# Latest

**URL:** https://discourse.numenta.org/latest.md?no_subcategories=false&page=2

[Latest](https://discourse.numenta.org/latest.md) · [Categories](https://discourse.numenta.org/categories.md) · [Tags](https://discourse.numenta.org/tags.md)

**Page:** 3

---

## [Tried feeding Grok the forum with a question](https://discourse.numenta.org/t/tried-feeding-grok-the-forum-with-a-question/12113)

<div class="topic-metadata">

**Author:** [@BrainVx](https://discourse.numenta.org/u/BrainVx)\
**Replies:** 2\
**Last updated:** [November 20, 2025, 12:00am UTC](https://discourse.numenta.org/t/tried-feeding-grok-the-forum-with-a-question/12113 "2025-11-20T00:00:33Z")

</div>

Thought this might be of interest and stoke some ideas. Having used Grok for a number of different areas (legal, code, physics, conceptual, etc.) I thought I would see what Grok made of the site and the posts in relatio…

---

## [Free ML, Web Development, Data, Computer Science Books for All](https://discourse.numenta.org/t/free-ml-web-development-data-computer-science-books-for-all/8694)

<div class="topic-metadata">

**Author:** [@Bitking](https://discourse.numenta.org/u/Bitking)\
**Replies:** 4\
**Last updated:** [October 14, 2025, 12:28pm UTC](https://discourse.numenta.org/t/free-ml-web-development-data-computer-science-books-for-all/8694 "2025-10-14T12:28:43Z")

</div>

Because education should be free…( download today!!) Who does not like great free books?

---

## [Does predictive coding have a future?](https://discourse.numenta.org/t/does-predictive-coding-have-a-future/4364)

<div class="topic-metadata">

**Author:** [@Bitking](https://discourse.numenta.org/u/Bitking)\
**Replies:** 2\
**Last updated:** [September 29, 2025, 2:47pm UTC](https://discourse.numenta.org/t/does-predictive-coding-have-a-future/4364 "2025-09-29T14:47:08Z")

</div>

Clearly - in this forum, we take the answer in the affirmative! I think this Paper speaks directly to the theoretical underpinnings of HTM theory. Fristen talks of the difference between predictions and perceptions (I …

---

## [Hierarchical Models in the Brain](https://discourse.numenta.org/t/hierarchical-models-in-the-brain/12039)

<div class="topic-metadata">

**Author:** [@CollinsEM](https://discourse.numenta.org/u/CollinsEM)\
**Replies:** 2\
**Last updated:** [September 29, 2025, 11:30am UTC](https://discourse.numenta.org/t/hierarchical-models-in-the-brain/12039 "2025-09-29T11:30:37Z")

</div>

Hierarchical Models in the Brain Karl Friston Author Summary Models are essential to make sense of scientific data, but they may also play a central role in how we assimilate sensory information. In this paper, we intr…

---

## [Could Someone Give me Advice on Implementing HTM for Real-Time Anomaly Detection in Industrial IoT Systems?](https://discourse.numenta.org/t/could-someone-give-me-advice-on-implementing-htm-for-real-time-anomaly-detection-in-industrial-iot-systems/11491)

<div class="topic-metadata">

**Author:** [@roberrttt](https://discourse.numenta.org/u/roberrttt)\
**Replies:** 2\
**Last updated:** [September 13, 2025, 11:51pm UTC](https://discourse.numenta.org/t/could-someone-give-me-advice-on-implementing-htm-for-real-time-anomaly-detection-in-industrial-iot-systems/11491 "2025-09-13T23:51:24Z")

</div>

Hello there, I am currently working on a project that involves implementing Hierarchical Temporal Memory for real time anomaly detection in an Industrial Internet of Things system. The goal is to monitor sensor data fr…

---

## [SpikingBrain Technical Report: Spiking Brain-inspired Large Models](https://discourse.numenta.org/t/spikingbrain-technical-report-spiking-brain-inspired-large-models/12095)

<div class="topic-metadata">

**Author:** [@MTIzNDU2Nzg5](https://discourse.numenta.org/u/MTIzNDU2Nzg5)\
**Replies:** 2\
**Last updated:** [September 12, 2025, 4:22pm UTC](https://discourse.numenta.org/t/spikingbrain-technical-report-spiking-brain-inspired-large-models/12095 "2025-09-12T16:22:51Z")

</div>

:brain: SpikingBrain: brain-inspired principles in large language models Researchers from China have presented a technical report on SpikingBrain - a series of large language models inspired by how the brain works. Th…

---

## [A Frozen Neural Network](https://discourse.numenta.org/t/a-frozen-neural-network/12066)

<div class="topic-metadata">

**Author:** [@SeanOConnor](https://discourse.numenta.org/u/SeanOConnor)\
**Replies:** 4\
**Last updated:** [September 7, 2025, 3:10pm UTC](https://discourse.numenta.org/t/a-frozen-neural-network/12066 "2025-09-07T15:10:10Z")

</div>

I presented a type of frozen neural network a long time ago. I just re-blogged about it: https://ko-fi.com/post/Frozen-Neural-Networks-Fixed-Filter-Bank-Architec-Z8Z31KRNGP Anyway I asked chatGPT5 its thoughts about it…

---

## [Using a Classifier with htm.core for Skeleton-Based Human Activity Recognition](https://discourse.numenta.org/t/using-a-classifier-with-htm-core-for-skeleton-based-human-activity-recognition/11889)

<div class="topic-metadata">

**Author:** [@muhammadzulqarnain58](https://discourse.numenta.org/u/muhammadzulqarnain58)\
**Replies:** 2\
**Last updated:** [September 5, 2025, 4:48am UTC](https://discourse.numenta.org/t/using-a-classifier-with-htm-core-for-skeleton-based-human-activity-recognition/11889 "2025-09-05T04:48:01Z")

</div>

Hello Everyone, I am working on a research project involving Human Activity Recognition using the NTU RGB+D 3D skeleton dataset, and I am currently building a pipeline with htm.core to process the skeletal data. So far,…

---

## [Weight Switching in Neural Networks using Locality Sensitive Hashing](https://discourse.numenta.org/t/weight-switching-in-neural-networks-using-locality-sensitive-hashing/12067)

<div class="topic-metadata">

**Author:** [@SeanOConnor](https://discourse.numenta.org/u/SeanOConnor)\
**Replies:** 2\
**Last updated:** [September 5, 2025, 11:58pm UTC](https://discourse.numenta.org/t/weight-switching-in-neural-networks-using-locality-sensitive-hashing/12067 "2025-09-05T23:58:41Z")

</div>

You have a very wide range of choices if you use locality sensitive hashing to select weights in a neural network. Not only various select from a pool strategies, also various switching scopes (eg. input weights to a neu…

---

## [Local Hebb in apical dendrites](https://discourse.numenta.org/t/local-hebb-in-apical-dendrites/11819)

<div class="topic-metadata">

**Author:** [@bkaz](https://discourse.numenta.org/u/bkaz)\
**Replies:** 8\
**Last updated:** [September 3, 2025, 5:01am UTC](https://discourse.numenta.org/t/local-hebb-in-apical-dendrites/11819 "2025-09-03T05:01:31Z")

</div>

Distinct synaptic plasticity rules operate across dendritic compartments in vivo during learninghttps://pubmed.ncbi.nlm.nih.gov/40245144/ "we found that apical and basal dendrites of layer 2/3 (L2/3) pyramidal neurons s…

---

## [Learning in High Dimension Always Amounts to Extrapolation](https://discourse.numenta.org/t/learning-in-high-dimension-always-amounts-to-extrapolation/12012)

<div class="topic-metadata">

**Author:** [@SeanOConnor](https://discourse.numenta.org/u/SeanOConnor)\
**Replies:** 12\
**Last updated:** [August 27, 2025, 1:09am UTC](https://discourse.numenta.org/t/learning-in-high-dimension-always-amounts-to-extrapolation/12012 "2025-08-27T01:09:35Z")

</div>

Here is a paper from a while ago: https://arxiv.org/abs/2110.09485 I was interacting with chatGPT about it and I pointed out that any function more complex than the simple weighted sum will become extremely local in hi…

---

## [How can HTM be used in Reinforcement Learning projects?](https://discourse.numenta.org/t/how-can-htm-be-used-in-reinforcement-learning-projects/12051)

<div class="topic-metadata">

**Author:** [@kayoj](https://discourse.numenta.org/u/kayoj)\
**Replies:** 1\
**Last updated:** [August 24, 2025, 1:53am UTC](https://discourse.numenta.org/t/how-can-htm-be-used-in-reinforcement-learning-projects/12051 "2025-08-24T01:53:05Z")

</div>

Hello I have been learning about reinforcement learning & started to wonder if Hierarchical Temporal Memory (HTM) ideas can fit into this field.:slightly\_smiling\_face: Most RL projects today use deep neural networks bu…

---

## [Winner Takes All and the Weighed Sum](https://discourse.numenta.org/t/winner-takes-all-and-the-weighed-sum/12037)

<div class="topic-metadata">

**Author:** [@SeanOConnor](https://discourse.numenta.org/u/SeanOConnor)\
**Replies:** 13\
**Last updated:** [August 22, 2025, 1:28pm UTC](https://discourse.numenta.org/t/winner-takes-all-and-the-weighed-sum/12037 "2025-08-22T13:28:36Z")

</div>

The weighted sum is used everywhere in ML though people rarely consider it as vector to scalar associative memory. Yes, its cheap and simple \<vector,scalar\> associative memory. However it has a number of non-ideal proper…

---

## [h&sELM (Hash & Select Extreme Learning Machines)](https://discourse.numenta.org/t/h-selm-hash-select-extreme-learning-machines/12040)

<div class="topic-metadata">

**Author:** [@SeanOConnor](https://discourse.numenta.org/u/SeanOConnor)\
**Replies:** 3\
**Last updated:** [August 22, 2025, 12:13pm UTC](https://discourse.numenta.org/t/h-selm-hash-select-extreme-learning-machines/12040 "2025-08-22T12:13:13Z")

</div>

This is where I am at with Extreme Learning Machines: https://sciencelimelight.blogspot.com/2025/08/hash-select-extreme-learning-machine.html It seems to be the best overall mix of things to do at the moment. I kinda …

---

## [A Hierarchy of Time-Scales and the Brain](https://discourse.numenta.org/t/a-hierarchy-of-time-scales-and-the-brain/12038)

<div class="topic-metadata">

**Author:** [@CollinsEM](https://discourse.numenta.org/u/CollinsEM)\
**Replies:** 2\
**Last updated:** [August 18, 2025, 9:48pm UTC](https://discourse.numenta.org/t/a-hierarchy-of-time-scales-and-the-brain/12038 "2025-08-18T21:48:19Z")

</div>

A Hierarchy of Time-Scales and the Brain Stefan J. Kiebel, Jean Daunizeau, Karl J. Friston Author Summary Currently, there is no theory that explains how the largescale organization of the human brain can be related to…

---

## [Cortical columns and knowledge representation, and shows why LLMs can never do this](https://discourse.numenta.org/t/cortical-columns-and-knowledge-representation-and-shows-why-llms-can-never-do-this/11962)

<div class="topic-metadata">

**Author:** [@Peter2](https://discourse.numenta.org/u/Peter2)\
**Replies:** 17\
**Last updated:** [August 15, 2025, 10:49pm UTC](https://discourse.numenta.org/t/cortical-columns-and-knowledge-representation-and-shows-why-llms-can-never-do-this/11962 "2025-08-15T22:49:02Z")

</div>

Check out what this old guy has to say in line with HTM.

---

## [Ensemble Neural Networks with Random Projections](https://discourse.numenta.org/t/ensemble-neural-networks-with-random-projections/12035)

<div class="topic-metadata">

**Author:** [@SeanOConnor](https://discourse.numenta.org/u/SeanOConnor)\
**Replies:** 1\
**Last updated:** [August 13, 2025, 11:30pm UTC](https://discourse.numenta.org/t/ensemble-neural-networks-with-random-projections/12035 "2025-08-13T23:30:04Z")

</div>

You could just ensemble a number of neural networks by averaging across dimensions. That is to miss some opportunities for neurons to help each other out, so to say. Or you can take a weak learner view point and then f…

---

## [Eigenspace Splitting for fun](https://discourse.numenta.org/t/eigenspace-splitting-for-fun/12024)

<div class="topic-metadata">

**Author:** [@SeanOConnor](https://discourse.numenta.org/u/SeanOConnor)\
**Replies:** 3\
**Last updated:** [August 11, 2025, 9:44am UTC](https://discourse.numenta.org/t/eigenspace-splitting-for-fun/12024 "2025-08-11T09:44:09Z")

</div>

I constructed these 2 related blogs: https://sciencelimelight.blogspot.com/2025/08/double-up-gaussian-random-numbers.html Where you can double up a vector into 2 vectors. I think the main application in machine lear…

---

## [Mental framing problems and ReLU as a switch](https://discourse.numenta.org/t/mental-framing-problems-and-relu-as-a-switch/12019)

<div class="topic-metadata">

**Author:** [@SeanOConnor](https://discourse.numenta.org/u/SeanOConnor)\
**Replies:** 2\
**Last updated:** [August 12, 2025, 12:38am UTC](https://discourse.numenta.org/t/mental-framing-problems-and-relu-as-a-switch/12019 "2025-08-12T00:38:45Z")

</div>

Blog - Mental framing problems and ReLU as a switch https://sciencelimelight.blogspot.com/2025/08/mental-framing-problems-viewing-relu-as.html I think the phrase “Missed Serendipity” about some of the early ReLU papers…

---

## [32.63% on ARC AGI 2 benchmark](https://discourse.numenta.org/t/32-63-on-arc-agi-2-benchmark/12011)

<div class="topic-metadata">

**Author:** [@Tofara\_Moyo](https://discourse.numenta.org/u/Tofara_Moyo)\
**Replies:** 0\
**Last updated:** [August 7, 2025, 3:51pm UTC](https://discourse.numenta.org/t/32-63-on-arc-agi-2-benchmark/12011 "2025-08-07T15:51:53Z")

</div>

Continuous-Time Graph Neural Networks for Abstraction and Reasoning Tasks author:Tofara Moyo I achieved 32.63% on Arc AGI2 dataset, state of the art with a model with 50k parameters. code here : https://colab.researc…

---

## [Help to understand and replicate the HOTGYM algorithm](https://discourse.numenta.org/t/help-to-understand-and-replicate-the-hotgym-algorithm/11942)

<div class="topic-metadata">

**Author:** [@MSc.Lic\_P\_Raquel\_Fru](https://discourse.numenta.org/u/MSc.Lic_P_Raquel_Fru)\
**Replies:** 3\
**Last updated:** [August 7, 2025, 5:19pm UTC](https://discourse.numenta.org/t/help-to-understand-and-replicate-the-hotgym-algorithm/11942 "2025-08-07T17:19:14Z")

</div>

Hello to all the members of the forum.To the authors! Congratulations for your publication, which is really very interesting, I am studying a PhD in Computer Science, in Paraguay, and I am very interested in the articles…

---

## [HTM Studio link](https://discourse.numenta.org/t/htm-studio-link/11930)

<div class="topic-metadata">

**Author:** [@Chris\_Misztur](https://discourse.numenta.org/u/Chris_Misztur)\
**Replies:** 5\
**Last updated:** [August 6, 2025, 5:08pm UTC](https://discourse.numenta.org/t/htm-studio-link/11930 "2025-08-06T17:08:59Z")

</div>

Does anyone have a working download link to HTM Studio for Windows? Thanks Chris

---

## [Multisynapse optical network outperforms digital AI models](https://discourse.numenta.org/t/multisynapse-optical-network-outperforms-digital-ai-models/11943)

<div class="topic-metadata">

**Author:** [@SeanOConnor](https://discourse.numenta.org/u/SeanOConnor)\
**Replies:** 28\
**Last updated:** [August 6, 2025, 8:03am UTC](https://discourse.numenta.org/t/multisynapse-optical-network-outperforms-digital-ai-models/11943 "2025-08-06T08:03:06Z")

</div>

I found this on the internet: https://phys.org/news/2025-07-multisynapse-optical-network-outperforms-digital.html Basically they are doing an optical (quasi) random projection using light and then a digital computer re…

---

## [Rare information about the history of Associative Memory](https://discourse.numenta.org/t/rare-information-about-the-history-of-associative-memory/11994)

<div class="topic-metadata">

**Author:** [@SeanOConnor](https://discourse.numenta.org/u/SeanOConnor)\
**Replies:** 5\
**Last updated:** [August 5, 2025, 12:55pm UTC](https://discourse.numenta.org/t/rare-information-about-the-history-of-associative-memory/11994 "2025-08-05T12:55:51Z")

</div>

Here is an article that contains rare information about the history of associative memory. The Hopfield net cannot work well because while the weighted sum can provide error correction (to allow attractor states) it is …

---

## [ReLU as a switch](https://discourse.numenta.org/t/relu-as-a-switch/11967)

<div class="topic-metadata">

**Author:** [@SeanOConnor](https://discourse.numenta.org/u/SeanOConnor)\
**Replies:** 1\
**Last updated:** [July 30, 2025, 1:33am UTC](https://discourse.numenta.org/t/relu-as-a-switch/11967 "2025-07-30T01:33:39Z")

</div>

I wrote this blog: https://sciencelimelight.blogspot.com/2025/07/relu-as-switch.html

---

## [Learning probabilistic representations with randomly connected neural circuits](https://discourse.numenta.org/t/learning-probabilistic-representations-with-randomly-connected-neural-circuits/11964)

<div class="topic-metadata">

**Author:** [@SeanOConnor](https://discourse.numenta.org/u/SeanOConnor)\
**Replies:** 1\
**Last updated:** [July 29, 2025, 3:06pm UTC](https://discourse.numenta.org/t/learning-probabilistic-representations-with-randomly-connected-neural-circuits/11964 "2025-07-29T15:06:14Z")

</div>

This paper has nice graphics: https://www.biorxiv.org/content/10.1101/478545v1.full Thank you.

---

## [Interesting idea - temporal difference imaging](https://discourse.numenta.org/t/interesting-idea-temporal-difference-imaging/11959)

<div class="topic-metadata">

**Author:** [@SeanOConnor](https://discourse.numenta.org/u/SeanOConnor)\
**Replies:** 2\
**Last updated:** [July 25, 2025, 10:41pm UTC](https://discourse.numenta.org/t/interesting-idea-temporal-difference-imaging/11959 "2025-07-25T22:41:37Z")

</div>

This is kind of interesting. Especially the exponential effect from using multiple cameras: https://youtu.be/zFiubdrJqqI?si=JfSwgauxXGRihYXO

---

## [Neural Networks Likely Operate as Hierarchical Associative Memory](https://discourse.numenta.org/t/neural-networks-likely-operate-as-hierarchical-associative-memory/11961)

<div class="topic-metadata">

**Author:** [@SeanOConnor](https://discourse.numenta.org/u/SeanOConnor)\
**Replies:** 0\
**Last updated:** [July 25, 2025, 10:37pm UTC](https://discourse.numenta.org/t/neural-networks-likely-operate-as-hierarchical-associative-memory/11961 "2025-07-25T22:37:10Z")

</div>

I wrote this blog listing 4 points supporting the idea that deep neural networks are likely hierarchical associative memory. https://sciencelimelight.blogspot.com/2025/07/artificial-neural-networks-likely.html

---

## [World’s first Large Visual Memory Model](https://discourse.numenta.org/t/world-s-first-large-visual-memory-model/11960)

<div class="topic-metadata">

**Author:** [@MTIzNDU2Nzg5](https://discourse.numenta.org/u/MTIzNDU2Nzg5)\
**Replies:** 0\
**Last updated:** [July 25, 2025, 12:14am UTC](https://discourse.numenta.org/t/world-s-first-large-visual-memory-model/11960 "2025-07-25T00:14:44Z")

</div>

Memories.ai: we built the world’s first Large Visual Memory Model - to give AI human-like visual memories. https://x.com/shawnshenjx/status/1948424066853122129?

---

## [Hierarchical Reasoning Model](https://discourse.numenta.org/t/hierarchical-reasoning-model/11958)

<div class="topic-metadata">

**Author:** [@cezar\_t](https://discourse.numenta.org/u/cezar_t)\
**Replies:** 1\
**Last updated:** [July 23, 2025, 2:16pm UTC](https://discourse.numenta.org/t/hierarchical-reasoning-model/11958 "2025-07-23T14:16:12Z")

</div>

HRM excels at tasks that demand extensive search and backtracking. Using only 1,000 input-output examples, without pre-training or CoT supervision, HRM learns to solve problems that are intractable for even the most adv…

[Previous page](https://discourse.numenta.org/latest.md?no_subcategories=false&page=1)

[Next page](https://discourse.numenta.org/latest.md?no_subcategories=false&page=3)