Press "Enter" to skip to content

Uvodne misli

prva2013Pojmi umetna inteligenca, strojno učenje, nevronske mreže in obsežni podatki pri številnih ljudeh zbujajo velika pričakovanja pa tudi strahove. Malokdo ve, za zakaj pri tem gre. Še manj ljudi zna na tem področju kaj narediti. Zato nas pri vzgoji in izobraževanju tudi v tem smislu čakajo pomembni izzivi.

Nekateri dosežki umetne inteligence so resnično osupljivi. Tako npr. program AlfaZero igra go, šah in shogi bolje kot katerikoli človek ali drug računalniški program. Izhajal je le iz pravil igre in igral s samim seboj. Generiral je ogromne količine podatkov, iz katerih se je učil. Na žalost je to znanje, vsaj za enkrat, pred človekom zaklenjeno v črni škatli. Svojih partij ne zna razložiti na človeku razumljiv način. Spominja nas na odličnega »rokodelca«, ki vrhunsko opravlja svoje delo. Vendar ne zna eksplicitno razložiti, kako dela, kar dela. Za človeka strokovnjaka je značilno, da to kar zna, zna tudi razložiti na ekspliciten simbolični način, kar s pridom uporablja za reševanje drugih problemov.

Zato je pomembno, da umetno inteligenco predstavimo na pravi in uporabni način. Strojno učenje je le eno izmed številnih področij umetne inteligence in nevronske mreže spadajo v to področje. Po mnenju strokovnjakov nevronske mreže uspešno rešujejo 90% problemov razpoznavanja vzorcev, npr. človeških obrazov. Šibkost predstavljata razumevanje in razlaga. Učenci in dijaki lahko tudi praktično spoznajo uporabo nekaterih metod in tehnik umetne inteligence. Tak primer orodja je Orange (https://orange.biolab.si). Za prikaz delovanja globokih nevronskih mrež pri razpoznavanju vzorcev lahko uporabimo katerega izmed demonstracijskih programov na spletu (npr. http://playground.tensorflow.org).

Za zaključek ponovimo, da je pri vzgoji in izobraževanju v informacijski družbi najpomembneje, da vzgajamo kritično misleče strokovnjake, ki imajo znanje in razumejo kako in zakaj delajo to kar delajo. Pri tem nam računalništvo in informatika ponuja lepe možnosti reševanja problemov, od osnovnih konceptov, preko programiranja, do metod umetne inteligence. Za obvladovanje sprememb, ki jih prinaša sodobna digitalna transformacija, igra osrednjo vlogo človek. Morda bolj, kot kdaj koli doslej, tehnologija spreminja nas in odnose med nami.

Uredniški odbor


druga2013Introductory thoughts

Terms, such as artificial intelligence, machine learning, neural networks and big data give rise to major expectations as well as fears in many people. Only a few people understand what this is all about. Even fewer are able to solve problems in these fields. In this context, education is facing important challenges.

Some of the achievements of artificial intelligence are truly amazing. For example, the AlfaZero program plays go, chess and shogi better than any human being or another computer program. It started with basic rules and then played against itself. It generated large amounts of data from which it learned. Unfortunately, this knowledge is locked in a black box and is not available to humans. It cannot explain its games and moves in a human-understandable way. It reminds us of a great “craftsman” who excels at her/his work. But it cannot explicitly explain how it does what it does. The characteristic of a human craftsman is that (s)he can explain what (s)he knows in an explicit symbolic way, which (s)he can use to solve other problems.

Therefore it is important to present artificial intelligence in the right and useful way. Machine learning is just one of many areas of artificial intelligence, and neural networks are part of this field. According to experts, 90% of the problems of pattern recognition are successfully solved by neural networks, for example, recognizing human faces. Its weaknesses are understandability of the models and explanation of results. Pupils and students can practically learn the use of certain methods and techniques of artificial intelligence. An example of such tools is Orange (https://orange.biolab.si). To demonstrate the performance of deep neural networks for pattern recognition we may use online demonstration programs, such as http://playground.tensorflow.org.

To conclude, in education in the information society, the most important thing is to develop critical thinking in professionals who possess the knowledge and understand how and why they do what they do. In doing so, computer science and informatics offer us a great opportunity to solve problems, from basic concepts through programming to methods of artificial intelligence. A human plays a central role in coping with change management brought by modern digital transformation. Perhaps more than ever before, technology is changing us and relationships among us.

Editorial board