Picture for Subutai Ahmad

Subutai Ahmad

Numenta

Avoiding Catastrophe: Active Dendrites Enable Multi-Task Learning in Dynamic Environments

Add code
Dec 31, 2021
Figure 1 for Avoiding Catastrophe: Active Dendrites Enable Multi-Task Learning in Dynamic Environments
Figure 2 for Avoiding Catastrophe: Active Dendrites Enable Multi-Task Learning in Dynamic Environments
Figure 3 for Avoiding Catastrophe: Active Dendrites Enable Multi-Task Learning in Dynamic Environments
Figure 4 for Avoiding Catastrophe: Active Dendrites Enable Multi-Task Learning in Dynamic Environments
Viaarxiv icon

Two Sparsities Are Better Than One: Unlocking the Performance Benefits of Sparse-Sparse Networks

Add code
Dec 27, 2021
Figure 1 for Two Sparsities Are Better Than One: Unlocking the Performance Benefits of Sparse-Sparse Networks
Figure 2 for Two Sparsities Are Better Than One: Unlocking the Performance Benefits of Sparse-Sparse Networks
Figure 3 for Two Sparsities Are Better Than One: Unlocking the Performance Benefits of Sparse-Sparse Networks
Figure 4 for Two Sparsities Are Better Than One: Unlocking the Performance Benefits of Sparse-Sparse Networks
Viaarxiv icon

Grid Cell Path Integration For Movement-Based Visual Object Recognition

Add code
Feb 17, 2021
Figure 1 for Grid Cell Path Integration For Movement-Based Visual Object Recognition
Figure 2 for Grid Cell Path Integration For Movement-Based Visual Object Recognition
Figure 3 for Grid Cell Path Integration For Movement-Based Visual Object Recognition
Figure 4 for Grid Cell Path Integration For Movement-Based Visual Object Recognition
Viaarxiv icon

Long Distance Relationships without Time Travel: Boosting the Performance of a Sparse Predictive Autoencoder in Sequence Modeling

Add code
Dec 02, 2019
Figure 1 for Long Distance Relationships without Time Travel: Boosting the Performance of a Sparse Predictive Autoencoder in Sequence Modeling
Figure 2 for Long Distance Relationships without Time Travel: Boosting the Performance of a Sparse Predictive Autoencoder in Sequence Modeling
Figure 3 for Long Distance Relationships without Time Travel: Boosting the Performance of a Sparse Predictive Autoencoder in Sequence Modeling
Figure 4 for Long Distance Relationships without Time Travel: Boosting the Performance of a Sparse Predictive Autoencoder in Sequence Modeling
Viaarxiv icon

How Can We Be So Dense? The Benefits of Using Highly Sparse Representations

Add code
Apr 02, 2019
Figure 1 for How Can We Be So Dense? The Benefits of Using Highly Sparse Representations
Figure 2 for How Can We Be So Dense? The Benefits of Using Highly Sparse Representations
Figure 3 for How Can We Be So Dense? The Benefits of Using Highly Sparse Representations
Figure 4 for How Can We Be So Dense? The Benefits of Using Highly Sparse Representations
Viaarxiv icon

Real-Time Anomaly Detection for Streaming Analytics

Add code
Jul 08, 2016
Figure 1 for Real-Time Anomaly Detection for Streaming Analytics
Figure 2 for Real-Time Anomaly Detection for Streaming Analytics
Figure 3 for Real-Time Anomaly Detection for Streaming Analytics
Figure 4 for Real-Time Anomaly Detection for Streaming Analytics
Viaarxiv icon

How do neurons operate on sparse distributed representations? A mathematical theory of sparsity, neurons and active dendrites

Add code
May 13, 2016
Figure 1 for How do neurons operate on sparse distributed representations? A mathematical theory of sparsity, neurons and active dendrites
Figure 2 for How do neurons operate on sparse distributed representations? A mathematical theory of sparsity, neurons and active dendrites
Figure 3 for How do neurons operate on sparse distributed representations? A mathematical theory of sparsity, neurons and active dendrites
Figure 4 for How do neurons operate on sparse distributed representations? A mathematical theory of sparsity, neurons and active dendrites
Viaarxiv icon

Continuous online sequence learning with an unsupervised neural network model

Add code
Apr 28, 2016
Viaarxiv icon

Porting HTM Models to the Heidelberg Neuromorphic Computing Platform

Add code
Feb 09, 2016
Figure 1 for Porting HTM Models to the Heidelberg Neuromorphic Computing Platform
Figure 2 for Porting HTM Models to the Heidelberg Neuromorphic Computing Platform
Figure 3 for Porting HTM Models to the Heidelberg Neuromorphic Computing Platform
Figure 4 for Porting HTM Models to the Heidelberg Neuromorphic Computing Platform
Viaarxiv icon

Why Neurons Have Thousands of Synapses, A Theory of Sequence Memory in Neocortex

Add code
Dec 01, 2015
Figure 1 for Why Neurons Have Thousands of Synapses, A Theory of Sequence Memory in Neocortex
Figure 2 for Why Neurons Have Thousands of Synapses, A Theory of Sequence Memory in Neocortex
Figure 3 for Why Neurons Have Thousands of Synapses, A Theory of Sequence Memory in Neocortex
Figure 4 for Why Neurons Have Thousands of Synapses, A Theory of Sequence Memory in Neocortex
Viaarxiv icon