Current Projects
- IQ Test Solver
- Image recognizer
- General constructor of systems of differential equations
- Data Mining and general prediction
IQ Test Solver
|
The Intelligence Quotient has traditionally been considered the best indicator for scientifically evaluating natural intelligence. Is this indeed the best way to measure this human capacity? Could a machine emulate a human being solving traditional intelligence tests? If so, could we affirm that a machine possesses an intelligence equivalent to that of a human? |
|
KITBIT explores some of these possibilities.
Psychologists agree that various forms of intelligence exist: verbal, logical, numerical reasoning, musical or spatial among others.
If we assume that general intelligence is a combination of logical, numerical and verbal intelligence, a test which evaluates these three abilities should obtain a relatively good indication of intelligence.
KITBIT attempts to emulate human intelligence and its logical and numerical capacities, exactly as they are measured by IQ Tests.
On our web-page, TheIQChallenge.com, we challenge our visitors to put KITBIT to the test in solving numerical and logical problems which have the exact same format as traditional intelligence tests used by psychologists.
Some screenshots from our web games based on traditional IQ tests.
KITBIT actually solves the following test types:
We will progressively add other similar tests to those already mentioned. Tests related to verbal logic remain beyond KITBIT's reach, for the moment.
Image recognizer
Artificial intelligence techniques are increasingly solving the problem of image recognition more precisely. The efficiency of these techniques is acceptable when the sample images have a certain uniformity (for example, in face recognition). The problem becomes much more complex when the sample images are less uniform.
The KITBIT project of image identification in dissimilar image universes attempts to obtain a general algorithm which, in principle, facilitates their classification by establishing categories and hierarchies.
Additionally, the KITBIT model will allow us to work on a sequence of images, observing and predicting the future behavior of the elements which compose the sequence.
General constructor of systems of differential equations
One of the principle tasks which natural intelligence confronts is that of formulating problems from the analysis of their solutions. Traditionally, we call this task "to generalize".
The majority of measurable phenomena are approximate solutions of differential equations with one or several variables. KITBIT attempts to define these differential equations and then use them, in a particular case, to obtain the descriptions of the observed phenomena.
This mechanism of reengineering allows us to obtain the general models or laws which the registered experiences follow.
In areas of well-defined knowledge KITBIT has already obtained some results.
Data Mining and general prediction
KITBIT can act as an organizing agent for databases. It can find affinities among sets of data. It also acts as a prediction agent in temporal series. Due to its ability to learn from experience, its efficiency in these tasks is increasing.
Currently we are investigating and developing applications for spreadsheets and databases which integrate useful functions for analysis.