CSIM: Readout Class Reference

Readout Class Reference

#include <readout.h>

Inheritance diagram for Readout:

List of all members.

Detailed Description

Base class of all readouts.

This class implements readouts within the simulation. Readouts take the responses of the liquid and calculate a new function with these inputs. Readouts can be trained to approximate some target function. Readouts can be connected to physical models or feedback neurons for closed-loop simulations. Therefore the interface of readouts is similar to that of synapses. The output of a readout can be recorded.

Public Member Functions

Readout (void)
Constructor: Initializes a new readout.
~Readout (void)
Destructor: Frees the memory allocated by the readout.
virtual int advance ()
Advances the readout.
virtual int addIncoming (Advancable *Incoming)
Connects this readout with filters, preprocessors and algorithm objects.
virtual int addOutgoing (Advancable *a)
Connects the readout to a target object which can be a feedback-neuron or a physical model.
virtual int addInputField (csimClass *O, char *fieldname)
Adds a new field to record input data to the readout from object O.
virtual void reset ()
void setAnalogFilter (AnalogFilter *af)
void setSpikeFilter (SpikeFilter *sf)
int appendPreprocessor (Preprocessor *pp)
int prependPreprocessor (Preprocessor *pp)
int insertPreprocessor (Preprocessor *pp, int position)
int getNumberPreprocessors ()
int importPreprocessor (int ind, double *rep, int rep_length)
double * exportPreprocessor (int ind, int *rep_length)
string getPreprocessorFormatDescription (int ind)
int setLearningAlgorithm (Algorithm *learner)
int importAlgorithm (double *rep, int rep_length)
double * exportAlgorithm (int *rep_length)
string getAlgorithmFormatDescription ()
void setRange (double lower_bound, double upper_bound)
uint32 getAnalogFilterID (void)
Get the Id of the Analog-Filter.
bool hasAnalogFilter (void)
Checks whether an Analog-Filter exists.
uint32 getSpikeFilterID (void)
Get the Id of the Spike-Filter.
bool hasSpikeFilter (void)
Checks whether a Spike-Filter exists.
uint32 getAlgorithmID (void)
Get the Id of the Algorithm.
bool hasAlgorithm (void)
Checks whether a Spike-Filter exists.
void getPreprocessorIDs (uint32 *idx)
Get the Ids of the Preprocessors.

Public Attributes

int enabled
Flag: 0 ... readout disabled, 1 ... readout enabled [range=(0,1)];.
double offset
Offset of output values.
double range
Range of output values.

Protected Member Functions

int add (RecField *f)
Add a new field to the recoder.

Protected Attributes

int nRecFields
Number of fields recorded.
int lRecFields
Number of fields allocated.
RecField ** recFields
Array of fieldpointers.
int nAnalogFields
Number of analog (i.e. nonspiking fields) recorded.
int lAnalogFields
Number of analog (i.e. nonspiking fields) allocated.
int * analogIndices
Array of indices for the analog fields.
int nSpikingFields
Number of spiking fields recorded.
int lSpikingFields
Number of spiking fields allocated.
int * spikingIndices
Array of indices for the spiking fields.
SynapseTarget ** outgoing
Pointers to outgoing targets.
int nOutgoing
Number of outgoing targets.
int lOutgoing
Number of outgoing targets allocated.
list< Preprocessor * > pre_processors
List of preprocessors.
AnalogFilter * analogFilter
Filter for analog input.
SpikeFilter * spikeFilter
Filter for spiking input.
Algorithm * algorithm
Learned algorithm.

Static Protected Attributes

static double * m_response = 0
Static working memory for filters and preprocessors.
static double * m_target = 0
Static working memory for filters and preprocessors.
static int lReservedMemory = 0
Length of the reserved memory.
static pthread_mutex_t mutex_temp_memory
Mutex for controlling access to shared memory.
static int nReadouts = 0
Number of readouts created.

Friends

class csimClassInfoDB

Classes

struct RecField
Stores information about a field to record from. More...

Constructor & Destructor Documentation

Readout::~Readout ( void )

Destructor: Frees the memory allocated by the readout.
Frees the memory.

Member Function Documentation

int Readout::advance ( void ) [virtual]

Advances the readout.
Advances the readout
Implements Advancable.

int Readout::appendPreprocessor ( Preprocessor * pp )

Appends a new preprocessor at the end of the preprocessor list. The input will be processed by all preprocessors, in the sequence in which they are stored in the preprocessor list.
Parameters:

pp The new preprocessor.

Returns:
-1 if an error occured, 1 for success.

double * Readout::exportAlgorithm ( int * rep_length )

Exports a parameterized version of the current learning algorithm.
Parameters:

rep_length Length of the representation vector.

Returns:
A representation of the learning algorithm parameters.

double * Readout::exportPreprocessor ( int ind,

int * rep_length

)

Exports a parameterized version of one preprocessor.
Parameters:

ind Index of the preprocessor in the preprocessor list.

rep_length Length of the representation vector.

Returns:
A representation of the preprocessor parameters.

string Readout::getAlgorithmFormatDescription ( )

Returns a string description of the input/output format of the current learning algorithm.
Returns:
A string describing the input/output format of the current learning algorithm.

int Readout::getNumberPreprocessors ( )

Returns the length of the preprocessor list.
Returns:
The length of the preprocessor list.

string Readout::getPreprocessorFormatDescription ( int ind )

Returns a string description of the input/output format of one preprocessor.
Parameters:

ind Index of the preprocessor in the preprocessor list.

Returns:
A string describing the input/output format of the preprocessor.

void Readout::getPreprocessorIDs ( uint32 * idx )

Get the Ids of the Preprocessors.
Get the Ids of the Preprocessors.
Parameters:

idx Pointer to the memory where to store the Ids.

int Readout::importAlgorithm ( double * rep,

int rep_length

)

Imports a parameterized version of the current learning algorithm.
Parameters:

rep Representation of the learning algorithm parameters.

rep_length Length of the representation vector.

Returns:
-1 if an error occured, 1 for success.

int Readout::importPreprocessor ( int ind,

double * rep,

int rep_length

)

Imports a parameterized version of one preprocessor.
Parameters:

ind Index of the preprocessor in the preprocessor list.

rep Representation of the preprocessor parameters.

rep_length Length of the representation vector.

Returns:
-1 if an error occured, 1 for success.

int Readout::insertPreprocessor ( Preprocessor * pp,

int position

)

Inserts a new preprocessor at the specified position within the preprocessor list. The input will be processed by all preprocessors, in the sequence in which they are stored in the preprocessor list.
Parameters:

pp The new preprocessor.

position The position at which the new preprocessor should be inserted.

Returns:
-1 if the position is invalid, 1 for success.

int Readout::prependPreprocessor ( Preprocessor * pp )

Adds a new preprocessor at the beginning of the preprocessor list. The input will be processed by all preprocessors, in the sequence in which they are stored in the preprocessor list.
Parameters:

pp The new preprocessor.

Returns:
-1 if an error occured, 1 for success.

void Readout::reset ( ) [virtual]

Resets the state of the readout and all filters, preprocessors and algorithms.
Implements Advancable.

void Readout::setAnalogFilter ( AnalogFilter * af )

Sets a new filter for analog inputs.
Parameters:

af Pointer to new analog filter. If NULL, analog signals will not be filtered.

int Readout::setLearningAlgorithm ( Algorithm * learner )

Sets a new learning algorithm. This algorithm computes the output of the readout.
Parameters:

learner A pointer to the algorithm object.

Returns:
-1 if an error occured, 1 for success.

void Readout::setRange ( double lower_bound,

double upper_bound

)

Set the range of output values.
Parameters:

lower_bound Lower bound for output values.

upper_bound Upper bound for output values.

void Readout::setSpikeFilter ( SpikeFilter * sf )

Sets a new filter for spiking inputs.
Parameters:

sf Pointer to new spike-filter. If NULL, spiking signals will not be filtered.

last modified 07/10/2006