I’ve created various models where after the model has run for some time the classifier confidence value starts returning NaN. As a sanity check, I’ve gone back to the standard hotgym example ( http://nupic.docs.numenta.org/stable/quick-start/network.html ) and found that it too suffers from this prob…

This is probably a bug, so I created a bug report . I wonder if the SDR Classifier in the community fork has the same problem? cc @David_Keeney @breznak @dmac

[image] PhilGoddard: Nothing else is being changed from the default hotgym code/model. Specifically, at about the 19000th data point, the 5-step ahead confidence starts returning NaN. Can you try to replicate this and give us a test? Problem is you’re probably using python OPF framework, whi…

We’re using the Network API version (as per the link in my original posting) not the OPF version. The only thing changed is that a longer data set is passed through the model. I’ve attached a code package with example data to the bug report.

[image] PhilGoddard: We’re using the Network API version (as per the link in my original posting) not the OPF version. Perfect, that means we can run it easily, we should be 100% NetworkAPI compatible! Would you be able to create a PR to Pull requests · htm-community/htm.core · GitHub ? If n…

yep, I think community repo should be ready to give your example a shot [image] Issue: Implement NetworkAPI example opened by breznak on 2019-05-08 A user is having a problem with SDRClassifier on numenta's, we'd like to replicate the provided experiment (runnable source …

@PhilGoddard Assuming this is the code we’re talking about, can you get access to some of these values during a compute cycle where predictionConfidence is NaN? Does classifierRegion.getOutputData("probabilities") or classifierRegion.getOutputData("actualValues") look any different in structure? d…

Yes, that’s the function being used. Not sure what you mean by “different in structure” – they are the expected size and data type. Superficially the problem is that probabilities has at least one NaN value in it, and argmax is returning the index of that NaN, so predictedConfidence is NaN. An in…

[image] PhilGoddard: but I am replicating/extending the data to have 30000 time steps QQ: are you still sending in timestamps in the extended data? Are they duplicates of existing data the model has seen? Or do they continue forward in time logically?

They continue forward in time logically. See the example input data file included in the code package attached to the bug report.

Phil, I tried debugging this today for awhile (video coming soon). It looks like a bug in the C++ SDRClassifier. If you change your implementation to py, does it work? clParams: implementation: py

SDRClassifier returning confidence of NaN

Engineering

pulinagrawal May 10, 2019, 12:33am 21

I cannot be sure yet because I have not been able to give enough time to cpp implementation, but this could possibly be related to the issue. If some one gets to debugging this in cpp, this may be something to look at.

Topic		Replies	Views
Confused with SDR classifier and very small alpha values NuPIC question , classification	4	864	July 31, 2016
SDR Classifier for spatial pooler label Engineering	12	1036	August 16, 2020
CLA / SDR Classifier Bug in nupic.core NuPIC	11	1354	October 26, 2016
Asymmetric confidence when predicting a symmetric signal Numenta Theory	41	2603	May 10, 2019
Implementing the Hot Gym Example NuPIC	3	655	April 15, 2020

SDRClassifier returning confidence of NaN

Related topics