This project explores active learning techniques, focusing on query strategies to optimize informative data selection for model training. It aims to reduce labeled data while improving model performance, especially with imbalanced datasets where certain classes are underrepresented.

This project demonstrates active learning for text classification using the Small-Text library on the IMDB dataset. A logistic regression model is trained iteratively, selecting the most uncertain samples for labeling with a smart query strategy. The approach highlights efficient learning with minimal labeled data, improving model performance.

主动学习相关

I built a data labaling platform with active learing and human feedback in the loop.

python code + pytorch + active learning + LeNet - 5 model

This project explores the implementation of active learning techniques, focusing on various query strategies to optimize the selection of informative data points for model training. It aims to reduce the amount of labeled data required while improving model performance, especially in scenarios with limited labeled data.

I applied active learning to the IMDB dataset for sentiment analysis. Starting with a small labeled subset, I trained a model and used uncertainty sampling to select and label challenging reviews. This iterative process improved performance while reducing labeling effort.

robotics

Classifying motion of an autonomous robot as translation or rotation - a project which was part of Machine Learning course at BITS Pilani

Code base for actively eliciting expensive/ cumbersome features of the most informative data instance. (IJCAI 2018)

When Active Learning and Data Augmentation meet at Object Detection

Official implementation of our AISTATS 2023 paper "FAIR: Fair Collaborative Active Learning with Individual Rationality for Scientific Discovery".

QBC research project at Cornell

The code in this repository makes it possible to reproduce the results of my master thesis "Evaluation of the performance of neural network models and classical models in the context of Active Learning for Systematic Reviewing"

Strata 2018 Tutorial: R and Python for Scalable Data Science