LADA: Look-Ahead Data Acquisition via Augmentation for Deep Active Learning (NeurIPS 2021)

Requirements

To install requirements:

pip install -r requirements.txt

Training

To train the model(s) in the paper, run this command:

python main.py --data Cifar10 --method LADA

Evaluation

• Data will be downloaded to folder 'data'.
• Result will be recorded to folder 'Results'.

Results

Our model achieves the following performance on active learning settings:

| Model name | FashionMNIST | SVHN | CIFAR-10 | CIFAR-100 | | ----------- |-------------- | ------------- | ------------- | ------------- | | LADA | 83.68% | 75.72% | 53.45% | 46.92% |