Computer Understanding of Human Language

Expressions of Interests in Funding this project - please press the Contact Me button above.

What is it to understand ?

The dictionary (www.dictionary.com) states that understanding is to perceive the meaning of; grasp the idea of; comprehend. Comprehension is a complex process and is somewhat difficult to understand.

So what are the steps between some stimulus and the human brain, well in rough forms here goes:

Sensory - We use our sight, our hearing or any of our senses to translate physical stimulus into electrical signals into the brain
Interpretive - These electrical signals are combined/compared/associated with our past and create interpreted information, for example we see a matrix of black and white and associate that with our learning of language and form language.
Memory - These little snippets of interpreted sensations get aggregated and stored along side similar memories from your past and adding to your future perceptions.
Understanding - It's one thing to interpret a sequence of pixels as a word, sentences and paragraphs. It's another to actually turn those little facts and figures into, well rules, ideas and establish learning patterns for the future. This is the challenge as yet unmet by computational devices.

Well - that was until now.

The algorithms, designs and methods have been established - establishing a process for computational understanding and simplification of language. The uses of this system include:

Word Processing, verification of the meaning of a document - comparing the document with other documents for similar intent and concept.
Security and government - reading a document for content to determine if the content's meaning is illegal or problematic in some way.
Spam - reading e-mail content to determine if the e-mail as an intent which you do not wish to receive.
Advertising - strongly targeting advertising based upon very clear meanings of text, creating conceptual not just textual links to advertising.
Simplification of Human Language : For children, or for adults reading complex information.
How is it done ?

A complex data structure is filled with a significant subset of the English language (the system works on character based, not symbolic based languages) which is unique to the CUHL process. The structure of this database is significantly different to any other technique currently used for computer language processing.

From this point, incoming text is cleaned in typical fashions, then mapped as a whole onto the existing data structure using a recursive algorithm with no specific exit conditions other than 'how accurate you want to get'. In essence, this is the same as a person who gets a rough understanding quickly, and further understanding is gained in time.

The process itself is designed to interwork with existing computer reading and language analysis systems as a complementary technology.

Why isn't it on the market - or - how can you invest?

Well, I need a partner, someone willing to put some money and/or skills into the mix to assist me in creating reality out of this solid design and concept. The following resources are required to turn this into a prototype able to be demonstrated

Chief Linguist : 2 years part time and/or 1 year full time expert Linguist specializing in Morphology, Historical Linguistics and Second Language Acquisition
Assistant Linguist : 2 years part time and/or 1 year full time postgraduate linguist to assist Chief Linguist
Computer Programmer : 2 years part time and/or 1 year full time specialist in data structures, object orient programming prefer Delphi
Seed Funding : US$50,000 for equipment, sourcing companion software etc.

Total estimated investment approximately US$350,000.

What do you get for your money ?

You get (after signing of appropriate non-disclosure agreements and contracts)

a 49.9% permanent share in the business.
shared ownership of all intellectual property and software developed by the project
shared ownership specifically of the design documents and concept itself including any patent jointly applied for.
a solid chance to create the next breakthrough application

I have done a basic cost benefit analysis based upon the software design, expertise of the linguistics team and competition and the following are the possible outcomes (with percentage likelyhood):

30% of extreme success, with the technology fully living up to it's expectations with likely profits within 10 years in excess of US$5 billion
60% of moderate success, with the technology partially living up to it's expectations with likely profits within 10 years in excess of US$5 million
10% of failure, with the technology itself, the implementation or some other issue causing loss of at least 50% of the invested money.

If you are interested, please contact me via the contact button above, to obtain more details on the design and process. A fee of US$2000 will be charged for legal and administrative costs in drawing up contracts, non-disclosure agreements and other documentation. If you just want to have a chat about the concept, feel free to e-mail me with questions.

Christopher Lawrence Burke

CRAZNAR.COM