Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

SP return const random representation on empty input & SP anomaly #550

Open
3 tasks
breznak opened this issue Jul 9, 2019 · 1 comment
Open
3 tasks
Assignees
Labels
anomaly research new functionality of HTM theory, research idea SP

Comments

@breznak
Copy link
Member

breznak commented Jul 9, 2019

Problem: empty SP output on low overlap

Currently, when the overlap with input (unexpected input, unlearned, ..?) is too low, SP returns 0 active columns.
This causes problems in TM, and Anomaly

Proposal

I want to change SP such that

  • aways returns numDesired (keeping sparsity) outputs, see SP: remove numActiveColumnsPerInhArea #549 , and this issue for cases with low overlap.
  • if overlap is too low:
    • detect spatial anomaly!
      • implement the same way is TM.anomaly, used in NAB for example.
    • return a const, random SDR that designates "none, or unrecognized input"
    • may call bumpUpWeakColumns to have some imput.

This should be tested on anomaly datasets, so ideally once #205

@breznak breznak self-assigned this Jul 9, 2019
@breznak breznak changed the title SP return const random representation on empty input SP return const random representation on empty input & SP anomaly Jul 24, 2019
@breznak breznak added the research new functionality of HTM theory, research idea label Jun 2, 2020
@breznak
Copy link
Member Author

breznak commented Jun 2, 2020

Related #463

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
anomaly research new functionality of HTM theory, research idea SP
Projects
None yet
Development

No branches or pull requests

1 participant