Locality Preserving Projection for Domain Adaptation with Multi-Objective Learning

Authors: Le Shu, Tianyang Ma, Longin Jan Latecki

AAAI 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The effectiveness of our approach is justified by both theoretical analysis and empirical results on real world data sets. The new feature representation shows better prediction accuracy as our experiments demonstrate.
Researcher Affiliation Academia Le Shu, Tianyang Ma, Longin Jan Latecki Computer and Information Sciences, Temple University Philadephia, PA, 19122, USA {slevenshu,ma.tianyang}@gmail.com, latecki@temple.edu
Pseudocode Yes Algorithm 1: Locality Preserving Projection for Domain Adaption with Multi-Objective Learning
Open Source Code No The paper does not provide concrete access to source code for the methodology described in this paper.
Open Datasets Yes The data set we evaluate first is the USPS handwritten digit database (Hull 1994). We then evaluate our algorithm on the 14 tumor data sets which were published by Ramaswamy et al. (Statnikov et al. 2005), and we downloaded them in the preprocessing version from Statnikov (Statnikov et al. 2005).
Dataset Splits No The paper discusses the impossibility of performing cross-validation for parameter tuning due to the unsupervised nature of the target domain (no labels available), stating: 'A standard way to find the good parameter is through cross-validation. However, we argue that such paradigm may not be suitable for the unsupervised domain adaptation task, because the target data labels are unavailable, which makes it impossible to perform the cross-validation.' However, it does not provide any specific dataset split information (percentages, sample counts, or explicit validation set usage) for its experiments.
Hardware Specification No The paper does not provide specific hardware details (e.g., CPU, GPU models, or memory specifications) used for running the experiments.
Software Dependencies No The paper does not provide specific software dependencies with version numbers.
Experiment Setup No The paper mentions using a '1-nearest neighbor classifier' but does not provide specific experimental setup details such as hyperparameter values, training configurations, or model initialization settings.