IRIS: New version of free Demo now available!

Hi Everybody! (@cmaver @rhyolight), (a business partner of Numenta), is proud to announce version 2 (v2.0.2) of Iris is now available for download!

Iris is a free application written in JavaFX which demonstrates the Natural Language Understanding (NLU) features of’s REST oriented Retina API; and demonstrates the way the Retina API may be used within your applications to bring NLU capabilities to your users. ( also has a more advanced streaming engine called the Retina Library, which can be used for Big Data applications.)

The new version of Iris includes the “Classification” feature, which enables the ability to create categories which can be stored and used as filters to gather metrics describing the degree to which that text adheres to a particular category; and in 8 different languages!

Iris internally uses the “Compare feature” of the same REST API available to its users. Because Iris is written using JavaFX, it may be run on Mac OS, Linux or Windows. The Mac OS installer has the correct version of Java (Java 8) pre-bundled so no Java installation is required. Linux and Windows users will have to download Java 8 to run it however.

We are still working on usage instructions for the new Classification feature, but for now here’s basic instructions for trying out the new feature:

  1. Iris uses Input Windows to offer ways to input text and expression data; and Output Windows to manipulate the display of results and specify variations in output attributes. This is so multiple inputs may be reused along with many outputs offering a variety of ways to view results. (Some features such as the “Compare” feature, require more than 1 input).
  2. In the Output Window click on the “Classify” tab; then in the table click on “Add / Edit New Filters”; or in the main menu, click on “Edit Classify Filters…”.
  3. Click on “New”. To create a Category, at least one body of text representing a “positive” example must be used (but many may be used to refine the requirements for a given category). Also, “negative” example text may be specified.
  4. Give the Category a name and press Enter.
  5. Click the “+” button below in the bar entitled “Positive Examples”; and paste or type in some text which typifies the semantic context of the category you are creating.
  6. Generate a Fingerprint (SDR) for the category by pressing the “Generate icon” next to the category name, which looks like a lightning bolt.
  7. When you are finished entering examples, click on Save; then close the Category Composer and return to the Output Window. Your new Category will be available for selection.
  8. In addition a Threshold may be specified to refine how closely text must adhere to the category to pass.
  9. When the filter matches the category, that row will be turned green.

Now when you apply input using the Input Window, any selected Categories will be automatically run in your Output Window…

For general usage information, the Iris project landing page has links to both documentation and an introductory video. In addition, I (@cogmission) may be contacted here in the HTM Forum for any extra tips or help you may need… (Please no personal messages (they will not be answered), let’s create forum topics instead, so that other users may benefit from your inquiries :slight_smile: )