Deep Learn­ing: Us­ing Al­go­rithms to Make Ma­chines Think

Deep learn­ing is part of the broader fam­ily of ma­chine learn­ing meth­ods. It was in­tro­duced with the ob­jec­tive of mov­ing ma­chine learn­ing closer to its main goal—that of ar­ti­fi­cial in­tel­li­gence.

OpenSource For You - - Contents -

The hu­man brain has evolved over many, many years and is one of our most im­por­tant organs. The brain per­ceives ev­ery smell, taste, touch, sound and sight. Many de­ci­sions are taken by the brain ev­ery nano sec­ond, with­out our knowl­edge.

Hav­ing evolved over sev­eral thou­sands of years, the hu­man brain has be­come a very so­phis­ti­cated, com­plex and in­tel­li­gent ma­chine. What was not pos­si­ble even as a dream dur­ing the 18th cen­tury and the be­gin­ning of the 19th cen­tury has be­come child’s play now in terms of tech­nol­ogy. Many adult brains can recog­nise mul­ti­ple com­plex sit­u­a­tions and take de­ci­sions very, very fast be­cause of this evo­lu­tion. The brain learns new things very fast now and takes de­ci­sions quickly, com­pared to those taken a few decades ago.

A hu­man now has ac­cess to vast amounts of in­for­ma­tion and pro­cesses a huge amount of data, day af­ter day, and is able to digest all of it very quickly.

Our brain is made up of ap­prox­i­mately 100 bil­lion nerve cells, called neu­rons, which have the amaz­ing abil­ity to gather and trans­mit elec­tro­chem­i­cal sig­nals. We can think of them as the gates and wires in a com­puter. Each of our ex­pe­ri­ences, senses and var­i­ous nor­mal func­tions trig­ger a lot of neu­ron based re­ac­tions/com­mu­ni­ca­tions. Fig­ure 1 shows the parts of a ba­sic neu­ron.

The hu­man brain and its neu­ral net­work have been the sub­ject of ex­ten­sive re­search for the last sev­eral years, lead­ing to the de­vel­op­ment of AI and ma­chine learn­ing tech­nolo­gies. The decade-long dream of build­ing in­tel­li­gent ma­chines with brains like ours has fi­nally ma­te­ri­alised. Many com­plex prob­lems can be now solved us­ing deep learn­ing tech­niques and al­go­rithms. The sim­u­la­tion of hu­man brain-like ac­tiv­i­ties is be­com­ing more plau­si­ble ev­ery mo­ment.

How dif­fer­ent is deep learn­ing com­pared to ma­chine learn­ing

Ma­chine learn­ing was de­fined by Arthur Sa­muel as, “The

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.