Introspective Classification with Convolutional Nets

Authors: Long Jin, Justin Lazarow, Zhuowen Tu

NeurIPS 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct experiments on benchmark datasets including MNIST, CIFAR10, and SVHN using state-of-the-art CNN architectures, and observe improved classification results.
Researcher Affiliation Academia Long Jin UC San Diego longjin@ucsd.edu Justin Lazarow UC San Diego jlazarow@ucsd.edu Zhuowen Tu UC San Diego ztu@ucsd.edu
Pseudocode Yes Algorithm 1 Outline of the reclassification-by-synthesis algorithm for discriminative classifier training.
Open Source Code No The paper does not provide concrete access to its own source code for the methodology described.
Open Datasets Yes We conduct experiments on three standard benchmark datasets, including MNIST, CIFAR-10 and SVHN.
Dataset Splits Yes We use the standard MNIST [24] dataset, which consists of 55, 000 training, 5, 000 validation and 10, 000 test samples.
Hardware Specification No The paper does not provide specific hardware details (exact GPU/CPU models, processor types with speeds, memory amounts, or detailed computer specifications) used for running its experiments.
Software Dependencies No The paper mentions the use of 'SGD optimizer' and 'Adam optimizer [17]' but does not provide specific version numbers for these or other software libraries/dependencies.
Experiment Setup Yes In our experiments, for the reclassification step, we use the SGD optimizer with mini-batch size of 64 (MNIST) or 128 (CIFAR-10 and SVHN) and momentum equal to 0.9; for the synthesis step, we use the Adam optimizer [17] with momentum term β1 equal to 0.5.