SBMT 2021: Can Neuroscience Insights Transform AI? - Lawrence Spracklen

•

1 recomendación•370 vistas

Numenta's Director of ML Architecture Lawrence Spracklen presented a talk at the SBMT Annual Congress on July 10th, 2021. He talked about how neuroscience principles can inspire better machine learning algorithms.

Ciencias

Can Neuroscience
insights transform AI?
Dr. Lawrence Spracklen
Director, ML Architecture

Numenta
Developing machine intelligence
through neocortical theory
• Understand how the brain works
• Apply neocortical principles to AI
Developed the “Thousand Brains”
theory of how the neocortex works

Artificial Neural Networks (ANNs)
Layer 1
Layer 2
Layer 3
Layer N
Input
Output Dense, fully-connected and computationally expensive

Traditional approach to ANNs
Perform matrix multiplications very fast
• GPUs have become AI workhorses
• 500+ trillion arithmetic operations per
second per card
• Hardware performance doubles
every few years
• Hardware cannot keep pace with
growth in model size
• Exploding AI costs
• 2018 : BERT cost $6K+ to train
• 2020 : GPT-3 cost $10M+ to train
3-years
17,000X
increase
Figure credit

AI Today
Incredible progress, but, at what cost….
• Vast models
• Trillions of parameters
• Expensive training
• Massive compute, power & training data requirements
• Catastrophic forgetting
• Static task-specific models that can’t learn
• Fragility
• Significant real-world dangers
Still a long way from AGI (Artificial General Intelligence)
• Can we continue down this current path?

Going forward
1. Improve model performance
2. Decrease frequency of retraining
3. Decrease training complexity
• Both Algorithms and Hardware need to evolve
• Focus on just one dimension doesn’t solve the problem
• “Faster Horse” issues
• Ensuring synergy provides a lasting solution
• Hardware feasibility needs to influence algorithm evolution
• And vice versa

Can Neuroscience help?
Examine the Neocortex
• Neuron interconnections are sparse
• Neuron activations are sparse
• Neurons are significantly more complex than AI’s point
neuron abstraction
• Humans can learn from very limited examples
Numenta’s Roadmap

Make models fast
Sparse models
• Deliver comparable accuracy with up to 20X fewer parameters
• Also leverage activation sparsity for multiplicative benefits
• 100X+ reduction in compute costs
• Hardware needs to be capable of exploiting sparsity
• Efficiently avoid multiplying by the zeros!
• 100X on FPGAs, 20X on CPUs with Numenta’s sparsity

Always be learning
Active dendrites
• Point neurons only incorporate proximal synapses
• Small proportion of neuron’s total synapses
• Extend artificial neurons to incorporate distal synapses
• Basal synapses used to modulate neuron behavior
• Applying context signals enables networks to learn multiple
tasks and facilitates online continuous learning
• Primes relevant neurons based on context
• Unsupervised determination of context is critical
image credit

Reduce learning repetition
Reference frames
• Training an ANN to recognize even a cup requires many images
• 100s of pictures of cups at different orientations, distances, designs and colors
• Separate problem into two base components
• Invariant representation of object
• Understanding of positional relationship to object
• Create robust position independent representations of objects
• Make observer orientation and distance explicit considerations
• Inspired by human grid cells
• Object independent
• Significantly reduces number of training examples

Conclusions
• Continued progress in AI is threatened by exponentially
increasing costs
• Insights from the Neocortex provide critical insights for how to
evolve AI
• Numenta has developed neocortex inspired roadmap to AGI
• Already demonstrated 100X AI model speedups using brain
inspired sparsity
• Working to incorporate continual learning and positionally
invariant representations into AI systems
• Reduce both retraining frequency & number of training examples
• Cumulative benefits reduce AI costs by many orders of
magnitude

THANK YOU
Questions?
lspracklen@numenta.com
https://numenta.com

Más contenido relacionado

La actualidad más candente

Hard & soft computing

SCAROLINEECE

KeithWiley_NeuromorphicComputing_and_CM1K_and_emulator_talk_wide

Keith Wiley

What the Brain says about Machine Intelligence

Numenta

Loihi many core_neuromorphic_chip

Mehmood Saleem

The Blue Brain Project is the first attempt to reverse-engineer the brain of mammalian, so that simulations of the function of brain can be understood. BLUE BRAIN is the name of the world's first virtual brain, which is a machine that can function as human brain. Today, scientists are attempting to create an artificial brain that can think, respond, take decision, and store anything in memory as like humans do. The primary goal of this project is to preserve the knowledge, intelligence, personalities, feelings and memories of a person that can be used for the development of the human society.

The blue brain project.docx

Sadhana Singh

Neuromorphic computing

Artificial brains

Ibm truenorth

Digital immortality

Introduction to Soft Computing

Mind uploading

soft-computing

soft computing manoj

Ed Safford III MetroCon 2015 Verification, Validation, and Deployment of Hybr...

Edward L S Safford III

From Cognitive Computing to Artificial Intelligence Dr Peter WaggettDire...

Sudeep Sakalle

Blue Brain

Gaurav Hajare

Most of today's AI technologies are extensions of ANN models that were envisioned twenty years ago. While they have some amazing capabilities, they are not truly intelligent. To attain truly intelligent machines, our approach is to understand the only thing we can all agree is intelligent today: the human brain. HTM is an AI technology that is biologically-constrained. All major algorithms at work in an HTM system were uncovered by years of intensive neuroscience research. In this presentation, I'll describe the major mechanism of HTM at a high level, and lay a path towards the future of truly intelligent machines. http://numenta.com/open-source-community/ Matt Taylor, Numenta

Hierarchical Temporal Memory: Computing Like the Brain - Matt Taylor, Numenta

WithTheBest

Deep Networks with Neuromorphic VLSI devices

Giacomo Indiveri

Introduction to soft computing

Siksha 'O' Anusandhan (Deemed to be University )

Geek Night 17.0 - Artificial Intelligence and Machine Learning

GeekNightHyderabad

La actualidad más candente (20)

Hard & soft computing

KeithWiley_NeuromorphicComputing_and_CM1K_and_emulator_talk_wide

What the Brain says about Machine Intelligence

Loihi many core_neuromorphic_chip

The blue brain project.docx

Neuromorphic computing

Artificial brains

Ibm truenorth

Digital immortality

Introduction to Soft Computing

Mind uploading

soft-computing

soft computing manoj

Ed Safford III MetroCon 2015 Verification, Validation, and Deployment of Hybr...

From Cognitive Computing to Artificial Intelligence Dr Peter WaggettDire...

Blue Brain

Hierarchical Temporal Memory: Computing Like the Brain - Matt Taylor, Numenta

Deep Networks with Neuromorphic VLSI devices

Introduction to soft computing

Geek Night 17.0 - Artificial Intelligence and Machine Learning

Similar a SBMT 2021: Can Neuroscience Insights Transform AI? - Lawrence Spracklen

In this talk, after a brief overview of AI concepts in particular Machine Learning (ML) techniques, some of the well-known computer design concepts for high performance and power efficiency are presented. Subsequently, those techniques that have had a promising impact for computing ML algorithms are discussed. Deep learning has emerged as a game changer for many applications in various fields of engineering and medical sciences. Although the primary computation function is matrix vector multiplication, many competing efficient implementations of this primary function have been proposed and put into practice. This talk will review and compare some of those techniques that are used for ML computer design.

Computer Design Concepts for Machine Learning

Facultad de Informática UCM

HiPEAC 2019 Workshop - Real-Time Modelling Visual Scenes with Biological Insp...

Tulipp. Eu

For the full video of this presentation, please visit: http://www.embedded-vision.com/platinum-members/embedded-vision-alliance/embedded-vision-training/videos/pages/may-2015-embedded-vision-summit-baidu For more information about embedded vision, please visit: http://www.embedded-vision.com Dr. Ren Wu, former distinguished scientist at Baidu's Institute of Deep Learning (IDL), presents the keynote talk, "Enabling Ubiquitous Visual Intelligence Through Deep Learning," at the May 2015 Embedded Vision Summit. Deep learning techniques have been making headlines lately in computer vision research. Using techniques inspired by the human brain, deep learning employs massive replication of simple algorithms which learn to distinguish objects through training on vast numbers of examples. Neural networks trained in this way are gaining the ability to recognize objects as accurately as humans. Some experts believe that deep learning will transform the field of vision, enabling the widespread deployment of visual intelligence in many types of systems and applications. But there are many practical problems to be solved before this goal can be reached. For example, how can we create the massive sets of real-world images required to train neural networks? And given their massive computational requirements, how can we deploy neural networks into applications like mobile and wearable devices with tight cost and power consumption constraints? In this talk, Ren shares an insider’s perspective on these and other critical questions related to the practical use of neural networks for vision, based on the pioneering work being conducted by his former team at Baidu. Note 1: Regarding the ImageNet results included in this presentation, the organizers of the ImageNet Large Scale Visual Recognition Challenge (ILSVRC) have said: “Because of the violation of the regulations of the test server, these results may not be directly comparable to results obtained and reported by other teams.” (http://www.image-net.org/challenges/LSVRC/announcement-June-2-2015) Note 2: The presenter, Ren Wu, has told the Embedded Vision Alliance that “There was some ambiguity with the rules. According to the ‘official’ interpretation of the rules, there should be no more than 52 submissions within a half year. For us, we achieved the reported results after 200 tests total within a half year. We believe there is no way to obtain any measurable gains, nor did we try to obtain any gains, from an 'extra' hundred tests as our networks have billions of parameters and are trained by tens of billions of training samples.”

"Enabling Ubiquitous Visual Intelligence Through Deep Learning," a Keynote Pr...

Edge AI and Vision Alliance

Vertex Perspectives | AI Optimized Chipsets | Part II

Vertex Holdings

Deep learning for smart manufacturing

Sunil Kumar Pradhan

In this talk, an overview of current trends in machine learning will be discussed with an emphasize on challenges and opportunities facing this field. It will focus on deep learning methods and applications. Deep learning has emerged as one of the most promising research fields in artificial intelligence. The significant advancements that deep learning methods have brought about for large scale image classification tasks have generated a surge of excitement in applying the techniques to other problems in computer vision and more broadly into other disciplines of computer science. Moreover, the impact of machine learning on education, research, and economy will be briefly presented. The rapid growth of machine learning is positioned to impact our lives in a way that we have not been able to fully imagine. It behooves government leaders to take a lead in developing the necessary resources to ride the projected benefits of machine learning.

Machine Learning Challenges and Opportunities in Education, Industry, and Res...

Facultad de Informática UCM

Introduction to deep learning

Abhishek Bhandwaldar

Zaikun Xu from the Università della Svizzera Italiana presented this deck at the 2016 Switzerland HPC Conference. “In the past decade, deep learning as a life-changing technology, has gained a huge success on various tasks, including image recognition, speech recognition, machine translation, etc. Pio- neered by several research groups, Geoffrey Hinton (U Toronto), Yoshua Benjio (U Montreal), Yann LeCun(NYU), Juergen Schmiduhuber (IDSIA, Switzerland), Deep learning is a renaissance of neural network in the Big data era. Neural network is a learning algorithm that consists of input layer, hidden layers and output layers, where each circle represents a neural and the each arrow connection associates with a weight. The way neural network learns is based on how different between the output of output layer and the ground truth, following by calculating the gradients of this discrepancy w.r.b to the weights and adjust the weight accordingly. Ideally, it will find weights that maps input X to target y with error as lower as possible.” Watch the video presentation: http://insidehpc.com/2016/03/deep-learning/ See more talks in the Swiss Conference Video Gallery: http://insidehpc.com/2016-swiss-hpc-conference/ Sign up for our insideHPC Newsletter: http://insidehpc.com/newsletter

Tutorial on Deep Learning

inside-BigData.com

■ You’ve probably heard that Deep Learning is making news across the world as one of the most promising techniques in machine learning, especially for analyzing image data. With every industry dedicating resources to unlock the deep learning potential, to be competitive, you will want to use these models in tasks such as image tagging, object recognition, speech recognition, and text analysis. In this training session you will build deep learning models using neural networks, explore what they are, what they do, and how. To remove the barrier introduced by designing, training, and tuning networks, and to be able to achieve high performance with less labeled data, you will also build deep learning classifiers tailored to your specific task using pre-trained models, which we call deep features. Also, you’ll develop a clear understanding of the motivation for deep learning, and design intelligent systems that learn from complex and/or large-scale datasets.

Deep learning internals

Anand Narayanan

Digit recognition

btandale

Neural Networks-1

Sai Kumar Dwivedi

Unit one ppt of deeep learning which includes Ann cnn

kartikaursang53

Data Science, Machine Learning and Neural Networks

BICA Labs

Deep learning introduction

Adwait Bhave

MDEC Data Matters Series: machine learning and Deep Learning, A Primer

Poo Kuan Hoong

Human brain based circuit systems

Hariharasudhan Mani

FPGA Hardware Accelerator for Machine Learning

Dr. Swaminathan Kathirvel

NEUROMORPHIC COMPUTING.pptx

komalpawooskar1

Deep learning at the edge: 100x Inference improvement on edge devices

Numenta

Introduction to deep learning

Vishwas Lele

Similar a SBMT 2021: Can Neuroscience Insights Transform AI? - Lawrence Spracklen (20)

Computer Design Concepts for Machine Learning

HiPEAC 2019 Workshop - Real-Time Modelling Visual Scenes with Biological Insp...

"Enabling Ubiquitous Visual Intelligence Through Deep Learning," a Keynote Pr...

Vertex Perspectives | AI Optimized Chipsets | Part II

Deep learning for smart manufacturing

Machine Learning Challenges and Opportunities in Education, Industry, and Res...

Introduction to deep learning

Tutorial on Deep Learning

Deep learning internals

Digit recognition

Neural Networks-1

Unit one ppt of deeep learning which includes Ann cnn

Data Science, Machine Learning and Neural Networks

Deep learning introduction

MDEC Data Matters Series: machine learning and Deep Learning, A Primer

Human brain based circuit systems

FPGA Hardware Accelerator for Machine Learning

NEUROMORPHIC COMPUTING.pptx

Deep learning at the edge: 100x Inference improvement on edge devices

Introduction to deep learning

Más de Numenta

Meetup page: https://www.meetup.com/Brains-Bay/events/284481247/ Neuromodulators are signalling chemicals in the brain, which control the emergence of adaptive learning and behaviour. Neuromodulators including dopamine, acetylcholine, serotonin and noradrenaline operate on a spectrum of spatio-temporal scales in tandem and opposition to reconfigure functions of biological neural networks and to regulate global cognition and state transition. Although neuromodulators are important in shaping cognition, their phenomenology is yet to be fully realized in deep neural networks (DNNs). In this talk, we will give an overview of the biological organizing principles of neuromodulators in adaptive cognition and highlight the competition and cooperation across neuromodulators.

Brains@Bay Meetup: A Primer on Neuromodulatory Systems - Srikanth Ramaswamy

Numenta

Meetup page: https://www.meetup.com/Brains-Bay/events/284481247/ A hallmark of intelligence is the ability to learn new flexible, cognitive behaviors - that is, behaviors that require discovering, storing and exploiting novel information for each new instance of the task. In meta-learning, agents are trained with external algorithms to learn one specific cognitive task. However, animals are able to pick up such cognitive tasks automatically, as a result of their evolved neural architecture and synaptic plasticity mechanisms, including neuromodulation. Here we evolve neural networks, endowed with plastic connections and reward-based neuromodulation, over a sizable set of simple meta-learning tasks based on a framework from computational neuroscience. The resulting evolved networks can automatically acquire a novel simple cognitive task, never seen during evolution, through the spontaneous operation of their evolved neural organization and plasticity system. We suggest that attending to the multiplicity of loops involved in natural learning may provide useful insight into the emergence of intelligent behavior.

Brains@Bay Meetup: How to Evolve Your Own Lab Rat - Thomas Miconi

Numenta

We receive information about the world through our sensors and influence the world through our effectors. Such low-level data has gradually come to play a greater role in AI during its 70-year history. I see this as occurring in four steps, two of which are mostly past and two of which are in progress or yet to come. The first step was to view AI as the design of agents which interact with the world and thereby have sensorimotor experience; this viewpoint became prominent in the 1980s and 1990s. The second step was to view the goal of intelligence in terms of experience, as in the reward signal of optimal control and reinforcement learning. The reward formulation of goals is now widely used but rarely loved. Many would prefer to express goals in non-experiential terms, such as reaching a destination or benefiting humanity, but settle for reward because, as an experiential signal, reward is directly available to the agent without human assistance or interpretation. This is the pattern that we see in all four steps. Initially a non-experiential approach seems more intuitive, is preferred and tried, but ultimately proves a limitation on scaling; the experiential approach is more suited to learning and scaling with computational resources. The third step in the increasing role of experience in AI concerns the agent’s representation of the world’s state. Classically, the state of the world is represented in objective terms external to the agent, such as “the grass is wet” and “the car is ten meters in front of me”, or with probability distributions over world states such as in POMDPs and other Bayesian approaches. Alternatively, the state of the world can be represented experientially in terms of summaries of past experience (e.g., the last four Atari video frames input to DQN) or predictions of future experience (e.g., successor representations). The fourth step is potentially the biggest: world knowledge. Classically, world knowledge has always been expressed in terms far from experience, and this has limited its ability to be learned and maintained. Today we are seeing more calls for knowledge to be predictive and grounded in experience. After reviewing the history and prospects of the four steps, I propose a minimal architecture for an intelligent agent that is entirely grounded in experience.

Brains@Bay Meetup: The Increasing Role of Sensorimotor Experience in Artifici...

Numenta

In this talk, I will propose a conceptual framework sketching a path toward open-ended skill acquisition through the coupling of environmental, morphological, sensorimotor, cognitive, developmental, social, cultural and evolutionary mechanisms. I will illustrate parts of this framework through computational experiments highlighting the key role of intrinsically motivated exploration in the generation of behavioral regularity and diversity. Firstly, I will show how some forms of language can self-organize out of generic exploration mechanisms without any functional pressure to communicate. Secondly, we will see how language — once invented — can be recruited as a cognitive tool that enables compositional imagination and bootstraps open-ended cultural innovation. For more:

Brains@Bay Meetup: Open-ended Skill Acquisition in Humans and Machines: An Ev...

Numenta

Most current deep neural networks learn from a static data set without active interaction with the world. We take a look at how learning through a closed loop between action and perception affects the representations learned in a DNN. We demonstrate how these representations are significantly different from DNNs that learn supervised or unsupervised from a static dataset without interaction. These representations are much sparser and encode meaningful content in an efficient way. Even an agent who learned without any external supervision, purely through curious interaction with the world, acquires encodings of the high dimensional visual input that enable the agent to recognize objects using only a handful of labeled examples. Our results highlight the capabilities that emerge from letting DNNs learn more similar to biological brains, though sensorimotor interaction with the world. For more:

Brains@Bay Meetup: The Effect of Sensorimotor Learning on the Learned Represe...

Numenta

Jeff Hawkins presents a talk on "How the Brain Uses Reference Frames to Model the World, Why AI Needs to do the Same." In this talk, he gives an overview of The Thousand Brains Theory and discusses how machine intelligence can benefit from working on the same principles as the neocortex. This talk was first presented at the NAISys conference on November 10, 2020. You can find a re-recording of the talk here: https://youtu.be/mGSG7I9VKDU

Jeff Hawkins NAISys 2020: How the Brain Uses Reference Frames, Why AI Needs t...

Numenta

In this research meeting, guest Stephen Omohundro gave a fascinating talk on GPT-3, the new massive OpenAI Natural Language Processing model. He reviewed the network architecture, training process, and results in the context of past work. There was extensive discussion on the implications for NLP and for Machine Intelligence / AGI. Link to GPT-3 paper: https://arxiv.org/abs/2005.14165 Link to YouTube recording of Steve's talk: https://youtu.be/0ZVOmBp29E0

OpenAI’s GPT 3 Language Model - guest Steve Omohundro

Numenta

CVPR 2020 Workshop: Sparsity in the neocortex, and its implications for conti...

Numenta

Sparsity In The Neocortex, And Its Implications For Machine Learning

Numenta

Recent advances in reverse engineering the neocortex reveal that it is a highly-distributed sensory-motor modeling system. Each cortical column learns complete models of observed objects through movement and sensation. The columns use long-range connections to vote on what objects are currently being observed. In this talk, we introduce the key elements of this theory and describe how these elements can be introduced into current machine learning techniques to improve their capabilities, robustness, and power requirements.

The Thousand Brains Theory: A Framework for Understanding the Neocortex and B...

Numenta

Jeff Hawkins Human Brain Project Summit Keynote: "Location, Location, Locatio...

Numenta

Location, Location, Location - A Framework for Intelligence and Cortical Comp...

Numenta

Have We Missed Half of What the Neocortex Does? A New Predictive Framework ...

Numenta

Locations in the Neocortex: A Theory of Sensorimotor Prediction Using Cortica...

Numenta

The Predictive Neuron: How Active Dendrites Enable Spatiotemporal Computation...

Numenta

These are Jeff Hawkins' slides from the Computational Theories of the Brain Workshop held at the Simons Institute at UC Berkeley on April 17, 2018. Abstract: In this talk, I propose that the neocortex learns models of objects using the same methods that the entorhinal cortex uses to map environments. I propose that each cortical column contains cells that are equivalent to grid cells. These cells represent the location of sensor patches relative to objects in the world. As we move our sensors, the location of the sensor is paired with sensory input to learn the structure of objects. I explore the evidence for this hypothesis, propose specific cellular mechanisms that the hypothesis requires, and suggest how the hypothesis could be tested. References: “A Theory of How Columns in the Neocortex Enable Learning the Structure of the World” by Jeff Hawkins, Subutai Ahmad, YuWei Cui (2017) “Place Cells, Grid Cells, and the Brain’s Spatial Representation System” by Edvard Moser, Emilio Kropff, May-Britt Moser (2008) “Evidence for grid cells in a human memory network” by Christian Doeller, Caswell Barry, Neil Burgess (2010)

Does the neocortex use grid cell-like mechanisms to learn the structure of ob...

Numenta

These are slides on a workshop Subutai Ahmad hosted on March 5, 2018 at the Computational and Systems Neuroscience Meeting (Cosyne) 2018. About: This workshop on long-range cortical circuits is focused on our peer-reviewed paper, “A Theory of How Columns in the Neocortex Enable Learning the Structure of the World.” Subutai discussed the inference mechanism introduced in the paper, our theory of location information, and how long-range connections allow columns to integrate inputs over space to perform object recognition.

Could A Model Of Predictive Voting Explain Many Long-Range Connections? by Su...

Numenta

This was a presentation given on February 8, 2018 at the European Institute for Theoretical Neuroscience (EITN)'s Dendritic Integration and Computation with Active Dendrites Workshop. The workshop is aimed at putting together experiments, models and recent neuromorphic systems aiming at understanding the computational properties conferred by dendrites in neural systems. It is focused particularly on the excitable properties of dendrites and the type of computation they can implement.

The Predictive Neuron: How Active Dendrites Enable Spatiotemporal Computation...

Numenta

This was a presentation given on December 15, 2017 at the MIT Center for Brains, Minds + Machines as part of their Brains, Minds and Machines Seminar Series. You can watch the recording of the presentation after Slide 1. In this talk, Jeff describes a theory that sensory regions of the neocortex process two inputs. One input is the well-known sensory data arriving via thalamic relay cells. We propose the second input is a representation of allocentric location. The allocentric location represents where the sensed feature is relative to the object being sensed, in an object-centric reference frame. As the sensors move, cortical columns learn complete models of objects by integrating sensory features and location representations over time. Lateral projections allow columns to rapidly reach a consensus of what object is being sensed. We propose that the representation of allocentric location is derived locally, in layer 6 of each column, using the same tiling principles as grid cells in the entorhinal cortex. Because individual cortical columns are able to model complete complex objects, cortical regions are far more powerful than currently believed. The inclusion of allocentric location offers the possibility of rapid progress in understanding the function of numerous aspects of cortical anatomy. Jeff discusses material from these two papers. Others can be found at https://numenta.com/papers A Theory of How Columns in the Neocortex Enable Learning the Structure of the World URL: https://doi.org/10.3389/fncir.2017.00081 Why Neurons Have Thousands of Synapses, A Theory of Sequence Memory in the Neocortex URL: https://doi.org/10.3389/fncir.2016.00023

Have We Missed Half of What the Neocortex Does? by Jeff Hawkins (12/15/2017)

Numenta

Marcus gave a talk called "Recognizing Locations on Objects" during the HTM Meetup on 11/03/2017. The brain learns and recognizes objects with independent moving sensors. It’s not obvious how a network of neurons would do this. Numenta has suggested that the brain solves this by computing each sensor’s location relative to the object, and learning the object as a set of features-at-locations. Marcus showed how the brain might determine this “location relative to the object.” He extended the model from Numenta’s recent paper, "A Theory of How Columns in the Neocortex Enable Learning the Structure of the World," so that it computes this location. This extended model takes two inputs: each sensor’s input, and each sensor’s “location relative to the body.” The model connects the columns in such a way that a column can compute its “location relative to the object” from another column’s “location relative to object.” When a column senses a feature, it recalls a union of all locations where it has sensed this feature, then the columns work together to narrow their unions. This extended model essentially takes its sensory input and asks, “Do I know any objects that contain this spatial arrangement of features?”

Recognizing Locations on Objects by Marcus Lewis

Numenta

Más de Numenta (20)