1.
Diagram of the Speech Mind-Module
/^^^^^^^^^\Speech module says auditory phonemes /^^^^^^^^^\
/ visual \ __________ / auditory \
/ memory \ | / Volition \ / memory \
| _______asso-|ciative | \__________/ | ___________ |
| /image \rec-|ognition | ___|_____ | / MindBoot \ |
| / percept \---|---------+ / EnThink \ || memory vault||
| \ engram / |tag c|f \_________/ || ||
| \_______/ | o|i ____|_______ || "students" ||
| | n|b / Indicative \ || "read" ||
| | c|e \____________/ || "books" ||
| | e|r _______|_______ | \___________/ |
| | p| / EnNounPhrase \ | |
| _______ | t| \_______________/ | ________ |
| /fresh \ | ____V_ _|_____ | / \ |
| / image \ | / Psy \ / \ | / aud \ |
| \ engram /---|---/concepts\ / Speech \---|-\ phonemes / |
| \_______/ | \________/ \_________/ | \________/ |
2. Purpose of the Speech Module
The purpose of the Speech module is to output words of natural language as text on-screen or as phonemic sounds.
3. Algorithm of the Speech Module of Strong AI with
NLU
3.A. The Speech module outputs any English or Russian word starting at a given point in auditory memory and continuing until a blank space is encountered at the end of the word in auditory memory.
3.B. So that the AI Mind may "hear itself think", the Speech module sends each phoneme of output back into the AudInput module in a process of re-entry.
4. Code of Speech() from ghost327.pl AI
source code in
Perl
# 2015may28 Requirements of Speech mind-module:
# Speech module shall output a single word starting at given @ear point.
# Outputted word shall be fed internally into the ReEntry module.
# Inflected nouns and verbs shall be spoken from NounGen or VerbGen.
sub Speech() { # http://ai.neocities.org/Speech.html
if ($age > $cns) { die "TERMINATING: memory full \n"; } # 2016jan28
$pho = " "; AudInput(); # 2017-12-07: prime AudInput with a 32=SPACE
$audstop = 0; # 2017-03-31: Initially false value of flag
$pov = 1; # 2017-03-27: only internal POV thinking uses speech module.
$t2s = $aud; # 2016feb13: start from the auditory recall-vector $aud
$rv = 0; # 2017-04-07: [MindForth] Let AudInput designate recall-vector "rv".
do { # PERL by Example (2015), p. 194: do/until Loops
my @aud=split(',',$ear[$t2s]); # 2016jan28: expose @ear values
$output = "$output" . "$aud[0]"; #2016apr09: commenting back in
print "$aud[0]"; # 2016mar29: say each character
$pho = $aud[0]; # 2017-03-27: for reentry into AudInput.
if ((ord $pho) == 0) { # 2017-03-31: if space detected...
$pho = " "; # 2017-03-31: blank SPACE-32
$audstop = 1; # 2017-03-31: flag to end a word with a space
} # 2017-03-31: end of test for a blank space.
if ((ord $pho) == 13) { # 2017-03-31: including any 32=SPACE...
$pov = 2; # 2018-07-01: after 32=CR (carriage return)
print $pho; # 2017-12-21: say or display "pho"; TEST
$lastpho = $pho; # 2017-03-31: to avoid extra "S" on verbs
} # 2017-03-31: end of constantly keeping track of last pho.
if ((ord $pho) == 32 || (ord $pho) == 13) { # 2017-03-31: SPACE or CR
$audstop = 1; # 2017-03-31: flag for finding end of word
$spacegap = 1; # 2017-03-31: identify presence of a space
print $pho; # 2017-03-31: say or display "pho"
} # 2017-03-31: end of test for 32=SPACE or 13=CR.
AudInput(); # 2017-03-27: input each phoneme, not entire idea.
if ($audstop == 1) { # 2017-03-31: flag found for end of word?
if ($spacegap == 1) { # 2017-03-31: space found?
$t++; # 2017-03-31: increment internal AI time "t"
$audrun = 1; # 2017-03-31: resetting at end of internal word.
$pho = " "; # 2017-04-01: send SPACE-31 into AudInput?
AudInput(); # 2017-04-01: one last call to AudInput()
$len = 0; # 2017-06-20: at end of word, reset length $len; 2017-06-20
$t++; # 2017-06-20: increment time-t by one unit? TEST
$spacegap = 0; # 2017-03-31: reset
} # 2017-03-31: end of test for a space.
} # 2017-03-31: end of test for end-of-word flag.
$pov = 1; # 2016apr08: internal point-of-view during thinking.
$t2s++ ; # 2016mar10: Increment time-to-speech to advance thru memory.
if ($aud[0] eq " ") { # 2016apr29: if no character is found
$audrv = $t2s; # 2015may13: Prepare for next word in memory.
if ($pho eq chr(13)) { $pov = 2 } # 2017-03-31: for sake of pause-counter
return; # 2015may13: Exit at end of a word
}; # 2015may13: End of if-clause
} until $t2s eq $cns; # Show the whole array of AI Mind memory.
$phodex = 1; # 2019-10-22: reset for AudBuffer()
} # 2019-10-22: Speech() returns to EnNounPhrase() or other generative module.
5.
Variables for the Speech Module of the AI Mind
$age -- temporary age for loop-counting.
$audrun -- in the AudRecog() module, which increments audrun, it is a counter of how many times the incrementing module has been called. If the audrun count is at unitary one or below two, then the initial character at the start of a word is being processed. audrun is reset to one in the AudInput() and FileInput() modules.
$audrv -- auditory recall-vector for Speech mind-module.
$audstop -- flag to stop Speech module after one word. The Speech module lets a CR-13 carriage-return or a SPACE-32 change the audstop flag from its zero value at the start of Speech to a unitary one (1), so that audstop and spacegap may work together to send the final non-letter into the AudInput module.
$cns -- (adjustable) size of "central nervous system" memory in the First Working AGI.
$len -- length, for avoiding input non-words. The len variable increments with each additional character during AudInput, and is reset to zero at the end of OldConcept and NewConcept.
$lastpho -- device to avoid extra "S" on verbs.
$output -- output string as in JavaScript FirstWorkingAGI.html or ghost.pl AGI.
$pho --
a "phoneme" or character of auditory input.
$phodex -- pho-index counter for AudBuffer(); phodex is reset to unitary one ("1") in the AudInput module and in the Speech module.
$pov -- point-of-view: 1=self; 2=dual; 3=alien. When pov=1, the word "you" is somebody in the external word. When pov=2, the word "you" refers to the self-concept "I" in the AI. When pov=3, the word "you" is interpreted as part of a conversation by a third-party or as a word in a text, not referring to the self-concept of either the AI or of someone talking to the AI.
$spacegap -- a gap to add one post-word space in the Speech module.
$t2s -- auditory text-to-speech index.
6. Troubleshooting of Speech Module with
NLU
6.1.a. Symptom: (Something goes wrong.)
6.1.b. Solution: (AI Mind
Maintainer devises solution.)
7. Future Development of Speech Module with
NLU
The Speech module needs to be converted from displaying text on-screen to outputting phonemic speech sounds.
Roadmap to Artificial Intelligence
8. Resources for Speech Module with
NLU
9.
AiTree of Mind-Modules for
Natural Language Understanding