Holographic word forms

 

Cox, G., Kachergis, G., Recchia, G., & Jones, M. N. (2011). Towards a scalable holographic word-form representation. Behavior Research Methods, 43, 602-615.

Holoword Python Code:

Python code may be integrated into cognitive models to input actual word strings, and to produce both similarity metrics and distributed representations that contain the similarity structure to train your model on. [holoword.py]

The code supports various holographic encoding operators (circular convolution, noisy addition, binary spatter coding, and random permutation), ngram extraction schemes (open or terminal-relative encoding), chunk sizes, and dimensionality.


Holoword GUI:

For those who are more likely to use our system for selecting stimuli and are not interested in the code, we have made a simple GUI version. You may input word strings--output is the similarity matrix from the holographic encoding of those strings, and you also have the option to store the actual distributed representation vectors themselves to feed into a cognitive model. 

Currently, the GUI is for Windows only: Download [holoword_GUI.zip]

To install, unzip the zip file in its entirety and run holoword.exe. You will need a list of words to train on. A sample list, tasa.lemmas.mostfrequent.txt, is included in the zip.

If you are running a version of Windows that is older than Vista, you may receive an error message when you try to run the program. To resolve this, you may need to first download and install the .NET Redistributables from Microsoft Update.


Return to the CCL Code Repository