Unknown-Aware Graph Regularization for Robust Semi-supervised Learning from Uncurated Data

Authors: Heejo Kong, Suneung Kim, Ho-Joong Kim, Seong-Whan Lee

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

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments demonstrate that UAG surpasses state-of-the-art OSSL methods by a large margin across various protocols. Codes are available at https://github.com/heejokong/UAGreg.
Researcher Affiliation Academia 1Department of Brain and Cognitive Engineering, Korea University, Seoul, South Korea 2Department of Artificial Intelligence, Korea University, Seoul, South Korea
Pseudocode No The paper describes its approach and formulations in text and equations, but does not include any explicit pseudocode or algorithm blocks.
Open Source Code Yes Codes are available at https://github.com/heejokong/UAGreg.
Open Datasets Yes We conduct extensive experiments on CIFAR-10/100 (Krizhevsky, Hinton et al. 2009) and Image Net-30 (Hendrycks et al. 2019) datasets, following the previous benchmarks.
Dataset Splits Yes For CIFAR-10 (CIFAR-100), we divide classes into 6 known (60 known) and 4 unknown (40 unknown) classes. All samples, except for the labeled data, are designated as unlabeled data.
Hardware Specification No The paper does not explicitly describe the specific hardware (e.g., GPU model, CPU model, memory) used for running its experiments.
Software Dependencies No The paper mentions software frameworks like 'WRN-28' and 'Res Net-18' which are model architectures, but does not provide specific software dependencies with version numbers (e.g., Python, PyTorch, TensorFlow versions).
Experiment Setup Yes For an outlier detector, T and m are set to 1.5 and 0.9, respectively.