Towards Maximizing the Representation Gap between In-Domain & Out-of-Distribution Examples

Authors: Jay Nandy, Wynne Hsu, Mong Li Lee

NeurIPS 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct two sets of experiments: First, we experiment on a synthetic dataset. Next, we present a comparative study on a range of image classification tasks. 1 2
Researcher Affiliation Academia Jay Nandy Wynne Hsu Mong Li Lee National University of Singapore {jaynandy,whsu,leeml}@comp.nus.edu.sg
Pseudocode No The paper does not contain structured pseudocode or algorithm blocks.
Open Source Code Yes 2Code Link: https://github.com/jayjaynandy/maximize-representation-gap
Open Datasets Yes We conduct two sets of experiments... Next, we present a comparative study on a range of image classification tasks. ... We carry out experiments on CIFAR-10 and CIFAR-100 [28] and Tiny Image Net [29]... Image Net-25K is obtained by randomly selecting 25, 000 images from the Image Net dataset [30].
Dataset Splits No The paper mentions training and test examples but does not explicitly provide details about a separate validation set or specific train/validation/test splits within the main text.
Hardware Specification No The paper does not provide specific hardware details (e.g., CPU/GPU models, memory) used for running experiments.
Software Dependencies No The paper does not provide specific software dependencies with version numbers.
Experiment Setup No The paper mentions 'λin and λout are user-defined hyper-parameters' and refers to 'Appendix B.1 for additional details on experimental setup, hyper-parameters', but does not provide concrete hyperparameter values or detailed system-level training settings within the main text.