jan29ruai

Artificial Intelligence in Russian

1. Sun.29.JAN.2012 -- Verbs Without Direct Objects

Today in the Dushka Russian AI we begin to address a problem that occurs also in our English AI Mind. Sometimes a verb does not need an object, but the AI needlessly says "ОШИБКА" for "ERROR" after the verb. We need to make it possible for a verb to be used by itself, without either a direct object or a predicate nominative. One way to achieve this goal might be to use the jux flag in the Psi conceptual array to set a flag indicating that the particular instance of the verb needs no object.

We have previously used the "jux" flag mainly to indicate the negation of a verb. If we also use "jux" with a special number to indicate that no object is required, we may have a problem when we wish to indicate both that a verb is negated and that it does not need an object, as in English if we were to say, "He does not play."

One way to get double duty out of the "jux" flag might be to continue using it for negation by inserting the English or Russian concept-number for "NOT" as the value in the "jux" slot, but to make the same value negative to indicate that the verb shall both be negated and shall lack an object, as in, "He does not resemble...."

During user input, we could have a default "jux" setting of minus-one ("-1") that would almost always get an override as soon as a noun or pronoun comes in to be the direct object or the predicate nominative. If the user enters a sentence like "He swims daily" without a direct object, the "jux" flag would remain at minus-one and the idea would be archived as not needing a direct object.

2. Sun.29.JAN.2012 -- Using Parameters to Find Objects

While we work further on the problem of verbs without objects, we should implement the use of parameters in object-selection. First we have a problem where the AI assigns activation-levels to a three-word input in ascending order: 23 28 26. These levels cause the problem that the AI turns the direct object into a subject, typically with an erroneous sentence as a result.
In RuParser, let us see what happens when we comment out a line of code that pays attention to the "ordo" word-ordervariable. Hmm, we get an even more pronounced separation: 20 25 30.

Here we have a sudden idea: We may need to run incoming pronouns through the AudBuffer and the OutBuffer in order unequivocally to assign "dba" tags to them. When we were using separate "audpsi" concept-numbers to recognize different forms of the same pronoun, the software could pinpoint the case of a form. We no longer want different concept-numbers for the same pronoun, because we want parameters like "dba" and "snu" to be able to retrieve correct forms as needed. Using the OutBuffer might give us back the unmistakeable recognition of pronoun forms, but it might also slow down the AI program.

Before we got the idea about using OutBuffer for incoming pronouns, in the OldConcept module we were having some success in testing for "seqneed" and "pos" to set the "dba" at "4=acc" for incoming direct objects. Then we rather riskily tried setting a default "dba" of one for "1=nom" in the same place, so that other tests could change the "dba" as needed. However, we may obtain greater accuracy if we use the OutBuffer.

3. Mon.30.JAN.2012 -- Removing Engram-Gaps From Verbs

Yesterday in the Russian AI we experimented rather drastically with using the "ordo" counter to cause words of input to receive levels of activation on a descending slope, so that the AI would be inclined to generate a sentence of response starting with the same subject as the input. We discovered that the original JavaScript AI in English was not properly keeping track of the "ordo" values, so we made the simple but drastic change of incrementing "ordo" only within OldConcept and NewConcept, both of which are modules where an incoming word must go through the one or the other.

Today we have sidetracked into correcting a problem in the VerbGen module. After input with a fictitious verb, VerbGen was generating a different form of the made-up verb in response, but calls to ReEntry were inserting blank aud-engrams between the verb-stem and the new inflection in the auditory channel. By using if (pho != "") ReEntry() to conditionalize the call to ReEntry for OutBuffer positions b14, b15 and b16, we made VerbGen stop inserting blank auditory engrams. However, there was still a problem, because the AI was making up a new form of the fictitious verb but not recognizing it or assigning a concept-number to it as part of the ReEntry process.

jan26ruai

Artificial Intelligence in Russian

Thurs.26.JAN.2012 -- Insufficient Activation of Subjects

The most glaring problem in the Dushka Russian AI right now is that the AI does not fully activate the subject-pronoun when we type in a short sentence of subject, verb and object. Without a proper subject to provide parameters, the AI fails to select or generate a proper Russian verb-form.

When we type in "люди знают нас" ("People know us"), as an answer we get "ВАМ ЗНАЮТ ТЕБЯ" -- a mishmash of "to you" "they know" "you". In general, the AI seems to be taking the final object entered as input and trying to convert it into the subject for a response.

Thurs.26.JAN.2012 -- Using the "seqneed" Variable

The Russian AI is not setting a Psi "seq" flag when we enter a Russian word as the subject of a following verb. When we inspect the recent 10nov11A.F MindForth code for clues, we discover that in October of 2011 we made major improvements to the method of assigning "seq" tags. We began using the "seqneed" variable as a way of holding off on assigning a "seq" until either the desired verb or the desired noun/pronoun made itself available. However, apparently in the English JavaScript AI we wrote the "seqneed" code only for needing nouns and not yet for needing a verb. No, we did write the code, but it involved avoiding the English auxiliary verb "do", so we accidentally removed the verb-seqneed code from the RuAi. Let us put most of the code back in, and see what happens. Upshot: Once we put the code back into InStantiate, subjects of verbs once again began having a "seq" reference to the verb. The AI even skipped an adverb that we then inserted as a test.

jan13ruai

These notes record the coding of the Russian AI Mind Dushka in JavaScript for Microsoft Internet Explorer (MSIE).

1. Fri.13.JAN.2012 -- Re-thinking Word Recognition

For artificial intelligence in Russian we need to re-think the whole idea of word-recognition as previously implemented in our English AI Minds. In English we did not worry much about word-endings, but in Russian (or German) we need to recognize a verb-form regardless of the number and person in which it is encountered. Since we are using the OutBuffer mechanism to detect and recognize verb-endings, we would like to use the same mechanism to retroactively insert a provisional audpsi identifier on not just the final phoneme of an auditory word-engram but also on the final stem-phoneme and perhaps on each phoneme of the inflected verb-ending. Then we would like to modify the AudRecog module so that it holds onto the provisional audpsi and declares the recognition of a verb in whatever present-tense form it is encountered.

Now we have run the current AI with an Alert box to tell us what is the value of "audpsi" when a second-person singular verb-ending is detected. With the input of "ЗНАЕШЬ" there was no value given for "audpsi", but for "ДЕЛАЕШЬ" a value of "821" was indicated, because the verb-form in its various permutations is provided in the RuBoot sequence.

2. Sat.14.JAN.2012 -- Enhancing Auditory Input

Yesterday in the AudMem module we had difficulty in waiting for the deposition of an audpsi ultimate-tag and in trying retroactively to insert the tag on the penultimate phonemes of the Russian word being recognized. We were obtaining values for audpsi at times when we expected there to not yet be an audpsi.

Although we try zeroing out audpsi at the end of AudMem, it looks as though further use of "audpsi" is required in the AudListen module and in the AudInput module, where finally audpsi is converted to
oldpsi for use in the OldConcept module.

It turns out that AudListen calls AudInput when a space-bar is reached during keyboard entry of a word. The AudInput module, without using AudMem, directly stores an audpsi ultimate-tag retroactively by using the "tult" value. Therefore we should be trying to insert additional "audpsi" tags in AudInput and not in AudMem.

3. Sun.15.JAN.2012 -- Auditory Stem-Tagging

We have gradually learned that the AudInput module will not let us readjust values of audpsi on a word from within an if-clause testing for a value of zero on the "aud4" or ctu continuation-flag. Therefore we may need to introduce a secondary if-clause in order to make each phoneme of the word carry the audpsi tag.

We developed a suspicion that something was not letting a positive audpsi be inserted after any phoneme with an "aud4" continuation-flag "ctu" of one. We searched for "aud4 ==" and in audDamp we found the conditional "if (aud4 == 1) aud5 = 0". This obscure line of code made us spend one or two days of work in trying to comprehend why we could not "backfill" the audpsi value onto phonemes prior to the final phoneme of a word.

When we commented out the offending line in audDamp, we began to notice unwarranted carry-overs of an old audpsi onto the first phoneme of the subsequent word. To correct that problem, audpsi will need to be reset to zero in at least one additional location. Actually, we had to reset "morphpsi" to zero at the end of AudRecog to solve the problem.

4. Sun.15.JAN.2012 -- Russian Verb Stem Recognition

Now in AudRecog we need to set up provisional recognition of Russian verb-stems. We create a "provrec" variable for "provisional recognition" and we use it to detect the early presence of "audpsi" tags before the end of a word is reached. Dushka begins to recognize incoming Russian verbs and to generate incorrect but on-target sentences using the recognized verb in the infinitive form. It remains to use the AudBuffer mechanism and the parameters of person and number to generate the output of a Russian verb in the proper gramatical form.

Table of Contents (TOC)

• Fri.13.JAN.2012 -- Re-thinking Word Recognition

• Sat.14.JAN.2012 -- Enhancing Auditory Input

• Sun.15.JAN.2012 -- Auditory Stem-Tagging

• Sun.15.JAN.2012 -- Russian Verb Stem Recognition

jan12ruai

These notes record the coding of the Russian AI Mind Dushka in JavaScript for Microsoft Internet Explorer (MSIE).

Thurs.12.JAN.2012 -- Parsing Russian Verb-Endings

In our Russian JavaScript AI code heretofore we have merged the English and Russian AI Minds and we have eliminated or deactivated all the code for thinking in English. For the Dushka AI to think properly in Russian, we need to implement the OutBuffer mechanism for dealing with the inflectional endings of Russian verbs and nouns. Since we are not sure where to begin, we will present ourselves with the problem of dealing with the input of a previously unknown Russian verb.

Pressing Alt-Shift to toggle into Russian input, we ran the AI and we typed in the word "ЗНАТЬ", which the AI properly recognized as bootstrap concept #840. The AI responded with an ungrammatical sentence of "Я ЗНАТЬ МЕНЯ".

Then we typed in the word "ЗНАЮ", which the AI failed to recognize as a form of ЗНАТЬ, assigning instead a concept number of 882, as if the item was a brand new word being learned by the AI. We will try setting the value of "nru" to 900 at the end of RuBoot, so that new concepts will be learned with concept numbers starting at #901. Now we typed in "ЗНАЮ" and it was assigned #901 as a concept number. Next we typed in "ЗНАЕШЬ", and it, too, was assigned #901 as a new concept.

If we want the OutBuffer mechanism to recognize a personal verb form as such, we will need to go back to a version of the Russian AI which was sending input into the buffers. On the Packard-Bell desktop computer in the 25dec11A.F MindForth, we used the "abc" transfer-variable in the AudListen module to capture input keystrokes and move the characters into
a buffer. In the JavaScript AI, we will need to use the area of AudListen() where "pho = pho.toUpperCase()" turns each keystroke into an uppercase Cyrillic letter. From there we also call AudBuffer so that the "abc" values are transferred into the buffer.

We should probably not call OutBuffer from the "CR()" carriage-return module, which deals with the moment after an incoming word has gone into the AudMem() module. Instead we should probably deal in AudListen() directly with the input of a space-bar or a carriage-return.

We need a suitable location to reset the "phodex" counter back to zero after the end of a word of input. The "CR()" carriage-return module does not seem to effect the hange promptly enough. Let us try resetting "phodex" in the AudListen module when a carriage-return or a space-bar is entered. That method seems to work well, and somehow the AudBuffer and the OutBuffer apparently get cleared out.

Now we need to choose where the testing for any particular verb-ending in the OutBuffer will take place. It could maybe take place in the AudMem module. No, it turns out that it is somehow too late to test for a "b16" ending in AudMem. It works better if we test for "b16" in AudListen, before the character even goes into AudMem. It also turns out that we can use "if (b16==String.fromCharCode(1070))" as a way to test for an actual Russian character.

In AudListen we have now managed to build up code that tests the final three right-justified spaces in the OutBuffer and recognizes a second-person singular Russian verb-ending during keyboard input. Within the same test-code we have set the "dba" as "2" for second person and the part-of-speech "bias" and "pos" at "8" for a verb. The set values carried over into the memory arrays. Thus we expanded and improved the RuParser function. The same mechanism that recognizes a verb-ending, also parses the word as a verb.