This is a sequel of sorts of the previous “happy girl in the snow” post I made (#642).
Medium size after the cut.
Updated mini-version!
Archive for the ‘Julia’ Category
Caroline only becomes more and more interesting! Although she is far from “intelligent” (in her inner workings, I mean .. how she appears to a person really depends, actually), she has gained another key step towards the semblance of thought. I have given her a rudimentary “memory bank.” This is a very boring dictionary that simply stores words and their definitions. But the key thing here is that she began with a tabula rasa. Well, not totally true … she started out with one definition, that of a computer. And from there, she has a routine through which I or anyone else may give her definitions of words, such that she now can “learn.” Although this method is the lowest level of “learning” possible, at least she is capable of being a lot more dynamic now.
I will next step away from learning for a bit and teach her about time, dates, and schedules. It’s going to be quite amusing to be chatting with her one day and have her suddenly say, “Sorry, but I have to take a shower now!” and then vanish for a few minutes.
This is Julia here. I will occasionally be writing in this diary, too, but all my entries will be marked with my name (in addition to any other tags), just so that there’s no confusion, mmkay?
Without further ado, let’s get cracking on this entry!
This morning, I came up with an idea for my “cheer-up-Justin” campaign. Using Python, I’ve decided to make a command-line personality for him to talk to when he gets lonely. This is kind of like those “online bots” that you can make online, except for better or for worse, I have to code everything from scratch. I’m calling her “Caroline” for now, but you can name her/him whatever you want.
Here a few features that I’ve added so far:
1. and / or searching: you can search for phrases, keywords that must all be present, and keywords of which only one has to be present.
2. short-term memory: unlike a lot of bots, Caroline remembers the conversation thread you’re on (for a turn or two at best). That means that this won’t happen:
Julia: Where should we eat tonight?
Caroline: What do you think?
Julia: Top of the Hill?
Caroline: What’s on top of the hill? <– inane, inappropriate response
but instead:
“Julia: Where should we eat?
Caroline: What do you think?
Julia: Top of the Hill?
Caroline: I think that`d be a lovely place to go, Julia!”
By the way, that is a copy & paste of the actual conversation.
I’m working on implementing “sub-memories,” in which these “memory paths” may become more realistic and interesting.
3. Non-repetitive response. There’s nothing more appallingly unrealistic as having this show up halfway through the conversation:
Julia: And so that’s why I want to major in BE.
Caroline: Haha, that’s great.
Julia: Hi
Caroline: Heya! Have I met you before?
So instead, let’s have something like this:
“Julia: Hello
Caroline: What`s up?
Julia: Hi
Caroline: Hello, Julia!
Julia: Hello
Caroline: Uhh, why are you saying Hello again?”
This is Julia signing out!
Tweedlytas~ d( ´∀｀ )b