/team_neural_network

Primary LanguageJupyter Notebook

Neural Network Team

This is the home repo for the Neural Network team.

Team Members

Adam Stafford, Tianhao Zang

Objective

To explore the implementation of nucleotide data into a neural network.

To do

Priority Items

  • Make subsets of sequences to do diagnostics on
  • Diagnostics: Build nueral network and run
    • try a few different architectures, no more than three layers
    • assess time on different number of subsets, quantify and visualize
    • assess memory on different number of subsets, quantify and visualize

Other

  • Explore other options for data structuring of nucleotide data beyond one hot encoding
    • read section 2.2 in Deep Learning in Python
    • Make notes on what you are thinking and upload to notes.md
  • Identify network Architecture, which layers?
    • What sort of layers and in which confirmation are we going to use?
  • Get data into one hot encoding format
    • classification 1: binary (presence or absence of GAL4 signal)
    • classification 2: classification of stage of signal

The data

The data is located on Google Drive. Please use data/3.species_24_only_07March2018

Key points to keep in mind

  • Structual organization of nucleotide input must be upheld.

Platforms

Main Language: Python Implementation: TensorFlow, Keras

Data Management