Skip to content

Glossar

Core-Structure

Neuron

Neuron are the smalles entity within the core-structure, which provides the base of the neuronal network. A Neuron is connected one or more Synapses to other nodes.

Synapse

Synapses are the connetions between the nodes. They are not static and can be create, destroyed and modified.

Hexagon

A Hexagon consist of multiple nodes, which are connected by synapses ot nodes in other hexagons. There are basically 3 types:

  1. Input-Hexagons

  2. Output-Hexagons

  3. Core-Hexagons

Cluster

Cluster are again a collection of multiple Hexagons

Data-Set

A Data-Set defines a set of train-data or data, which should be used for requests against the neuronal network. This can be a table in CSV-format or an MNIST-dataset for now.

Infrastrure

Tasks

Tasks an asynchronous operation of the network based on a given Data-Set. There are two types:

  1. Train-Task

For Train-Tasks input-data and desired output must exist in the Data-Set in order to update the neuronal network based on the data.

  1. Request-Task

In Request-Tasks only the input-data are provided for the network in order to generate a output of the previouly trained network. The output is stored in Shiori as Request-Result.

Checkpoint

Checkpoints are the serializied version of a Cluster. The Cluster will be converted into one single blob, written to disc and registered in the database.

Cluster-Template

Cluster-Templates are a custom-formated string, which defines the structure of the Cluster. Basically it describes the sizes and order of the hexagons. See the docu of the cluster-templates.