Alexander Lavin

Results 14 issues of Alexander Lavin

How could one set the wallpaper to be [https://vincentgarreau.com/particles.js/](particle.js) background instead of a static image?

The algorithm runs online, but with the assumption we have the length of the dataset a priori. What about streaming scenarios where we have a continuous stream of data? Is...

enhancement

A good functionality for [NuPIC Random](https://github.com/numenta/nupic/blob/master/nupic/bindings/math.i) would be to allow for seeding with a string. This can be done by hashing the string to an integer seed with hashlib.

newbie

The delta encoder should not subclass from and use the adaptive scalar encoder -- [here](https://github.com/numenta/nupic/blob/master/src/nupic/encoders/delta.py#L29). The adaptive scalar encoder can give poor results particularly for short data streams because the...

type:bug
type:docs
subject:encoders
priority:3
triage

The simple labels and hello classification tests for sensor_knn.json config fail -- 52.98% vs 66.2%, and 60% vs 80%, respectively. @subutai would you please confirm you get the same?

With the param `generateData` set the HTMRunner will create `self.dataDict` as it reads in the CSV and writes CSVs formatted for the network API. If you pass an already formatted...

bug
NLP

When creating a network with a TP region in the sequence classification factory, you can only specify the TP params and not the underlying SP params. That is, the TP...

NLP
Sequence Classification

Instead of using dynamic imports when registering regions ([here](https://github.com/numenta/nupic.research/blob/master/htmresearch/frameworks/classification/classification_network.py#L163)) the user should pass in the region to the factory. That is, whatever runner is creating the network instance should be...

Sequence Classification

Similar to [fluent #137](https://github.com/numenta/nupic.fluent/pull/137), extend the use of enums for classifier types (`"CLA"` vs. `"KNN"`), model types, etc.

NLP

The confusion matrix (or now matrices) should reflect errors where actual and predicted labels are e.g. [2, 3, 7] and [3, 4], respectively; this example highlights mismatches between vector size...

NLP