CLA Glossary for ML

Fergal Byrne edited this page Apr 3, 2014 · 4 revisions
Clone this wiki locally

CLA Glossary

:::::::::::::::::::::::::::::::::::META::::::::::::::::::::::::::::::::::

What should go in this area?

This is a re-write of the CLA Whitepaper Glossary (plus any other terms/concepts we want to highlight) where the explanations are targeted at ML/AI researchers.

Terms, concepts and strategies that appear in CLA documentation, presentations, or code should be listed here.

For each term:

  • Similar terms or strategies in the traditional community should be briefly discussed.
    • Citations for descriptions of ML/AI terms. e.g. Links to scholarly articles or wikipedia articles.
  • The distinction between the CLA term and the traditional term should be carefully noted.
  • Links to NuPIC documentation, or code that explore this idea further.

:::::::::::::::::::::::::::::::::::END - META:::::::::::::::::::::::::::

Active State

n. A state in which Cells are active due to Feed-Forward input.

  • NN equivalent: Output of a neuron. Binary valued.

Boosting

TODO

Bottom-Up

adj. Synonym for Feed-Forward

Cells

n. HTM equivalent of a Neuron in NNs. Basic unit of computation in CLA. Cells are more complex than NN neurons, in that they have two types of Dendrites - Proximal (feed-forward) and Distal (predictive) - and a more complex internal state.

Instead of having a scalar output as in NN, Cells have a binary Active State. Again unlike NNs, Cell activity is based on a combination of factors, not just feed-forward input.

Cells are organized into Columns in HTM regions. In NuPIC, all Cells in a Column share FF input (ie they have a single proximal dendrite segment for feed-forward).

Coincident Activity

n. Two or more Cells are active at the same time.

Column

n. A group of one or more Cells that function as a unit in an HTM Region

Cells within a column represent the same feed-forward input, but in different contexts.

Dendrite Segment

n. A unit of integration of Synapses associated with Cells and Columns

HTMs have two different types of dendrite segments:

  • Distal segments are associated with lateral connections to a cell, and measure coincident activity. When the number of active synapses on the dendrite segment exceeds a threshold, the associated cell enters the predictive state.
  • The Proximal dendrite is associated with feed-forward connections to a column (ie shared by the column's cells). The number of active synapses is summed to generate the feed-forward activation of a column.

Desired Density [NuPIC-specific]

param. Desired percentage of Columns active due to Feed-Forward input to a Region

The percentage only applies within a radius that varies based on the fan-out of feed-forward inputs. It is desired because the percentage varies some based on the particular input.

The desired density of active columns within a local inhibition area (the size of which is set by the internally calculated inhibitionRadius, which is in turn determined from the average size of the connected potential pools of all columns). The inhibition logic will ensure that at most N columns remain on within a local inhibition area, where N = localAreaDensity * (total number of columns in inhibition area).

Dimensions, Input, Spatial Pooler [NuPIC-specific]

param. A number, list or numpy array representing the dimensions of the input vector.

Format is [height, width, depth, ...], where each value represents the size of the dimension. For a topology of one dimension with 100 inputs use 100, or [100]. For a two dimensional topology of 10x5 use [10,5]

Dimensions, Columns, Spatial Pooler [NuPIC-specific]

param. A number, list or numpy array representing the dimensions of the columns in the region.

Format is [height, width, depth, ...], where each value represents the size of the dimension. For a topology of one dimesion with 2000 columns use 2000, or [2000]. For a three dimensional topology of 32x64x16 use [32,64,16]

Duty Cycle [NuPIC-specific]

param.

TODO

Duty Cycle, Minimum Percent Overlap [NuPIC-specific]

param. A number between 0 and 1.0, used to set a floor on how often a column should have at least stimulusThreshold active inputs.

TODO specify param name minSomething

Periodically, each column looks at the overlap duty cycle of all other column within its inhibition radius and sets its own internal minimal acceptable duty cycle to: minPctDutyCycleBeforeInh * max(other columns' duty cycles).

On each iteration, any column whose overlap duty cycle falls below this computed value will get all of its permanence values boosted up by synPermActiveInc. Raising all permanences in response to a sub-par duty cycle before inhibition allows a cell to search for new inputs when either its previously learned inputs are no longer ever active, or when the vast majority of them have been "hijacked" by other columns.

Duty Cycle, Minimum Percent Active

TODO

A number between 0 and 1.0, used to set a floor on how often a column should be activate. Periodically, each column looks at the activity duty cycle of all other columns within its inhibition radius and sets its own internal minimal acceptable duty cycle to: minPctDutyCycleAfterInh * max(other columns' duty cycles). On each iteration, any column whose duty cycle after inhibition falls below this computed value will get its internal boost factor increased.

Duty Cycle, Period

TODO The period used to calculate duty cycles. Higher values make it take longer to respond to changes in boost or synPerConnectedCell. Shorter values make it more unstable and likely to oscillate.

Encoder

obj. Unlike in NNs, inputs to HTMs are binary vectors (or higher-dimensional matrices). Encoders are needed to convert real-world or machine-generated data into bit arrays for use by HTM Regions. The output of an encoder is similar to a Sparse Distributed Representation (SDR) but usually is not nearly as sparse.

Encoders in NuPIC are often nested to combine different fields in the data (e.g. timestamp and temperature), and also to combine different semantic aspects of a single value (e.g. timestamp.timeofday and timestamp.isweekend). When creating an encoder, you usually specify the overall width of the output n bits, and the width of the window of w on-bits in each encoding.

Encoder, Category

obj. Encodes each of a list of categories with a unique set of bits. In NuPIC, a simple category encoder might have the following outputs:

Input Output (n = 16, w = 4)
Animal 1111000000000000
Vegetable 0000111100000000
Mineral 0000000011110000
Other 0000000000001111

Encoder, Passthrough

obj. Passes its input straight through to its receiver. What you use when the input is already an SDR, for example when the output of one Region is passed to a higher-level one.

Also used when the input to the whole system is an SDR, eg. for CEPT Retina representations of text.

Encoder, Scalar

obj. Encodes a scalar (floating-point) value.

Feed-Forward (aka FF)

Moving in a direction away from an input, or from a lower Level to a higher Level in a Hierarchy (sometimes called Bottom-Up).

In CLA, feed-forward inputs to a Cell are connected to its

Feedback

TODO moving in a direction towards an input, or from a higher Level to a lower level in a Hierarchy (sometimes called Top-Down)

First Order Prediction

TODO a prediction based only on the current input and not on the prior inputs – compare to Variable Order Prediction

Hierarchical Temporal Memory (HTM)

TODO a technology that replicates some of the structural and algorithmic functions of the neocortex

Hierarchy

TODO a network of connected elements where the connections between the elements are uniquely identified as Feed-Forward or Feedback

HTM Cortical Learning Algorithms

TODO the suite of functions for Spatial Pooling, Temporal Pooling, and learning and forgetting that comprise an HTM Region, also referred to as HTM Learning Algorithms

HTM Network

TODO a Hierarchy of HTM Regions

HTM Region

TODO the main unit of memory and Prediction in an HTM

An HTM region is comprised of a layer of highly interconnected cells arranged in columns. An HTM region today has a single layer of cells, whereas in the neocortex (and ultimately in HTM), a region will have multiple layers of cells. When referred to in the context of it’s position in a hierarchy, a region may be referred to as a level.

Inference

TODO recognizing a spatial and temporal input pattern as similar to previously learned patterns

Inhibition

TODO

Inhibition, Number Active Per Area - Spatial Pooler

TODO

An alternate way to control the density of the active columns. If numActivePerInhArea is specified then localAreaDensity must less than 0, and vice versa. When using numActivePerInhArea, the inhibition logic will insure that at most 'numActivePerInhArea' columns remain ON within a local inhibition area (the size of which is set by the internally calculated inhibitionRadius, which is in turn determined from the average size of the connected receptive fields of all columns). When using this method, as columns learn and grow their effective receptive fields, the inhibitionRadius will grow, and hence the net density of the active columns will decrease. This is in contrast to the localAreaDensity method, which keeps the density of active columns the same regardless of the size of their receptive fields.

Inhibition, Global - Spatial Pooler

TODO defines the area around a Column that it actively inhibits

If true, then during inhibition phase the winning columns are selected as the most active columns from the region as a whole. Otherwise, the winning columns are selected with resepct to their local neighborhoods. using global inhibition boosts performance x60.

Inhibition Radius

TODO defines the area around a Column that it actively inhibits

Input, Encoder

TODO

Input - Spatial Pooler

TODO

Input - Temporal Pooler

TODO

Lateral Connections

TODO connections between Cells within the same Region

Level

TODO an HTM Region in the context of the Hierarchy

Neuron

TODO an information processing Cell in the brain

In this document, we use the word neuron specifically when referring to biological cells, and “cell” when referring to the HTM unit of computation.

Overlap - Spatial Pooler

TODO

Permanence

TODO a scalar value which indicates the connection state of a Potential Synapse

A permanence value below a threshold indicates the synapse is not formed. A permanence value above the threshold indicates the synapse is valid. Learning in an HTM region is accomplished by modifying permanence values of potential synapses.

Permanence Decrement

TODO The amount by which an inactive synapse is decremented in each round.

Permanence Decrement, Orphaned columns

TODO The amount by which an inactive synapse is decremented in each round.

The amount by which to decrease the permanence of an active synapse on a column which has high overlap with the input, but was inhibited (an "orphan" column).

Permanence Increment

TODO The amount by which an active synapse is incremented in each round.

Potential Synapse (Pool)

TODO the subset of all Cells that could potentially form Synapses with a particular Dendrite Segment

Only a subset of potential synapses will be valid synapses at any time based on their permanence value.

Potential Radius - Spatial Pooler

TODO the subset of all Cells that could potentially form Synapses with a particular Dendrite Segment

This parameter determines the extent of the input that each column can potentially be connected to. This can be thought of as the input bits that are visible to each column, or a 'receptiveField' of the field of vision. A large enough value will result in the 'global coverage', meaning that each column can potentially be connected to every input bit. This parameter defines a square (or hyper square) area: a column will have a max square potential pool with sides of length 2 * potentialRadius + 1.

Potential Percent - Spatial Pooler

The percent of the inputs, within a column's potential radius, that a column can be connected to. If set to 1, the column will be connected to every input within its potential radius. This parameter is used to give each column a unique potential pool when a large potentialRadius causes overlap between the columns. At initialization time we choose ((2*potentialRadius + 1)^(# inputDimensions) * potentialPct) input bits to comprise the column's potential pool.

Prediction

TODO activating Cells (into a predictive state) that will likely become active in the near future due to Feed-Forward input

An HTM region often predicts many possible future inputs at the same time.

Receptive Field

TODO the set of inputs to which a Column or Cell is connected

If the input to an HTM region is organized as a 2D array of bits, then the receptive field can be expressed as a radius within the input space.

Sensor

TODO a source of inputs for an HTM Network

Sparse Distributed Representation (SDR)

TODO representation comprised of many bits in which a small percentage are active and where no single bit is sufficient to convey meaning

Kanerva - http://en.wikipedia.org/wiki/Sparse_distributed_memory

Spatial Pooler

TODO

Spatial Pooling

TODO the process of forming a sparse distributed representation of an input

One of the properties of spatial pooling is that overlapping input patterns map to the same sparse distributed representation.

Sub-Sampling

TODO recognizing a large distributed pattern by matching only a small subset of the active bits in the large pattern

Synapse

TODO connection between Cells formed while learning

Temporal Pooler

TODO

Temporal Pooling

TODO the process of forming a representation of a sequence of input patterns where the resulting representation is more stable than the input

Threshold, Connected - Spatial Pooler

TODO

The default connected threshold. Any synapse whose permanence value is above the connected threshold is a "connected synapse", meaning it can contribute to the cell's firing.

Threshold, Stimulus - Spatial Pooler

TODO

This is a number specifying the minimum number of synapses that must be on in order for a columns to turn ON. The purpose of this is to prevent noise input from activating columns.

Top-Down

TODO synonym for Feedback

Variable Order Prediction

TODO a prediction based on varying amounts of prior context – compare to First Order Prediction

It is called “variable” because the memory to maintain prior context is allocated as needed. Thus a memory system that implements variable order prediction can use context going way back in time without requiring exponential amounts of memory.