# Need new ideas for a new HTM project

**URL:** <https://discourse.numenta.org/t/need-new-ideas-for-a-new-htm-project/4951>\
**Category:** Lounge\
**Created:** [November 14, 2018, 7:23am UTC](https://discourse.numenta.org/t/need-new-ideas-for-a-new-htm-project/4951 "2018-11-14T07:23:53Z")\
**Posts on this page:** 13\
**Page:** 1

<div class="post-metadata">

**Author:** ![marty1885](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.numenta.org/marty1885/32/1891_2.png) [@marty1885](https://discourse.numenta.org/u/marty1885)\
**Post date:** [November 14, 2018, 7:23am UTC](https://discourse.numenta.org/t/need-new-ideas-for-a-new-htm-project/4951/1 "2018-11-14T07:23:53Z")

</div>

Hi all.  
I have been experimenting and applying HTM to real problems recently. And now I want to use HTM in my final project for my machine learning class in my collage. Which could be anything regarding ML. Well… I’m honestly out of new ideas. I can’t think off anything besides anomaly detection, sequence prediction and data classification (by using SP and SDRClassifer).

Do anyone has any idea? Any idea I can borrow?

* * *

My previous projects  
[HTM Playing Rock, Paper, Scissor against RNN](https://discourse.numenta.org/t/htm-vs-gru-playing-rock-paper-scissors-htm-outperforming-by-7)  
[ECG Heartbeat Categorization as anomaly detection using HTM](https://discourse.numenta.org/t/learning-from-real-world-data-ecg-heartbeat-categorization-as-anomaly-detection-using-htm/)  
[Generating Esperanto-ish gibberish using HTM](https://discourse.numenta.org/t/esperanto-nlp-using-htm-and-my-finding)

---

<div class="post-metadata">

**Author:** ![Jose\_Cueto](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.numenta.org/jose_cueto/32/4052_2.png) [@Jose\_Cueto](https://discourse.numenta.org/u/Jose_Cueto)\
**Post date:** [November 14, 2018, 12:18pm UTC](https://discourse.numenta.org/t/need-new-ideas-for-a-new-htm-project/4951/2 "2018-11-14T12:18:49Z")

</div>

How about classifying patterns in an EEG data stream? If you have one of those toy EEG devices of course. It generates lots of data in real-time.

---

<div class="post-metadata">

**Author:** ![sebjwallace](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.numenta.org/sebjwallace/32/6411_2.png) [@sebjwallace](https://discourse.numenta.org/u/sebjwallace)\
**Post date:** [November 14, 2018, 8:23pm UTC](https://discourse.numenta.org/t/need-new-ideas-for-a-new-htm-project/4951/3 "2018-11-14T20:23:20Z")

</div>

Stock market? Classic.

---

<div class="post-metadata">

**Author:** ![Tachion](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.numenta.org/tachion/32/3257_2.png) [@Tachion](https://discourse.numenta.org/u/Tachion)\
**Post date:** [November 14, 2018, 8:57pm UTC](https://discourse.numenta.org/t/need-new-ideas-for-a-new-htm-project/4951/4 "2018-11-14T20:57:31Z")

</div>

Writing a proper audio encoder for HTM, the one that mat did way back on youtube lacked a feature that is present in the cochlea and that would be the “heat map of intensity” of a particular tone.

let me give you a clue of how this might look in a list form:  
20hz range: [0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0] - low intensity  
20hz range: [0, 0, 1, 1, 1, 1, 1, 1, 1, 0, 0] - high intensity

this form of representation is superior to the classic dial thats sort of similar to a dial you would find on a music mix set found on the scalar encoder

This saves the semantic of the tone while embedding the intensity of the tone that can be later used as a location compass in other projects. Unfortunately you would need to use some Fourier transforms for this kind of encoding. I imagine the human (or any other animals) ear uses an aproximator of that kind to represent soundbytes

Also it would be interesting to see results done both using and not using topology

---

<div class="post-metadata">

**Author:** ![thanh-binh.to](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.numenta.org/thanh-binh.to/32/1793_2.png) [@thanh-binh.to](https://discourse.numenta.org/u/thanh-binh.to)\
**Post date:** [November 14, 2018, 9:51pm UTC](https://discourse.numenta.org/t/need-new-ideas-for-a-new-htm-project/4951/5 "2018-11-14T21:51:21Z")

</div>

You can use frequency encoder as Some Numenta guys did last year

---

<div class="post-metadata">

**Author:** ![Tachion](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.numenta.org/tachion/32/3257_2.png) [@Tachion](https://discourse.numenta.org/u/Tachion)\
**Post date:** [November 14, 2018, 10:26pm UTC](https://discourse.numenta.org/t/need-new-ideas-for-a-new-htm-project/4951/6 "2018-11-14T22:26:57Z")

</div>

yes, completely forgot about it

Need to address that my addition also adds intensity, think of it as “volume” and it gives HTM additional attack vectors, one of them being orientation on which it will build new patterns while still remaining on that frequency region

I think just changing the encoding from what they already have built to what i’ve added is basically 10-15 min work

---

<div class="post-metadata">

**Author:** ![Bitking](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.numenta.org/bitking/32/3099_2.png) [@Bitking](https://discourse.numenta.org/u/Bitking)\
**Post date:** [November 14, 2018, 10:36pm UTC](https://discourse.numenta.org/t/need-new-ideas-for-a-new-htm-project/4951/7 "2018-11-14T22:36:11Z")

</div>

How about something fun like a turtle robot brain or a lunar lander pilot?

> **[Turtle (robot)](https://en.wikipedia.org/wiki/Turtle_(robot))**
>
> Turtles are a class of educational robots designed originally in the late 1940s (largely under the auspices of researcher William Grey Walter) and used in computer science and mechanical engineering training. These devices are traditionally built low to the ground with a roughly hemispheric (sometimes transparent) shell and a power train capable of a very small turning radius. The robots are often equipped with sensor devices which aid in avoiding obstacles and, if the robot is sufficiently sophi...

> **[Lunar Lander (1979 video game)](https://en.wikipedia.org/wiki/Lunar_Lander_(1979_video_game))**
>
> Lunar Lander is a single-player arcade game in the Lunar Lander subgenre. It was developed by Atari, Inc. and released in August 1979. In the game, the player controls a lunar landing module as viewed from the side and attempts to land safely on the Moon. The player burns fuel points to rotate the module and fires a thruster to counteract gravity, earning points based on the skillfulness and difficulty of the landing. The game resets the module after every successful landing or crash with a new...

Or pong?

---

<div class="post-metadata">

**Author:** ![TotalMegaCool](https://avatars.discourse-cdn.com/v4/letter/t/d9b06d/32.png) [@TotalMegaCool](https://discourse.numenta.org/u/TotalMegaCool)\
**Post date:** [November 14, 2018, 11:34pm UTC](https://discourse.numenta.org/t/need-new-ideas-for-a-new-htm-project/4951/8 "2018-11-14T23:34:48Z")

</div>

One of the things I wanted to do was to try and get the computer to build a model of the world and make predictions about what will happen. I was thinking of having a simple 2D world with a ball that falls and reacts with an environment.

Step01:

 ![Step01](https://canada1.discourse-cdn.com/flex030/uploads/numenta/original/2X/7/7ed3d8caef3887c0a2a1a9f545b5f964da6f4370.png)  
Step02:  
 ![Step02](https://canada1.discourse-cdn.com/flex030/uploads/numenta/original/2X/d/dd83a34c976cc1e6c6ff0858dff3a8d98f85549d.png)  
Step03:  
 ![Step03](https://canada1.discourse-cdn.com/flex030/uploads/numenta/original/2X/2/20501e4cec8c508eb1b873c8b05ae86a86d843a5.png)

I don’t really know if it would work or what the results would be but I was thinking of once the agent had learned a model of the world I could use it to then drive the simulation and get the ball to fall and react in a realistic way using the HTM prediction as the next frame of the simulation and recursively repeating the prediction.

---

<div class="post-metadata">

**Author:** ![marty1885](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.numenta.org/marty1885/32/1891_2.png) [@marty1885](https://discourse.numenta.org/u/marty1885)\
**Post date:** [November 15, 2018, 1:16am UTC](https://discourse.numenta.org/t/need-new-ideas-for-a-new-htm-project/4951/9 "2018-11-15T01:16:28Z")

</div>

@Bitking The ideas you proposed are awesome. But I don’t think that we have sensor-motor inference working at the moment? Am I missing something?

---

<div class="post-metadata">

**Author:** ![Bitking](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.numenta.org/bitking/32/3099_2.png) [@Bitking](https://discourse.numenta.org/u/Bitking)\
**Post date:** [November 15, 2018, 1:19am UTC](https://discourse.numenta.org/t/need-new-ideas-for-a-new-htm-project/4951/10 "2018-11-15T01:19:45Z")

</div>

Sensor - yes. We call them encoders.  
Working up a good motor output may be breaking new ground.  
The joy of being a researcher!

---

<div class="post-metadata">

**Author:** ![Henry\_Minsky](https://avatars.discourse-cdn.com/v4/letter/h/51bf81/32.png) [@Henry\_Minsky](https://discourse.numenta.org/u/Henry_Minsky)\
**Post date:** [November 15, 2018, 1:49am UTC](https://discourse.numenta.org/t/need-new-ideas-for-a-new-htm-project/4951/11 "2018-11-15T01:49:01Z")

</div>

There is an OpenAI “gym” for some simple video games, like Super Mario

The API lets you run the video game emulator, and sends you a simplified ‘tile’ representation of the game,  
which might be easy to feed to an HTM.

I’d like to see something which predicts what will happen if one of the ‘goombas’ is seen, or if you move the player towards one of them, predicts that your player will die.

The sensor data is available as a 2D array of ‘tiles’, each one represents what object is at a location on a maybe 40 x 20 grid or something.

I don’t know the best way to encode that into HTM sparse data inputs, but it would be nice to separate it out into X and Y velocity of objects or something which reasonably closely matches what a real visual system of an animal might be putting out.

---

<div class="post-metadata">

**Author:** ![Tachion](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.numenta.org/tachion/32/3257_2.png) [@Tachion](https://discourse.numenta.org/u/Tachion)\
**Post date:** [November 15, 2018, 6:38am UTC](https://discourse.numenta.org/t/need-new-ideas-for-a-new-htm-project/4951/12 "2018-11-15T06:38:45Z")

</div>

I had a few ideas regarding how a primitive sensory-motor inference could be applied without any neuroscience to back it up  
In practice its going to look very shoehorned in

But my problem is representing displacement cells, any ideas how could that be done?

---

<div class="post-metadata">

**Author:** ![way-sal](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.numenta.org/way-sal/32/3028_2.png) [@way-sal](https://discourse.numenta.org/u/way-sal)\
**Post date:** [November 15, 2018, 1:28pm UTC](https://discourse.numenta.org/t/need-new-ideas-for-a-new-htm-project/4951/13 "2018-11-15T13:28:21Z")

</div>

what about HTM + automotive transport system