bit Networks Quantization

Robust Depth Completion with Uncertainty-Driven Loss Functions