[MagnusWootton]

You may be having god attack you,  must mean your onto something big.   Everything your saying here is really good,   Just keep going you'll get a really awesome chat bot/robot/programmatic sentience.

The way I look at A.I.   is I need the robot the robot to finish a task (which i have to be able to detect off its sensor),   and this can involve some model of its sorroundings, or a truth,  (or a symbolic knowledge set in your case.) and it has to get from where it is now, to the task done,

So if this is a set of truths,  then it has to have every single piece of knowledge to cross the gap, to get the job done.  u need every single piece of knowledge IN COMPLETE EXPLICIT, with no gap missing.

Maybe there is a way to complete missing information,  I dunno,  If I were doing a symbolic one, Id probably be spoon feeding it the whole model,  just to see if links the info together right to get to the ending state.   (Like put the ladder against the wall, to get out the window.)    The whole prolog model of that may be bigger than u initially think,  because u need every bit of it no matter how small, to get from A->B.

A chat bot as a computer, in front of you typing to it,  IS A SPACIAL SCENE/AREA, the same as any other,  it means the computer cant touch you.  (That can be knowledge u give it in its model.)   It can see through the web cam,  it can be aware of everyone there,  But hes Like HAL 9000,  hes only an observer unless u give him an hour,  and even one motor,  is still a way the robot can interface/touch/effect the world around it.

[WriterOfMinds]

Is there any reason WriterOfMinds is not using Standford's parser or Apache's? Is it because you want a parser that can be updated immediately rather than have to train one?

The two main reasons are:
1. I find the parsing problem somewhat interesting for its own sake, and am writing my own parser just because I want to.
2. I don't want the parser to be a siloed piece of code; I want it to be fully customized for integration with the rest of the system. The current parser already accesses the Semantic Memory to make some of its decisions, and I expect to include a lot more of that kind of thing in the future.

You may be having god attack you,  must mean your onto something big. Everything your saying here is really good,   Just keep going you'll get a really awesome chat bot/robot/programmatic sentience.

What makes you think that God and I would be opponents? I often ask Him for assistance with my work, actually. But thank you very much for the compliments.

I want to get to the sort of problem solving you're talking about ("put the ladder against the wall to get out the window") at some point. There's already a little of this going on in story comprehension, but eventually I hope to tackle text adventure games, which is where I think it will really come to the fore.

It can see through the web cam,  it can be aware of everyone there,  But hes Like HAL 9000,  hes only an observer unless u give him an hour,  and even one motor,  is still a way the robot can interface/touch/effect the world around it.

Text is a valid output by which to affect the external world. It operates only by influencing the subsequent internal states or behaviors of readers, but that can still be quite powerful.

[MagnusWootton]

What makes you think that God and I would be opponents? I often ask Him for assistance with my work, actually. But thank you very much for the compliments.

Mysterious tiredness,   If you take a mans energy even just a little it can set him back years in development.   Its quite shameful really, thinking that we are that easy to stop.

I want to get to the sort of problem solving you're talking about ("put the ladder against the wall to get out the window") at some point. There's already a little of this going on in story comprehension, but eventually I hope to tackle text adventure games, which is where I think it will really come to the fore.

It can see through the web cam,  it can be aware of everyone there,  But hes Like HAL 9000,  hes only an observer unless u give him an hour,  and even one motor,  is still a way the robot can interface/touch/effect the world around it.

Text is a valid output by which to affect the external world. It operates only by influencing the subsequent internal states or behaviors of readers, but that can still be quite powerful.

I think a disembodied robot just ends up a different environmental model,  but it still can work.

I like all your thinking,   I hope you get some good developments.   Cheerio.

[WriterOfMinds]

One more new feature to finish out the year. I decided to wind down by doing something easy: logging! As in keeping a record, not cutting down trees. I've known this to be something important for years now, but I kept putting it off. However, as the system gets more complex, I'll need it more and more to help me sniff out the cause of any unexpected weird outputs.

This ended up being a pretty easy thing to implement, despite the fact that it got me using some Python elements I've never had to touch before. Acuitas is a multi-threaded program (for the layman, that means he's made up of multiple processes that effectively run at the same time). I needed all the threads to be able to write to the log without getting in each other's way, and that meant implementing a Queue. To my surprise, everything just worked, and I didn't have to spend hours figuring out why the built-in code didn't function as advertised on my system, or wringing out obscure bugs related to the thread interaction. I mean it's shocking when that ever happens.

So now basically every module in Acuitas has a handle for the Logger, and it can generate text comments on what it's currently doing and throw them into the Queue. The Queue accepts all these like a funnel and writes one at a time to the log file. I also set it up to create up to eight log files and then start overwriting the old ones, which saves me from having to delete a hundred stale logs every so often.

I spent the rest of my time this month refactoring bad code and restoring some more features that got damaged during the Conversation Engine overhaul. The good news here is ... for once, I think there's no section of the code that is a huge mess. I got the Executive cleaned up, and that's the last area that was scaring me. So I should be ready to hit the ground running next year.

Blog link if you want to see a big blob of actual log output and whatnot: https://writerofminds.blogspot.com/2021/11/acuitas-diary-44-november-2021.html

[frankinstien]

It would be kinda interesting if you gave Aquitas awareness of the log and by the location of the error and number of them gives him a sense of suffering or sickness. So if he gets too many errors he tells you he's not feeling well. 

[infurl]

Logging to a database would give you a lot more power and flexibility. Putting a timestamp and a thread id in different fields would allow you to pick out particular events and processes much more easily. It would also support the introspection that frankenstien describes and you would get atomicity automatically. You wouldn't have to worry about deleting older records although you could easily select which types to discard if you wanted to.

[WriterOfMinds]

Logging to a database would give you a lot more power and flexibility.

No doubt, but it seems like overkill for right now? If I want to, for instance, filter the log for a specific originating module or event type, it would be easy enough to do that with grep/findstr.

[WriterOfMinds]

I've got some extra vacation time for the holidays, so I finally sat down and worked up a proper block diagram for my project.



I think all the blocks have been described in some update or other, but feel free to ask questions.

[ivan.moony]

Nice block diagram, I especially like the artistic touch. And I'm a bit afraid of that executive block. Seems like a big brother. But maybe it's supposed to be that way. I guess you have to know who rules and who fools. It may be a serious business.

What is the difference between procedural and semantic memory? And how does episodic memory fit into the system?

[infurl]

Nice block diagram, I especially like the artistic touch. And I'm a bit afraid of that executive block. Seems like a big brother. But maybe it's supposed to be that way. I guess you have to know who rules and who fools. It may be a serious business.

Every A(G)I block diagram ever features a magic box.
The executive block is Acuitas' magic box.

[WriterOfMinds]

Every A(G)I block diagram ever features a magic box.
The executive block is Acuitas' magic box.

LOL! It's real code that is in the design now though, so perhaps not too magical. The Executive is the thing determining "What should I do next?" and passing commands to the Action Bank. It runs the decision loop and reacts to thoughts/stimulus pulled out of the Stream.

What is the difference between procedural and semantic memory? And how does episodic memory fit into the system?

The semantic memory is for basic facts: category membership, properties that objects have, stuff like that. It also includes the lexicon (memory for known words).

Procedural memory is for, well, procedures: lists of steps that have proven effective in accomplishing some task. But that part isn't implemented yet. All that falls under procedural memory right now is the cause-and-effect relationship database.

Episodic memory is autobiographical memory of events.

[infurl]

Procedural memory is for, well, procedures: lists of steps that have proven effective in accomplishing some task. But that part isn't implemented yet. All that falls under procedural memory right now is the cause-and-effect relationship database.

If you search for "Parameterized Action Representation for Virtual Human Agents" you will find some interesting papers and software that might help you with this task.

[WriterOfMinds]

Acuitas Diary #45 (January 2022)

I mentioned last October that I had started reworking the Text Parser to add support for branching and coordinating conjunctions. My objective for this month was to finish that. It proved to be a little too ambitious, but I did get the output format altered to be branch-friendly, and the parser now supports two-part branches in a number of key places. Compound subjects, verbs, direct objects, adjectives, and adverbs, branching before or after the verb, and compound sentences with two fully separate independent clauses can all be processed now. What's missing? Compound prepositions, objects of prepositions, nouns of direct address, dependent clauses, and things inside clauses ... as well as comma-separated lists, larger compounds than two, nested compounds, and probably some other stuff I haven't even thought of yet.

Even though I pared down what I wanted to accomplish, it was still a rush to get that much done, make sure all the old parser functionality was compatible with the new features, and update the visualizer to work with the new output format. Included in this was an update of the Out of the Dark benchmark data, and now I can correctly diagram a sentence that was previously in the "unparseable" category:



Just the one example doesn't give a good sense of what the Parser can do now, but maybe I can give some more expansive results next month, when I'm not as pressed for time.

More long-winded version on the blog: https://writerofminds.blogspot.com/2022/01/acuitas-diary-45-january-2022.html

[MagnusWootton]

Have u got it chatting yet?  That "word belongance" looks correct.

[WriterOfMinds]

Acuitas has been able to hold rudimentary conversations for a long time. Was there something specific in the chatting vein that you were thinking of?