Out-of-Distribution Detection with An Adaptive Likelihood Ratio on Informative Hierarchical VAE

Authors: Yewen Li, Chaojie Wang, Xiaobo Xia, Tongliang Liu, xin miao, Bo An

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results demonstrate that our method can significantly outperform existing state-of-the-art unsupervised OOD detection approaches. 4 Experiments
Researcher Affiliation Collaboration 1Nanyang Technological University 2University of Sydney 3Amazon
Pseudocode No The paper does not contain structured pseudocode or algorithm blocks.
Open Source Code Yes Did you include the code, data, and instructions needed to reproduce the main experimental results (either in the supplemental material or as a URL)? [Yes] See supplemental material
Open Datasets Yes Fashion MNIST [39] (in) / MNIST [40] (out) and CIFAR10 [41] (in) / SVHN [42] (out)... we add KMNIST [43], not MNIST [44], Omniglot [45] and Small NORB [46] datasets; for CIFAR10/SVHN pair, we add Celeb A [47], Places365 [48], Flower102 [49] and LFWPeople [50] datasets.
Dataset Splits No The paper mentions "trained on the training split" and "evaluated on both the testing split" but does not explicitly provide details about a validation dataset split or its size/percentage for their own experiments.
Hardware Specification Yes All experiments are performed on a PC with an NVIDIA RTX 3090 GPU and the our code is implemented with Py Torch [53].
Software Dependencies Yes The models are implemented in PyTorch 1.10.1.
Experiment Setup Yes For optimization, we adopt the same Adam optimizer [52] with a learning rate of 3e-4. We train all models in comparison by setting the batch size as 128 and the max epoch as 1000.