Neural Active Learning Beyond Bandits

Authors: Yikun Ban, Ishika Agarwal, Ziwei Wu, Yada Zhu, Kommy Weldemariam, Hanghang Tong, Jingrui He

ICLR 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We use extensive experiments to evaluate the proposed algorithms, which consistently outperform state-of-the-art baselines.
Researcher Affiliation Collaboration 1University of Illinois Urbana-Champaign, 2IBM Research
Pseudocode Yes Algorithm 1 NEURONAL-S and Algorithm 2 NEURONAL-P are included in the paper.
Open Source Code No The paper does not provide any explicit statement about releasing code or a link to a code repository for its methodology.
Open Datasets Yes We evaluate NEURONAL for both stream-based and pool-based settings on the following six public classification datasets: Adult, Covertype (CT), Magic Telescope (MT), Shuttle [24], Fashion [61], and Letter [18].
Dataset Splits Yes The default label budget is 30% T. We perform hyperparameter tuning on the training set.
Hardware Specification No The paper describes neural network models and training, but it does not specify any hardware details like GPU models, CPU types, or memory used for the experiments.
Software Dependencies No The paper mentions using neural networks and activation functions (e.g., ReLU) but does not specify the versions of any software libraries, frameworks, or programming languages used (e.g., Python, PyTorch, TensorFlow).
Experiment Setup Yes For all NN models, we use the same width m = 100 and depth L = 2. We perform hyperparameter tuning on the training set. Each method has a couple of hyperparameters: the learning rate, number of epochs, batch size, label budget percentage, and threshold (if applicable). During hyperparameter tuning for all methods, we perform a grid search over the values {0.0001, 0.0005, 0.001} for the learning rate, {10, 20, 30, 40, 50, 60, 70, 80, 90} for the number of epochs, {32, 64, 128, 256} for the batch size, {0.1, 0.3, 0.5, 0.7, 0.9} for the label budget percentage and {1, 2, 3, 4, 5, 6, 7, 8, 9} for the threshold (exploration) parameter.