Efficient Active Learning

Efficient active learning aims to minimize the cost of labeling data for training machine learning models by strategically selecting the most informative unlabeled samples for annotation. Current research focuses on improving data selection strategies, often incorporating techniques like uncertainty sampling, diversity maximization, and leveraging pre-trained models (e.g., vision transformers, large language models) to accelerate the process. These advancements are crucial for addressing the high annotation costs associated with many machine learning tasks, particularly in domains like medical image analysis and autonomous driving, leading to more efficient and cost-effective model development.

Papers

September 14, 2024

MALADY: Multiclass Active Learning with Auction Dynamics on Graphs
Gokul Bhusal, Kevin Miller, Ekaterina Merkurjev
Graph Drawing Active Learning Semi Supervised Learning Efficient Active Learning Graph Based Active Learning Multiclass Learning

August 7, 2024

Leveraging Variation Theory in Counterfactual Data Augmentation for Optimized Active Learning
Simret Araya Gebreegziabher, Kuangshi Ai, Zheng Zhang, Elena L. Glassman, Toby Jia-Jun Li
Active Learning Text Classification Human Learning Annotated Training Data Counterfactual Data Augmentation Efficient Active Learning Variational Principle

June 13, 2024

May 15, 2024

May 1, 2024

Predictive Accuracy-Based Active Learning for Medical Image Segmentation
Jun Shi, Shulan Ruan, Ziqi Zhu, Minfan Zhao, Hong An, Xudong Xue, Bing Yan
Active Learning Medical Image Segmentation Segmentation Task Segmentation Accuracy Efficient Active Learning

April 8, 2024

AnchorAL: Computationally Efficient Active Learning for Large and Imbalanced Datasets
Pietro Lesci, Andreas Vlachos
Active Learning Imbalanced Datasets Imbalanced Classification Efficient Active Learning Pool Based Active Learning

March 2, 2024

Feature Alignment: Rethinking Efficient Active Learning via Proxy in the Context of Pre-trained Models
Ziting Wen, Oscar Pizarro, Stefan Williams
Active Learning Pre Trained Model Context Information Feature Alignment Optimization Proxy Transparent Proxy Server Architecture Efficient Active Learning

February 23, 2024

Towards Efficient Active Learning in NLP via Pretrained Representations
Artem Vysogorets, Achintya Gopal
Active Learning Text Classification NLP Field Pre Trained Representation Efficient Active Learning Annotation Effort

June 20, 2023

Annotation Cost Efficient Active Learning for Content Based Image Retrieval
Julia Henkel, Genc Hoxha, Gencer Sumbul, Lars Möllenbrok, Begüm Demir
Image Pair Deep Metric Learning Content Based Image Retrieval Efficient Active Learning Annotation Efficient Annotated Training Image

May 22, 2023

Data-efficient Active Learning for Structured Prediction with Partial Annotation and Self-Training
Zhisong Zhang, Emma Strubell, Eduard Hovy
Active Learning Pseudo Label Self Training Structured Prediction Selective Annotation Label Space Efficient Active Learning

June 17, 2022

Towards Efficient Active Learning of PDFA
Franz Mayr, Sergio Yovine, Federico Pan, Nicolas Basset, Thao Dang
Active Learning Multiplier Free Quantization Token Prediction Efficient Active Learning Probabilistic Deterministic Finite Automaton

March 31, 2022

Efficient Active Learning with Abstention
Yinglun Zhu, Robert Nowak
Active Learning Abstention Ability Efficient Active Learning Passive Learning Label Complexity

February 16, 2022

FAMIE: A Fast Active Learning Framework for Multilingual Information Extraction
Minh Van Nguyen, Nghia Trung Ngo, Bonan Min, Thien Huu Nguyen
BERT Model Information Extraction Individual Annotator Active Learning Framework Efficient Active Learning

December 15, 2021

Towards General and Efficient Active Learning
Yichen Xie, Masayoshi Tomizuka, Wei Zhan
Semantic Segmentation Active Learning Model Training General Text Efficient Active Learning