NeuroMem Knowledge Builder

To experience with the power and adaptivity of a NeuroMem neural network, General Vision has developed NeuroMem Knowledge Builder (NMKB), a simple framework running under Windows for training the neurons on your datasets, validating their accuracy on new datasets and generating detailed reports about throughput and accuracy.

NMKB is delivered with a cycle accurate simulation of 8000 neurons, but can also interface with a NeuroMem USB dongle with a fixed capacity of 2304 neurons.

NMKB import labeled datasets in txt or csv format including feature vectors and their categories. The vectors can derive from any data types like text, heterogeneous measurements, images, video and audio files.

The neurons can build a knowledge in a few seconds plotting their learning curve to monitor convergence. Diagnostics reports help determine if the training was appropriate sufficient, predict which categories are easy or difficult to discriminate, and more.

For the classification of new datasets, you can choose to use the neurons as a Radial Basis Function or K-Nearest Neighbor classifier. Other settings include the value K and a consolidation rule to produce a single output in case of uncertainties. The throughput and accuracy of the classification are detailed per categories. Also, the traceability of the knowledge built by the NeuroMem neurons is conveniently exploited by NMKB. For example, you can filter input vectors classified incorrectly and comprehend why by comparing their profile to the firing models. This utility may even pinpoint errors in the input datasets!

The application produces a data log to easily track and compare the settings and work flow which were tested on a given dataset.

The knowledge built by the neurons can be saved and loaded as a primitive knowledge on execution platforms which can themselves complement and enrich this knowledge if they are configured with a learning logic.