Task-Driven Common Representation Learning via Bridge Neural Network

Authors: Yao Xu, Xueshuang Xiang, Meiyu Huang5573-5580

AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The experiments on the tasks, including pair matching, canonical correlation analysis, transfer learning, and reconstruction demonstrate the state-of-the-art performance of BNN.
Researcher Affiliation Collaboration Yao Xu,1,2 Xueshuang Xiang,1,2, Meiyu Huang1 1Qian Xuesen Laboratory of Space Technology, China Academy of Space Technology, Beijing, 100190 2School of Aerospace Science and Technology, Xidian University, Xian, 710071
Pseudocode Yes Algorithm 1 Training process of BNN
Open Source Code No The paper does not contain any explicit statement about providing open-source code for the described methodology or a link to a code repository.
Open Datasets Yes We perform our experiments based on two datasets commonly used in the recent literature for CCA test: MNIST (Le Cun et al. 1998) half matching and X-Ray Microbeam Speech data (XRMB) (Westbury 1994).
Dataset Splits Yes The database contains 60,000 training images and 10,000 testing images of 28x28 pixels. (for MNIST) and In our experiment, 60,000 random samples are used for training and 10,000 for testing. (for XRMB).
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory) used for running the experiments, only general mentions of 'lightweight convolution layers'.
Software Dependencies No The paper mentions using 'Scikitlearn' but does not provide specific version numbers for any software dependencies or libraries.
Experiment Setup Yes The output dimension n of both networks is set to 50 for MNIST dataset, 112 for XRMB dataset, and the judging threshold γ is set to 0.5. and Define balance parameter α, learning rate η and NP ratio ξ; from Algorithm 1.