Neural Network Editor

Editor Description

The neural network editor allows users to graphically construct the neural network circuits that control the behavior of the biomechanical model of the organism. The user can do this by dragging neurons from the various neural plug-in modules shown in the toolbox onto the diagram and dropping them where they want them located. Then they can draw links between the neurons to add synapses. You can have multiple diagram sheets and they can be arranged in a hierarchal manner. New diagrams can be added manually through the menu, or you can add sub-diagrams. Each diagram shows up at the top of the page as a new tab.
The editor window is composed of a number of toolbars that allow you to control your circuits. Below is a list of each part of the editor window along with a description. You can also click on any part of the image in figure 1 to directly jump to a description.

To learn more about what a neural plug-in module is and which ones are available then click Neural Simulation Plug-ins
Figure 1. neural network editor Window. Click on any part of the window to jump to a description of that part.

Editor Parts

Figure 2. Toolbox.
This is a toolbar that has all the nodes (neurons) that you can add to the editor diagram. The toolbar has a band for each plug-in module that contains nodes that can be used. In this figure this is the grey bands named AnimatTools, Fast Neural Network, etc.. If you click on one of these bands it will display only the nodes available from that module. The band labeled Alphabetical Tools displays all nodes from all the modules in alphabetical order. For more information on what neural plug-in is you can click here.
Figure 3. Properties Toolbar.
This toolbar displays the properties of any selected item on the currently selected diagram. This allows you to view and modify the properties of the selected object. If no object is selected then the properties for the current diagram are displayed. The properties are split into functional groups like 'Graphical Properties', 'Neural properties' and so on. However, you can view an alphabetical list of the objects properties by clicking the AZ button that is just above the properties window.
Figure 4. Hierarchy Toolbar.
You can have as many diagram windows as you need to build your neural circuits, and you can arrange them in a hierarchal manner using sub-diagrams. This allows you to simply your designs by splitting up complex networks into groups of simpler networks that are connected to each other. It also helps reduce the time it takes to build a circuit because if you have a network module that is repeated, such as the central pattern generator for a leg, then you only have to build this once as a sub-diagram, then you can just copy and paste and modify it. This keeps you from having to lay everything out again by hand.

The hierarchy toolbar provides a tree view perspective of all the diagrams, nodes, and links in your neural circuitry. Each diagram has its own branch in the tree, and those branches can be nested. Underneath each diagram branch there are items for Links and Nodes. Under each of these items is listed all the links and branches for this diagram. You can click on any of these items to select that item. When you do that item will be highlighted in the diagram and its properties will appear in the property editor toolbar. You can expand and collapse any branch by right clicking on that item and displaying a pop-up menu. Menu items for 'Collapse All' and 'Expand All'. If you right click on any diagram node the pop-up menu also contains items to 'Insert Diagram' and 'Delete'. When you click on the diagram node that diagram is displayed. If you select 'Insert Diagram' a new diagram will be inserted as a child of the one you selected. If you select 'Delete' it will delete the selected diagram and ALL of its children. So be careful when deleting diagrams. If you right click on a node or a link a menu entry for deleting that item is also displayed.

Figure 5. Modules Toolbar.
This toolbar is used to view/edit the global properties of the neural plug-in modules for this organism. You can learn more about what neural plug-ins are and how to use them by clicking here. The plug-ins that have been discovered are listed in the top box named 'Neural Modules'. If you click on one of these the global properties for that module are displayed in the property box below that list so you can view/edit them. You can find out more details about the properties that are displayed for each module by going here.

Figure 6. Neural Network Errors Toolbar.
When a network grows large it can become difficult to pinpoint problems, or understand why something is not working the way you think it should. Sometimes this happens because you have made a simple mistake and overlooked something. For instance, You might run into a situation where you are trying to stimulate a muscle to contract and you have your neural network setup and you have looked at all the outputs and everything seems like it should be working, but you get no tension generated by the muscle. One possible reason could be that you simply forgot to associate your muscle node in the neural network editor with an actual muscle in the body of your organism. If you do not do this then the simulator will not know which muscle it is trying to contract.

The errors toolbar is meant to try and help you with problems like this one. It looks for some of the more likely potential errors and lists them for you. This includes things like not associating real body parts to items in the neural network editor and having neurons that are not connected to any other neurons. If you are having a weird problem with the behavior of your network you should first check this dialog to see if any errors appear. If they do then you can double click on the error and it will take you to the diagram where the relevant item is located and highlight it for you automatically so you can see its properties.


This is a series of tabs along the top of the window for each diagram in the editor. You can click on the left and right arrow buttons on the right side to scroll the view to tabs that are hidden from view because of the size of the screen.

Diagram
This is the working area for this diagram. It is where you add new nodes and create links between them.

Project Menus & Toolbars

The following items can be found in the toolbar, main menu, and also in pop-up menus for this editor. You can view pop-up menus by selecting items in the Hierarchy tree or on the diagrams and then right clicking on them to display the pop-up menu.


Add Stimulus

Opens the dialog to add a new stimulus to the selected item.


Delete Item

Deletes the currently selected item.


Select By Type

Selects all items of the same object type. Click here for more details.


Relabel Items

Use regular expressions to rename items and child subsystems. Click here for more details and a video tutorial.


Relabel Selected Items

Relabel and number selected items. Click here for more details and a video tutorial.


Zoom In

Zoom into the diagram.


Zoom out

Zoom out of the diagram.


Align Selected Items

Aligns the currently selected items. The items are aligned relative to the last item selected. There are several choices of how to align the parts.


Distribute Selected Items

Distributes the currently selected items. The items are distributed relative to the first and last items selected. There are several choices of how to distribute the parts.


Size Selected Items

Resizes the currently selected items. The items are resized relative to the last item selected. You can resize the length, height, or both.