# Decoding in HTM

**URL:** <https://discourse.numenta.org/t/decoding-in-htm/9173>\
**Category:** HTM.Java\
**Tags:** sdrs\
**Created:** [November 25, 2021, 7:59am UTC](https://discourse.numenta.org/t/decoding-in-htm/9173 "2021-11-25T07:59:57Z")\
**Posts on this page:** 1\
**Showing post:** 2

<div class="post-metadata">

**Author:** ![DanML](https://avatars.discourse-cdn.com/v4/letter/d/edb3f5/32.png) [@DanML](https://discourse.numenta.org/u/DanML)\
**Post date:** [December 1, 2021, 9:51pm UTC](https://discourse.numenta.org/t/decoding-in-htm/9173/2 "2021-12-01T21:51:02Z")

</div>

Hi, this thread may help you:

> [@Biological way to decode SDR back to Objects](https://discourse.numenta.org/t/biological-way-to-decode-sdr-back-to-objects/5546):
>
> In [https://github.com/htm-community/nupic.cpp/issues/297](https://github.com/htm-community/nupic.cpp/issues/297) I’m discussing how to properly decode (convert from SDR back to original Object) and what would be the biologically plausible way. Decoding, inverse of compute(), aka topdown compute This idea is about how we should properly obtain original value from SDR, how to “decode”. Methods There are 2 (3) methods doing decoding: a Classifier is trained on pair (SDR, value) and later can infer SDR-\> value. (in the past) there used to be top-down…

AFAIK current implementations use the Classifier (NN) to find the mapping. I assume the Java version is the same.  
Rhyolight’s point is also worth thinking about.

hth

---

_[View the full topic](https://discourse.numenta.org/t/decoding-in-htm/9173)._