Skip to main content

Artificial Intelligence Knowledge Information Framework

Project description

#AIKIF
#####Artificial Intelligence Knowledge Information Framework
*NOTE - this is very much an experimental work in progress* - the code runs, but wont do anything useful at this stage

##Overview
This is an example framework to capture the flow of information initially for personal data management, but ultimately useful for AI applications.<br />
Initially it will be populated and tested for human use, but includes tests and verification process for future ‘General AI’s.<br />
Functions (Octave, Python, SQL) are called at set stages of the AI process which log the results into a standard database schema.<br /><br />


###Quick Start
The goal is to get any set of information and parse it into a consistent format so a machine can read it.<br />
For example:<br />
Project Management<br/>
Code Management<br/>
Personal Information Management<br/>

##Programs
###Main Programs
AI.py - sample main program to show a trivial example of logging data<br />
view.py - simple command driven procedure to show various details of the system<br />
index.py - creates text indexes of all the files<br />
search.py - searches, using both indexes and ontologies<br />
go_web_aikif.bat- starts the web server for the AIKIF admin interface

###Toolbox
Various modules which contain generic functions

###DataTools
collection of modules to manage data transformations

###Standard Library Programs
AIKIF_utils.py - standard utils for the filelists<br />
fileMapping.py - main routine that decides what the output files will be called<br />
security.py - manages security, which will allow users to have private data (not the norm for this)<br />

###Data Load programs
These programs are used to load a specific dataset, the code used to parse each file is in a separate load procedure<br />
processRawData.py - this calls all data load programs and logs results<br />
create_word_lists.py - loads a list of nouns, verbs, adjectives from web into local structures<br />
loadCountry_Gdeltproject.py - loads a country reference file<br />
loadPIM_Filelist.py - loads a list of local files into objects, events, photos<br />
loadPIM_shopping.py - sample to show how a personal shopping list is loaded<br />

###Experimental programs - probably wont be used
addRawData.py - using word lists, this experiments with parsing information as a bag of words<br />
AIKIF_create.py - creates default set of filelists and data files (DONT run this if you start using the software)<br />


##Data
Raw Data - raw information from any source<br />
BIAS tables - weightings to rank data based on various criteria (source, person, format)<br />
Weighted Data - data ranked according to weightings / human verification results<br />
Algorithms - database of algorithms, split into componants <br />
Concepts - generic concepts about information<br />
Concepts_Data - links to concepts and data<br />
New_Concepts - randomly generated possible concept links based on ratings<br />

##Tracking generic AI concepts / Logging
###Goal Management
goal_types - 0=supergoal:{'Be Friendly to humans'}, 1=endgoals ['assist', 'solve', 'learn'], 2=goals [], 3=subgoals []<br />
goals - list of goals to achieve<br />
preferences - ranked order of topics to focus on<br />

###Decision Making
commands - requested commands from human operator<br />
action - list of actions (AI plans) <br />
outcomes - list of possible outcomes with impacts, liklihood, past stats, ratings<br />

###Source Data
-----------
rawData - raw text feed of data from datasets, web, social media<br />
websites - reference file on websites with biases<br />
people - reference file on people / usernames with biases<br />

###Data Processing Tables
----------------------
bias - details on bias's for a given source of rawData<br />
feedback - human reasoning behind various BIAS weightings and human votes (+/-) on rawData<br />
facts - result of processed rawData taking into account sources, biases and feedback<br />
knowledge - understanding of facts *(no idea how this will be implemented)<br />

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

AIKIF-0.0.1.zip (115.3 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page