Single-Photon Image Classification

Authors: Thomas Fischbacher, Luciano Sbaiz

ICLR 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We demonstrate the existence of a quantum computing toy model that illustrates key aspects of quantum information processing while being experimentally accessible with room temperature optics. ... We further demonstrate that a classifier that is permitted to employ quantum interference by optically transforming the photon state prior to detection can achieve a classification accuracy of at least 41.27% for MNIST (respectively 36.14% for Fashion-MNIST ).
Researcher Affiliation Academia Anonymous authors Paper under double-blind review. The paper states 'Anonymous authors Paper under double-blind review', which explicitly hides affiliation information. Therefore, a classification cannot be made.
Pseudocode No No structured pseudocode or algorithm blocks were found in the paper.
Open Source Code Yes Tensor Flow code to both train such a model and also evaluate its performance is included in the supplementary material. ... Tensor Flow2 code to reproduce the experiments of this work and all the figures is provided in the ancillary files together with the computed unitary tranformations for MNIST and Fashion-MNIST .
Open Datasets Yes On the MNIST handwritten digit dataset (Le Cun and Cortes (2010)) ... Fashion-MNIST dataset of Xiao et al. (2017)
Dataset Splits No The paper frequently refers to a 'test set' for evaluation but does not provide specific details about the training, validation, or test dataset splits (e.g., percentages or sample counts) used for reproducibility.
Hardware Specification No The paper does not provide specific details about the computational hardware (e.g., CPU/GPU models, memory, specific cloud instances) used for training the models. It only describes optical components for a potential physical experimental setup.
Software Dependencies No The paper mentions 'Tensor Flow' and 'Tensor Flow2' as the framework used, but does not provide specific version numbers for these or any other software dependencies required for reproducibility.
Experiment Setup No The paper describes the training process using cross-entropy loss and mentions using stochastic gradient descent, but it does not provide specific experimental setup details such as hyperparameter values (e.g., learning rate, batch size, number of epochs) or specific optimizer settings.