The new tech­nolo­gies that will com­plete you—and your busi­ness

The cre­ative new tech­nolo­gies that will fill in the blanks for us like never be­fore

Inc. (USA) - - CONTENTS - Amy Webb Amy Webb is an au­thor and fu­tur­ist and the founder of the Fu­ture To­day In­sti­tute, a lead­ing fore­cast­ing and strat­egy firm that re­searches tech­nol­ogy for a global client base. She is the au­thor of The Sig­nals Are Talk­ing: Why To­day’s Fringe Is To

EVERY YEAR, my ex­tended fam­ily gets to­gether for a pic­nic at which we al­ways take a group pic­ture. And every year, the shot is ru­ined by some­thing in the back­ground: a street lamp stick­ing out of my aunt’s head, strangers caught in the flash, cars in­trud­ing into peace­ful out­door set­tings.

But with a few clicks in Pho­to­shop, I can eas­ily re­move that street lamp with­out de­cap­i­tat­ing my aunt. Al­go­rithms an­a­lyze the photo, de­ter­mine how it would look if some ob­ject weren’t there, and then—voilà!—the im­age is fixed.

This trick seems sim­ple. But it’s an ex­am­ple of so­phis­ti­cated ma­chine-learn­ing al­go­rithms at work, and a sig­nal of what’s just over the hori­zon. A slew of new tech­nolo­gies will soon fill in the blanks not just in our pho­tos, but in many other places, too. They will au­to­mat­i­cally tell us who and what we’re see­ing, and they’ll an­tic­i­pate peo­ple’s re­ac­tions in a given sit­u­a­tion. Recog­ni­tion al­go­rithms— which an­a­lyze an im­age of, say, a ten­nis racket and iden­tify it—will even­tu­ally be in­te­grated into the de­vices con­sumers use. For good or ill, they’ll also be built into be­hav­ior-mon­i­tor­ing sys­tems avail­able to busi­nesses.

Within the next 24 months, con­sumers will have rudi­men­tary ac­cess to such tech­nolo­gies. Pin­ter­est re­cently launched Lens, a sort of Shazam for ob­jects; Blip­par, which Inc. wrote about last month, has sim­i­lar ca­pa­bil­i­ties. Sam­sung’s up­com­ing Galaxy S8 will al­low a mo­bile-phone cam­era to be used as a vis­ual search tool, too—and this on-the-fly recog­ni­tion sets the ground­work for where we’re headed to­mor­row. To­day, Face­book, Ap­ple, and Google auto-gen­er­ate movies from con­tent you up­load. In fu­ture it­er­a­tions, com­ple­tion al­go­rithms will pull footage from co-work­ers and so­cial me­dia con­nec­tions, stitch it to­gether with your own, and au­to­mat­i­cally cre­ate videos from events like work re­treats and fam­ily pic­nics. Mean­while, re­searchers at MIT’s Com­puter Science and Ar­ti­fi­cial In­tel­li­gence Lab have de­vel­oped an al­go­rithm that can fore­cast how hu­mans will in­ter­act with one an­other. Cap­ture two peo­ple in a video frame, and it can pre­dict whether they will shake hands, hug, kiss, or ig­nore each other. CSAIL has also de­vel­oped and trained a deep-learn­ing al­go­rithm that can rec­og­nize hu­man ac­tiv­ity well enough to gen­er­ate full videos from sin­gle images. Us­ing a photo of a trainer, a dog, and a jockey rid­ing a horse, a com­puter auto-com­pleted a video of the trainer lead­ing the horse into po­si­tion—with ac­tual peo­ple, trees, grass, and an­i­mals— to con­vinc­ingly de­pict a scene that never hap­pened in the real world. Such de­vel­op­ments im­ply a near-fu­ture full of fas­ci­nat­ing ap­pli­ca­tions, in which smart al­go­rithms pro­duce videos that fore­see, for ex­am­ple, how kids from At­lanta might re­spond to a new potato chip fla­vor, or how older Har­ley riders might re­act to a near-ac­ci­dent.

What all of this points to: Ma­chine vi­sion will soon be pow­er­ful enough to use our past be­hav­ior to pre­dict what we’ll do next, mak­ing it eas­ier to ob­serve crowds at events, mon­i­tor em­ploy­ees at work, and track cus­tomers as they shop.

By then, al­go­rithms will fo­cus not just on what, but on who, as well, by pair­ing our faces to the data we gen­er­ate on so­cial net­works and mo­bile apps. Smart cam­eras will de­tect a cus­tomer who’s likely to en­gage with your com­pany’s sales rep if ap­proached, and then feed that rep key data points about that cus­tomer in real time. The pri­vacy con­cerns such tech­nolo­gies stir may well make us all feel a lit­tle vul­ner­a­ble. But per­haps the even­tual killer app won’t be a nostalgic video maker, but rather a real-time in­for­ma­tion sys­tem that can help us re­late bet­ter to one an­other. Which could ul­ti­mately do more for all of us than fix­ing an an­nual fam­ily photo.

Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.