Stars

    When Isaac Newton wrote about the Universal Gravitation Law, he step in the shoulders of giants who looked at the stars. Somehow, he was able to find a pattern there that explained how the attraction of two bodies in space occurs, an interaction law to describe a behavior, even though he failed to explain why. Curiously, the laws of gravitation, electrostatic attraction (Coulomb), and magnetic pole attraction force have the same shape, the product of two properties with a constant divided by the squared distance. Often I wonder if the later ones could be understood without the model of the first one.

    In that regard I wonder what could have happened if instead of considering celestial bodies, scientist had focused their efforts in small objects here on Earth. Rocks, boulders, iron balls, you name it. Could it have been possible to propose the object gravitational attraction model with such small objects? Sometimes I think that this kind of discoveries were made because we are able to look at the stars.

    How is all that related with the mind and the cognitive process of language? Well, maybe it is not at all! However, sometimes I think that all those research on neurons, ANN, and brains (artificial or living, matters the same) to solve the AI problem is just trying to get gravitation law out of watching rocks and boulders (no offense to geologists). What if understanding how cognitive processes work requires going macro and study society instead of isolated individuals or just neurons. In the end, the only reason mankind evolved intelligence is because humans needed to communicate among them, because humans are social beings. Despite all bright ideas are born by introspection in the sacred cloister of a privileged mind, that mind requires previous interaction (namely, communication) with other information sources, and all discovery is futile if it is not transmitted. Therefore, interaction, communication, and isolation seem to be necessary for intelligence to emerge, and such patterns, those interaction laws between sentient bodies can only be observed in large systems.

    Maybe we need to take our eyes out of the screen and look again to the stars.

A forgotten goal

 The objective of Artificial Intelligence, as its name suggest, is to create another intelligence –not necessarily human—of artificial nature, one comparable with the one of its creator in order to help them to solve problems and find the answers evading the second one, either by means of the collaboration between two conscious beings, or by the study and analysis of the creator in an attempt to explain their imitating creation. This objective and the idea itself of creating beings endowed with reason is not new, nor a trend or an invention of the last three centuries, but it seems to have accompanied mankind since ancient times. The man has created gods that over time cased to be mere incomprehensible and mindless manifestations of natural forces to become beings at least as intelligent as their creators. These goods, either by whim, chance, or fate, create other beings in their image and likeness, laying this likeness not in the shape or substance, but in spirit (in the etymological sense of the word), since it endow of intellect. Being these creations machines like Talos, the giant bronze titanic automaton forged by Hephaestus (or Daedalus, regarding the read version), intelligent apes, aliens, or humans; the pattern seems to be destined to repeat itself, as in the case of the Archytas of Tarentum’s dove or the so-called Descartes’ automaton-daughter, in an cuasi-cyclical process.

Despite the ultimate goal consists in imitating the human cognitive process and give consciousness to a machine by means of ingenious devices, the task of provide a standard definition for intelligence has been left aside, making then necessary to define it. Defining intelligence might be, more than a philosophical thesis, the work of a lifetime, reason for which the author will not delve deeply about the consciousness’ mater or the ontology of the spirit. From a scientific point of view, Artificial Intelligence privileges the study of problem solving by means of rational processes, ignoring often other aspects of the human intelligence such as sensibility and artistic abilities. This often leads to think on determining intelligence as the problem solving ability, a skill developed by humans during their first years of life.

As part of its development and starting from their early childhood, humans play games. These games involve more than just songs and choreographies to exercise memory and psychomotor coordination, there are also mental skill games requiring puzzle-solving and imagine impossible situations for problem solving. Many of these problems are closely related to the identification and generation of pattern, to such degree that it seems the human mind finds pleasure in stuff that follow known patterns, such as the sound of consonant rhyming verses, the contrast changes in garden’s colors, the sound configuration and distribution or a symphony or musical piece, or when modeling a system under concrete physical laws with precise mathematical equations, just to name some examples. It might seem, then, that intelligence is related with the patterns mind is able to generate and recognize, establishing a directly proportional relation between intelligence and the complexity of the enjoyable pattern.

However, a machine that compulsorily analyzes and solves problems just because it was programmed for it doesn’t give the sensation of intelligence, even though its maker would be very happy. To strength this argument, suppose the one who solves the problems is a human being, but only those problems that are given to them to solve and not those of their own interest, having no other memory but the experience given by the enigmas previously solved. This human has no dreams or aspirations, not even think about questioning why those problem are being given to them. The reader hardly could consider such a person as a smart one, regardless how difficult are the problems that person-machine solve, not questioning automatically disqualify them for solving another set of problems. Being of such importance the rational thinking for mother science and the one Artificial Intelligence tries to mimic, it is almost intuitive to recall the famous cite cogito ergo sum (I think, therefore I am) of René Descartes, father of rationalism and who set the epistemological basis of rational thinking. Descartes center the methodic doubt at the core of his thinking: intelligent beings must be able to doubt, to question the certitude of the obtained data. Unable to doubt, creations would be mere slaves and not beings endowed with reason. If this doubt will lead the machine to question its own existence, acquiring consciousness in consequence, is beyond the scope of this dissertation that focuses in trying to understand and mimic the language abilities in the human’s cognitive process: one of the several areas in the Artificial Intelligence’s research field.

Originally, Artificial Intelligence encompassed the modeling of the human brain, and the cognitive functions of the human being at several abstraction layers. However, modern science has render incapable of propose replicable models that explain how the human mind works, forcing scientist to focus in solving problems increasingly more specific, causing the field to split and specialize in areas that study either specific human cognitive processes, or in optimize and improve the most broadly used techniques of the field, loosing in several cases the perspective and the original goal.

To take robots home, to have machines genuinely speaking with humans, to stop being the only conscious species known in this world, to be outsmarted and exterminated by our creations, it is important to always remember the original goal.

Yet another boring blog in english.

When traveling, the author of this blog use to communicate mostly English and, as long as the other people know some simple sentences, it had never been a problem... with the single exception of London, maybe. The author had several problems there to buy bus tickets, but he's not sure the employee was able to pronounce any other vowel but O.

Although it may be not perfect, the author of this blog feels quite confident on having certain proficiency on it.

However, according with Academia, even though the author is proficient in conversational English, he do know how to write not. To them, all English should follow the "main idea. Explanation." pattern for each written paragraph, in a 4-tuple of {Intoduction, argument, counter-argument, conclusion}, having each sentence no more than 20 words --unless strictly necessary--, and avoiding subordinates as much as possible. That's the way science, the literature for the world's brightest minds, must be written. The author don't see a Ferrary limiting itself to 20km/h or a CAT truck asking to be loaded under half a ton.

At this point the reader probably have guessed the author is not (always) following that schema, nor he (totally) agrees with it. The author finds it as a good guideline but also restrictive and reductionist since not all ideas can be presented that way. To him, the human mind is capable of understanding more than that. In that sense, this blog may prove useful, giving him freedom to explain his ideas without minding those constrains. the point is not to write another boring paper in english, but a (yes, must probably boring) blog in english... with a little bit of humor.

Such explanations should be unnecessary since the author expects none to read this and knows nobody cares about his research. However, the research conducted is related to language understanding. People don't use to follow the rules for writing a nice paper for the JUNICS (Journal of Useless Nonsenses and Incomprehensible Complex Stuff) academic journal in daily life, but the most efficient way to communicate a message. Same must apply to machines, it must be them who adapt to the way humans communicate and not otherwise.

This entry will remain here as a reminder that the author must not get stuck into a writing pattern, but understand language as something that changes over time.