# TM’s ability to generalize

**URL:** https://discourse.numenta.org/t/tm-s-ability-to-generalize/3630
**Category:** Numenta Theory
**Created:** [March 8, 2018, 7:48pm UTC](https://discourse.numenta.org/t/tm-s-ability-to-generalize/3630 "2018-03-08T19:48:08Z")
**Posts on this page:** 1
**Showing post:** 5

<div class="post-metadata">

### Author: ![sunguralikaan](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.numenta.org/sunguralikaan/32/585_2.png) [@sunguralikaan](https://discourse.numenta.org/u/sunguralikaan)
#### Post date: [March 12, 2018, 5:49pm UTC](https://discourse.numenta.org/t/tm-s-ability-to-generalize/3630/5 "2018-03-12T17:49:18Z")

</div>

> [@Why is HTM Ignored by Google DeepMind?](https://discourse.numenta.org/t/why-is-htm-ignored-by-google-deepmind/2589/41):
>
> The SP handles semantically similar static input states but not semantically similar temporal contexts

I think I understand what you mean but I do not think temporal similarity is possible without somehow reducing or sampling multiple transitions into a single activation. TM only samples from the previous activation so it cannot create that _semantically similar temporal context_. It can only represent a transition. I would speculate that Temporal Pooling is what you want.

---

_[View the full topic](https://discourse.numenta.org/t/tm-s-ability-to-generalize/3630)._