yayy I investigated some neural nets and some other alternatives like spacy etc, and then, decided to use both spacy and some neural nets.
hmm
I need to get expert in typed dependencies output of nn and also spacy.
I think I need to test all these with ambigious sentences to see how the results are, I dont get why there is no probabilistic multiple results returns for dep parsing.
hmm. I were thinking also adding some graph based probabilistic sentence decomposition also for dependency graphs. I think I can accomplish that still I mean probabilistic interpretation already needs to be present in knowledge graph.
so if i continue studying today, I would need to study to dependency graph tutorials for types there etc.
hmm also then I need to translate those known typed dependencies to entire sage algebraic representations. including prolog support systems at yet not representable Part of Speech and PTB tokens/definitions.
hmm seems as there is much to do:D
but ok. there is still 2 weeks to complete this tasks. hmm.
i think first week i need to accomplish these tasks of dep graph to algebraic representation in knowledge graph db and then later I need to also develop some ontology signature support system in interpretation. hmm i think that might not be ready in this initial 2 week iteration. I think excluding that ontological signature support system to interpretation, I think possibly the conversion to knowledge graph based definition sets and also generation of sage side inferencing could be readied (With also prolog support). Lets focus these first. Later focus the ontological db's support for ambiguity or correct parsing.
ok then 0.1 version wont include the ontology db supoprt system. lets add such systems in 0.2 yepp.
now the query is:
should I rest now :D or should I continue studying today :)
not that much I accomplished today other than assembling deciding which toolsets to utilize.
hmm I think I can not decide yet I think I fell asleep whilst listening a Star Trek novel yesterday. might first listen it then decide whether I continue studying or not afterwards.
hmm this day time working to data engineering tasks and evening time working to ML project's coding tasks is really challenging imho. I wish it were holiday this week that I could continue studying tomorrow morning. but ML project would resume earliest evening again tomorrow.
hmm so I guess I need to convert the noted dep graph tags to sage wise algebraic representation definitionsets. hmm lets focus on that tomorrow and other day. I think conversion all these to algebraic representation definition set for version 0.1 would at least take tuesday and wednesdays and maybe even thursday evening. then I might install neo4j to write to knowledge graph translated sage algebraic representations with a probabilistic definiton in graph. hmm then I might on weekend focus on wordnet to translate most of literals to sage representations. hmm then I guess I might replan the other week's task when monday comes on deciding which feature to priotize after that. I think next week i might priotize the hmm query input functionalities and hmm ontology data support system.
Yorumlar
Yorum Gönder