DRAW: A Recurrent Neural Network For Image Generation

Authors: Karol Gregor, Ivo Danihelka, Alex Graves, Danilo Rezende, Daan Wierstra

ICML 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Section 4 provides experimental results on the MNIST, Street View House Numbers and CIFAR-10 datasets, with examples of generated images; and concluding remarks are given in Section 5. We assess the ability of DRAW to generate realisticlooking images by training on three datasets of progressively increasing visual complexity: MNIST (Le Cun et al., 1998), Street View House Numbers (SVHN) (Netzer et al., 2011) and CIFAR-10 (Krizhevsky, 2009).
Researcher Affiliation Industry Karol Gregor KAROLG@GOOGLE.COM Ivo Danihelka DANIHELKA@GOOGLE.COM Alex Graves GRAVESA@GOOGLE.COM Danilo Jimenez Rezende DANILOR@GOOGLE.COM Daan Wierstra WIERSTRA@GOOGLE.COM Google Deep Mind
Pseudocode No The paper describes the model's equations and iterative steps but does not present them in a formally labeled pseudocode or algorithm block.
Open Source Code No The paper mentions an accompanying video but does not provide an explicit statement about releasing source code for the methodology or a link to a code repository.
Open Datasets Yes We assess the ability of DRAW to generate realisticlooking images by training on three datasets of progressively increasing visual complexity: MNIST (Le Cun et al., 1998), Street View House Numbers (SVHN) (Netzer et al., 2011) and CIFAR-10 (Krizhevsky, 2009).
Dataset Splits Yes The SVHN training set contains 231,053 images, and the validation set contains 4,701 images.
Hardware Specification No No specific hardware details (e.g., CPU, GPU models, or memory) used for running experiments were mentioned.
Software Dependencies No The paper mentions using LSTM and the Adam optimization algorithm but does not specify version numbers for any software dependencies or libraries.
Experiment Setup Yes Network hyper-parameters for all the experiments are presented in Table 3. The Adam optimisation algorithm (Kingma & Ba, 2014) was used throughout. Table 3 shows "Task #glimpses LSTM #h #z Read Size Write Size" with specific numerical values for each.