Uncertainty Based Active Learning

Uncertainty-based active learning (UAL) aims to efficiently train machine learning models by strategically selecting the most informative data points for labeling, minimizing annotation costs. Current research focuses on improving UAL's robustness to model mismatches and calibration issues, often employing Bayesian neural networks, ensemble methods, and techniques like Low-Rank Adaptation (LoRA) for large language models. These advancements are particularly impactful in resource-constrained domains like medical imaging and robotics, enabling faster and more cost-effective model development for applications requiring high accuracy with limited labeled data.

Papers

October 25, 2024

Efficient Biological Data Acquisition through Inference Set Design
Ihor Neporozhnii, Julien Roy, Emmanuel Bengio, Jason Hartford
Active Learning Drug Discovery Inference Framework Uncertainty Based Active Learning Data Acquisition

September 25, 2024

Reactive Multi-Robot Navigation in Outdoor Environments Through Uncertainty-Aware Active Learning of Human Preference Landscape
Chao Huang, Wenshuo Zang, Carlo Pinciroli, Zhi Jane Li, Taposh Banerjee, Lili Su, Rui Liu
Multi Robot System Preference Learning Multi Robot Navigation Robot Safety Reactive Behavior Uncertainty Based Active Learning Outdoor Environment

August 24, 2024

Understanding Uncertainty-based Active Learning Under Model Mismatch
Amir Hossein Rahmati, Mingzhou Fan, Ruida Zhou, Nathan M. Urban, Byung-Jun Yoon, Xiaoning Qian
Prediction Uncertainty Prediction Performance Uncertainty Based Active Learning Model Mismatch Model Capacity

July 2, 2024

CALICO: Confident Active Learning with Integrated Calibration
Lorenzo S. Querol, Hajime Nagahara, Hideaki Hayashi
Deep Learning Active Learning Uncertainty Based Active Learning Integrated Calibration

June 16, 2024

Incorporating uncertainty quantification into travel mode choice modeling: a Bayesian neural network (BNN) approach and an uncertainty-guided active survey framework
Shuwen Zheng, Zhou Fang, Liang Zhao
Uncertainty Quantification Bayesian Neural Network Travel Mode BNN Layer Uncertainty Based Active Learning Mode Choice

April 2, 2024

Uncertainty-aware Active Learning of NeRF-based Object Models for Robot Manipulators using Visual and Re-orientation Actions
Saptarshi Dasgupta, Akshat Gupta, Shreshth Tuli, Rohan Paul
3d Representation Robotic Manipulator Unseen 3D NeRF Method Unseen Object Uncertainty Based Active Learning Real World Reorientation Task Franka Emika Robot

March 2, 2024

STAR: Constraint LoRA with Dynamic Active Learning for Data-Efficient Fine-Tuning of Large Language Models
Linhai Zhang, Jialong Wu, Deyu Zhou, Guoqiang Xu
Active Learning Shot Learning Complex Reasoning Task Star Model Safe LoRA Uncertainty Based Active Learning Data Efficient Fine Tuning

February 23, 2024

Bidirectional Uncertainty-Based Active Learning for Open Set Annotation
Chen-Chen Zong, Ye-Wen Wang, Kun-Peng Ning, Haibo Ye, Sheng-Jun Huang
Active Learning Open Set Bidirectional Correspondence Uncertainty Based Active Learning Negative Learning Unknown Class

January 29, 2024

Breaking the Barrier: Selective Uncertainty-based Active Learning for Medical Image Segmentation
Siteng Ma, Haochang Wu, Aonghus Lawlor, Ruihai Dong
Active Learning Medical Image Segmentation Knowledge Barrier Uncertainty Based Active Learning

December 21, 2023

On the Convergence of Loss and Uncertainty-based Active Learning Algorithms
Daniel Haimovich, Dima Karamshuk, Fridolin Linder, Niek Tax, Milan Vojnovic
Loss Function Stochastic Gradient Descent Early Stage Convergence Interest Loss Uncertainty Based Active Learning Smooth Loss Function

July 26, 2023

Hybrid Representation-Enhanced Sampling for Bayesian Active Learning in Musculoskeletal Segmentation of Lower Extremities
Ganping Li, Yoshito Otake, Mazen Soufi, Masashi Taniguchi, Masahide Yagi, Noriaki Ichihashi, Keisuke Uemura, Masaki Takao, Nobuhiko Sugano, Yoshinobu Sato
Medical Image Datasets Efficient Hybrid Bayesian Active Learning Uncertainty Based Active Learning Lower Limb Musculoskeletal Segmentation Hybrid Frontier Sampling

March 22, 2023

Uncertainty Aware Active Learning for Reconfiguration of Pre-trained Deep Object-Detection Networks for New Target Domains
Jiaming Na, Varuna De-Silva
Object Detection Model Target Domain Active Learning Algorithm Pre Trained Object Uncertainty Based Active Learning

February 23, 2023

Active Prompting with Chain-of-Thought for Large Language Models
Shizhe Diao, Pengcheng Wang, Yong Lin, Tong Zhang
Chain of Thought Complex Reasoning Task Interactive Prompting Uncertainty Based Active Learning Complex Question Answering Prompt Example

February 11, 2023

Pushing the Accuracy-Group Robustness Frontier with Introspective Self-play
Jeremiah Zhe Liu, Krishnamurthy Dj Dvijotham, Jihyeon Lee, Quan Yuan, Martin Strobel, Balaji Lakshminarayanan, Deepak Ramachandran
Uncertainty Estimation Empirical Risk Minimization Self Play Group Robustness Uncertainty Based Active Learning Worst Group Accuracy

February 1, 2023

Active Uncertainty Reduction for Safe and Efficient Interaction Planning: A Shielding-Aware Dual Control Approach
Haimin Hu, David Isele, Sangjae Bae, Jaime F. Fisac
Interaction Aware Robust Planning Uncertainty Based Active Learning Dual Control

June 13, 2022

On the reusability of samples in active learning
Gijs van Tulder, Marco Loog
Active Learning External Sample Reusable Data Exhaust Uncertainty Based Active Learning

April 18, 2022

Active Learning Helps Pretrained Models Learn the Intended Task
Alex Tamkin, Dat Nguyen, Salil Deshpande, Jesse Mu, Noah Goodman
Active Learning Uncertainty Based Active Learning Task Ambiguity

February 15, 2022

Active Uncertainty Reduction for Human-Robot Interaction: An Implicit Dual Control Approach
Haimin Hu, Jaime F. Fisac
Human Robot Interaction Uncertainty Based Active Learning Dual Control

February 10, 2022

Improving performance of aircraft detection in satellite imagery while limiting the labelling effort: Hybrid active learning
Julie Imbert, Gohar Dashyan, Alex Goupilleau, Tugdual Ceillier, Marie-Caroline Corbineau
Active Learning System Performance Satellite Imagery Uncertainty Based Active Learning Labeling Effort Aircraft Detection

Uncertainty Based Active Learning

Papers

Efficient Biological Data Acquisition through Inference Set Design

Reactive Multi-Robot Navigation in Outdoor Environments Through Uncertainty-Aware Active Learning of Human Preference Landscape

Understanding Uncertainty-based Active Learning Under Model Mismatch

CALICO: Confident Active Learning with Integrated Calibration

Incorporating uncertainty quantification into travel mode choice modeling: a Bayesian neural network (BNN) approach and an uncertainty-guided active survey framework

Uncertainty-aware Active Learning of NeRF-based Object Models for Robot Manipulators using Visual and Re-orientation Actions

STAR: Constraint LoRA with Dynamic Active Learning for Data-Efficient Fine-Tuning of Large Language Models

Bidirectional Uncertainty-Based Active Learning for Open Set Annotation

Breaking the Barrier: Selective Uncertainty-based Active Learning for Medical Image Segmentation

On the Convergence of Loss and Uncertainty-based Active Learning Algorithms

Hybrid Representation-Enhanced Sampling for Bayesian Active Learning in Musculoskeletal Segmentation of Lower Extremities

Uncertainty Aware Active Learning for Reconfiguration of Pre-trained Deep Object-Detection Networks for New Target Domains

Active Prompting with Chain-of-Thought for Large Language Models

Pushing the Accuracy-Group Robustness Frontier with Introspective Self-play

Active Uncertainty Reduction for Safe and Efficient Interaction Planning: A Shielding-Aware Dual Control Approach

On the reusability of samples in active learning

Active Learning Helps Pretrained Models Learn the Intended Task

Active Uncertainty Reduction for Human-Robot Interaction: An Implicit Dual Control Approach

Improving performance of aircraft detection in satellite imagery while limiting the labelling effort: Hybrid active learning