‘Teach univer­sity stu­dents what ro­bots can­not do’

Times of Oman - - ROUND-UP -

MUS­CAT: With all the talk about cur­rent and fu­ture ap­pli­ca­tions of Ar­ti­fi­cial In­tel­li­gence (AI), one can­not help but won­ders what a univer­sity cur­ricu­lum would look like in the fu­ture. More specif­i­cally, what is the skills that univer­sity grad­u­ates need in or­der not to worry about some smart ro­bot tak­ing over their jobs at some point in their ca­reer?

The sci­ence of AI was cre­ated a long time ago; the term it­self was coined at a work­shop in the USA in 1956. How­ever, it is the field of ma­chine learn­ing and the ad­vance­ment in com­puter hard­ware that have brought AI to the level we are wit­ness­ing to­day. To put this into per­spec­tive, the word ma­chine in this con­text is equiv­a­lent to the no­tion of an al­go­rithm or a com­puter pro­gram.So, when we talk about ma­chine learn­ing, we are re­fer­ring to ma­chines driven by com­puter pro­grams ca­pa­ble of us­ing knowl­edge ac­cu­mu­lated from pre­vi­ous runs to han­dle new in­puts, ex­actly the same way hu­mans use pre­vi­ous ex­pe­ri­ences to han­dle new sit­u­a­tions.

The fa­mous AI project from IBM in the 1990s, named Deep Blue, was suc­cess­ful in beat­ing a num­ber of world-class masters of chess. This was pos­si­ble be­cause Deep Blue was taught how to make the (al­most) op­ti­mal move for each pos­si­ble sce­nario that may come up on a chess­board. Though very im­pres­sive, Deep Blue was not a learn­ing al­go­rithm be­cause it was not ca­pa­ble of learn­ing new moves on its own that were not part of its train­ing reper­toire.

An ANN is a com­pu­ta­tional model that tries to mimic the way the hu­man brain works, and it is only re­cently that ad­vance­ment in com­puter hard­ware has made it pos­si­ble to im­ple­ment such a model, where learn­ing is done by form­ing as­so­ci­a­tions, and not by per­form­ing com­pu­ta­tions, ex­actly the same way the hu­man brain works.

It is easy to un­der­stand why tech­nol­ogy com­pa­nies like Google, In­tel, Ap­ple, Ama­zon, and Or­a­cle may want to grab AI star­tups (in 2018, some of these ac­qui­si­tions had a price tag of over $1 bil­lion), but what is hap­pen­ing right now is that al­most all big cor­po­ra­tions across al­most every tra­di­tional in­dus­try are also in a rush to in­te­grate AI into their prod­ucts. So, what does that mean for univer­sity stu­dents and pro­fes­sors? What are the skills that univer­sity grad­u­ates need in or­der to suc­ceed in the fu­ture? The an­swer to this ques­tion may be found in the work of the Amer­i­can psy­chol­o­gist Howard Gard­ner on mul­ti­ple in­tel­li­gences. Ac­cord­ing to Gard­ner, in­tel­li­gence is not bound to log­i­cal and math­e­mat­i­cal in­tel­li­gence, which is the type of in­tel­li­gence that AI is very good at and will con­tinue to get bet­ter each day. Jobs that re­quire this type of in­tel­lec­tual in­tel­li­gence are tar­gets for AI ap­pli­ca­tions and most likely will be taken over by ma­chines in the fu­ture. The good news is that Gard­ner has also iden­ti­fied other types of in­tel­li­gence that hu­man be­ings can be very good at, while ma­chines are still lag­ging be­hind se­ri­ously and may never get closer to the abil­ity of a hu­man be­ing, at least not in the fore­seen fu­ture.

One ex­am­ple is the ver­ballinguis­tic in­tel­li­gence, which brings some peo­ple the gift of be­com­ing cre­ative writ­ers and very good at learn­ing other lan­guages. Al­though the AI area of Nat­u­ral Lan­guage Pro­cess­ing has made sig­nif­i­cant progress in re­cent years, we think it will be a long way be­fore ma­chines can write nov­els and po­ems and learn lan­guages on their own, the way we hu­mans do. An­other type of in­tel­li­gence is the in­ter­per­sonal in­tel­li­gence, which gives some of us the abil­ity to man­age and lead other peo­ple ef­fec­tively. On this front, the area of AI that at­tempts to recog­nise hu­man emo­tions is still in its in­fancy. A third ex­am­ple is an in­trap­er­sonal in­tel­li­gence, which makes some peo­ple more aware of their strengths and weak­nesses than oth­ers.

We in­vite univer­sity pro­fes­sors to help stu­dents suc­ceed in the fu­ture by de­sign­ing lessons that draw on all kinds of in­tel­li­gence. This way, each stu­dent will get the chance to dis­cover his or her own set of in­tel­lec­tual abil­i­ties, which will dif­fer from one stu­dent to an­other. Only then, we will be pre­par­ing our stu­dents for the fu­ture age of AI. We re­call a 15th-cen­tury motto, which states that the virtue of a teacher is to help stu­dents un­cover their dif­fer­ent abil­i­ties.

Fouad Che­did, the writer, is cur­rently the Deputy Vice Chan­cel­lor for Aca­demic Af­fairs at A’Shar­qiyah Univer­sity in the Sul­tanate of Oman. He has taught at sev­eral univer­si­ties in the USA, Ja­pan, Le­banon, and Oman.

Newspapers in English

Newspapers from Oman

© PressReader. All rights reserved.