An image representation based convolutional network for DNA classification

Authors: Bojian Yin, Marleen Balvert, Davide Zambrano, Alexander Schoenhuth, Sander Bohte

ICLR 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our experiments show that the method outperforms several existing methods both in terms of prediction accuracy and training time.
Researcher Affiliation Academia 1Centrum Wiskunde & Informatica (CWI), Amsterdam, The Netherlands 2Department of Biology, University of Utrecht, The Netherlands
Pseudocode No The paper does not contain any pseudocode or clearly labeled algorithm blocks.
Open Source Code Yes The complete model is presented in Table 1, and code is available on Github (https://github.com/Bojian/Hilbert-CNN/tree/master).
Open Datasets Yes We test the performance of our approach using ten publicly available datasets from Pokholok et al. (2005).
Dataset Splits Yes A randomly chosen 90% of the dataset is used for training the network, 5% is used for validation and early stopping, and the remaining (5%) is used for evaluation.
Hardware Specification No The paper does not specify the hardware used for running the experiments (e.g., GPU/CPU models, memory details).
Software Dependencies No The LSTM model was implemented in Keras (Chollet et al., 2015), all other models were implemented in Tensorflow (Abadi et al., 2015). (No version numbers provided for Keras or Tensorflow).
Experiment Setup Yes The learning rate is set to 0.003, the batch-size was set to 300 samples and the maximum number of epochs is 10. ... The complete model is presented in Table 1.