hmm I think I need to revise py skills to able to code nicely I have not been coding in py for a long while now and I think I need to revise some patterns there.
hmm yayy, I decided on an initial dependency graph to knowledge graph conversion scheme.
of the converted linguistics universe graph node types and the sage wise initial descriptions set. only for 0.0.1 version of ai project. I think after this 0.0.1 version is tomorrow trained to map the sage wise representations to talk with some training in Bert alike toolings. then might have a ml module that is interacted with talk.
but this initial algebraic/group theoretic representations might not be any elaborate yet. its only 0.0.1 ai version. maybe to be have this ai version be ready this weekend or maybe during this weekend. an ai which can interpret paragraph and queried and answered to from knowledge db it generates.
hmm then in tandem i would continue the study of getting expert in group theoretic studying from modules study first then group theory phd introduction book and then also rings/modules and geometry correspondence studies. then I think I might alter its sage representation definition set.
ayy it would be 0.0.1 coder ai version would be ready by this weekedn or during week :)
then I can also initially integrate with some sound module wise integrated module which helps assists coding studying. might even create its as chrome addon so that it assists in data engineering studies.
but this version would not be not 1.0 version, it would not be any wise nor very talented in nlp yet. but its 0.0.1 version. aha maybe i would tie it to a vocal interaction/communication system then i could have an google chroime assistant ai system that I can easily curate scala code examples when I need to search scala code examples. e.g helping my data engineering tasks during day.
so initial this 0.0.1 version lacks yet coding skills. as mentioned coding skills would be built 4 weeks later I think. or 6 weeks later. I dont know. I first need to finish getting expertise in group theoretic studies etc.
------------------------------------------------------------------------------------------------------------------
hmm since Idont know category theory yet, I think my ontology definitions would be group theoretic wise would be not nicely designed. its expected cause I am not any expert in group theory topics yet.
hmm.
in translating paragraph sentence's dependency graph to its sage wise definition, there: the verb parts represents a homomorphism and each would be stored in an activity named knowledge graph and initially there are also nouns that needs to have ontology wise either multiplication (product) wise might have specific sets of homomorphisms.
hmm then there is an ordered activity. where its an homomorphism with a usual cyclic permutation to activity knowledge graph nodes which is not conceptually actually for a product of multiple groups some internal equivalence class of multiplicated (product) wise groups (actually). so it can not be called homomorphism actually but index of activity groups. but inside activity groups there is surely homomorphisms of time based cyclic permutation of being homomorphism's left side and right side another group's generator being iterated along the time's cyclic permutation operator iterations.
so there could be also hierarchy of activities that again ordered activity could be again link to another ordered activity defs.
hmm I think I would initially define like this with this activities concept.
hmm iniitally very basic group theoretic constructs. I dont even know category theory, if i knew then it might be better designed. but does not matter, first would design like this but then later could design better when i get expert in group theory topics.
and nouns side and their ontology trees e.g. hypernyms or hyponyms etc those stuff like: e.g. actually what matters in this description is any ontology levels,. homomorphism sets. a.g. there is a H named homomorphism that works on AxB then e.g. one ontology levels one supported homomorphims being H means: --> the hyponyms of A ontology level would also have H homomorphism.
so the ontology iniitally would only work in this aspect of knowledge and definition sets of products of ontology levels' as being subject or object of homomorphisms (which is AxBxC 's e.g. A ) and such definition sets and the ontology hiearchy being defined in this homomorphism distributions.
I know this is initially a not a well designed eloquent representation but I am not expert in group theory nor I can currently defiune with category theory like languages. what is important that these could be translated to other representation either later. I can only define such representations yet being not any expert in group theory.
hmm first thing I would do is take notes of the design then move to task of hmm representing word.net in knowledge graph including category ontology information. then also for verbs starting building algebraic representations for their homomorphism definitions. and for all nouns etc defining the homomorphism constraints as above, e.g. whether e.g. do paper talk? e.g. talk is a thing that sapiens does and I mean there wont be any homomorphism for talk that has the object being paper right? I mean it would be of having known homomorphisms definition sets for with considering the ontology classes to construct both from wordnet and later extend with wikipedia ontology db data.
I think I would first take some design notes regarding translation to sage now and then: would start creating ontology graphs from wordnet and starting to construct homomorphism definitions from verbs and then also constructing known homomorphisms for regarding ontology node multiplications(products).
so this latter part would create the common sense. e.g. when a sentence is tried to be interpreted, when there is ambiguity e.g. this homomorphism def set would be a supprot system e.g. would eliminate incorrect representations e.g. we know paper does not talk :) I mean the common sense knowledge of universe.
so I might have multiple neural networks analyzing sentences and their multiple or singular interpretations would be supported with such support system. but of course context is important. paper can also talk if its inside tv show. I mean for ambiguous interpretations: context is important. I mean e.g. if its a movie paper can talk. so this general knowledge that common sense knowledge should be also present. e.g. since universe rules could change in movies. e.g. paper walking. I mean they could animate a paper to talk or walk in a movie. so its rather fiction. so the context should also include common knowledge e.g. movie scripts context has fictional diverging from usual universe contexts e.g. an animated paper that walks etc.
so this common sense support system for the ai is very important, or mostly frequent homomorphism sets, in various contexts etc.
e.g. if ai can not have such common sense support system and reads scripts of a movie then thinks papers could talk, then :D the person evaluating the ai might think ai is crazy :DDDD
I mean :D contextual support systems via such frequent or defined homomorphisms sets is very important.
or. in movies they usually give sapiens behaviours to inanimate entities e.g. to a paper lets hypothesize like this -->
then there is a homomorphism happening with animated paper's group theoretic set entities and humans specific group set entities.
hmm this kind of topics should be neatly designed otherwise ai might be thought insane :D i mean if ai thinks papers can talk :D
this fiction mechanisms invention of sapiens story telling skills: this definitely should be also included as logic common sense support system to ai modules.
----------------------------------------------------------------------------------------------------------------
Ayy so funny I sometimes forget a song is looping whilst i were intending to listen album and clicked repeat thinking it would repeat in song whilst I were trying to make songs of album repeat and then I am lost in design tasks so I dont even perceive same song is repeating :D funny :D I were trying to repeat songs of album of each song consecutively but have not even perceived a lot time same song is repeating instead since I were very focused to this design tasks of ml and ontology etc etc.
-------------------------------------------------------------------------------------------------------------------
now lets take notes of group theoretic definition sets for translations of sentence dependency graphs to sage wise definition sets via knowledge graph nodes (neo4j).
then lets start studying analyzing the ontology information from word net becoming expert or learning more how to use the cypher language.
----------------------------------------------------------------------------------------------------------------------------
Yorumlar
Yorum Gönder