Synthesizing Samples for Zero-shot Learning

Authors: Yuchen Guo, Guiguang Ding, Jungong Han, Yue Gao

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

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments on benchmarks demonstrate the superiority of the proposed approach to the state-of-the-art ZSL approaches.
Researcher Affiliation Academia School of Software, Tsinghua University, Beijing 100084, China School of Computing & Communications, Lancaster University, UK
Pseudocode No The paper describes methods using text and mathematical equations but does not contain any structured pseudocode or algorithm blocks.
Open Source Code No The paper does not provide an explicit statement about releasing open-source code or a link to a code repository for its methodology.
Open Datasets Yes In this paper, we adopt four benchmark datasets for ZSL. The first is Animal with Attributes (Aw A) [Lampert et al., 2014]... The second is a Pascal-a Yahoo [Farhadi et al., 2009]... The third is SUN scene recognition dataset [Patterson and Hays, 2012]... The fourth is Caltech-UCSD-Birds-200-2011 (CUB) [Wah et al., 2011].
Dataset Splits Yes The first is Animal with Attributes (Aw A) [Lampert et al., 2014] using a standard source-target split with 40 source classes and 10 target classes. The second is a Pascal-a Yahoo [Farhadi et al., 2009]... Following the standard setting, the a Pascal provides the source classes and the a Yahoo provides the target classes. The third is SUN scene recognition dataset [Patterson and Hays, 2012]... Following the standard setting [Jayaraman and Grauman, 2014], 707 scenes are source classes and 10 scenes are target classes. The fourth is Caltech-UCSD-Birds-200-2011 (CUB) [Wah et al., 2011]... We follow the suggested split by Akata et al. [2015] which uses 150 species as source classes and 50 species as target classes.
Hardware Specification No The paper does not explicitly state the specific hardware (e.g., CPU, GPU models, memory) used for running its experiments.
Software Dependencies No The paper mentions using the VGG-19 network and t-SNE, but it does not provide specific version numbers for these or any other software dependencies, making the software environment unreproducible.
Experiment Setup Yes For each target class, 500 samples are synthesized using the reconstruction based distribution. The second is to assume Σc = diag(σc1, ..., σcd) where we only consider the diagonal elements and the other elements are assumed to be 0. In this paper, we simply adopt three kinds of classifiers, SVM, Logistic Regression (LR) and 1NN.