AR­TI­FI­CIAL IN­TEL­LI­GENCE TAKES OVER PRODUCT DE­SIGN & DE­VEL­OP­MENT

Stitch World - - CONTENTS -

There is no de­bate that tech­nol­ogy has been dis­rupt­ing the ap­parel sup­ply chain, cov­er­ing man­u­fac­tur­ing, sourc­ing, lo­gis­tics and even product de­sign­ing. Brands and re­tail­ers have re­alised the im­por­tance of the use of tech­nol­ogy in product de­sign­ing ow­ing to the ever-chang­ing de­mands of the con­sumer and also be­cause of the need to speed up the time to mar­ket. Based on the re­port– ‘ The Fu­ture of Fash­ion: From De­sign to Mer­chan­dis­ing, How Tech Is Re­shap­ing the In­dus­try’– by CB In­sights, the ar­ti­cle ex­plores the dis­rup­tions in product de­sign­ing.

Un­der­stand­ing cus­tomers and their needs has been of more im­por­tance now than ever be­fore. Fash­ion brands and re­tail­ers across the world are us­ing ar­ti­fi­cial in­tel­li­gence to fig­ure out the de­signs of the fu­ture. Data col­lected from all the plat­forms are em­pow­er­ing brands to utilise th­ese to pre­dict what cus­tomers will want to wear next, with the help of Ar­ti­fi­cial In­tel­li­gence.

Google, the tech gi­ant, has part­nered with Ger­man re­tailer Za­lando in an ex­per­i­men­tal project to ex­plore the po­ten­tial of al­go­rithms in fash­ion de­sign. Dubbed as ‘Project Muze’, the project uses a neu­ral net­work trained with the de­sign pref­er­ences ( colour, tex­ture, and style) of more than 600 fash­ion trend­set­ters along with the fea­tures data from the Google Fash­ion Trend Re­port, in ad­di­tion to styles on Za­lando. The users, with the help of Project Muze, will be able to cre­ate vir­tual 3D fash­ion de­signs on their own.

Ama­zon is another com­pany that is lay­ing its hands on product de­sign by us­ing the con­cept of Ar­ti­fi­cial In­tel­li­gence. The com­pany’s re­search team at Lab126 has de­vel­oped an al­go­rithm that is ca­pa­ble of learn­ing about a par­tic­u­lar fash­ion style us­ing images. Through a cut­ting- edge tool known as Gen­er­a­tive Ad­ver­sar­ial Net­work ( GAN), the al­go­rithm can gen­er­ate unique images in sim­i­lar styles from scratch.

In another de­vel­op­ment, Ama­zon is in the process of de­vel­op­ing a ma­chine learn­ing al­go­rithm that will be able to de­cide whether a par­tic­u­lar ensem­ble can be con­sid­ered as stylish or not. By analysing a few la­bels at­tached to the images, the soft­ware can pro­vide users with feed­back or rec­om­men­da­tions for ad­just­ments.

How­ever, AI pow­ered de­signs need im­prove­ment be­fore brands can rely on th­ese al­go­rithms. The out­comes of Google’s Project Muze were just un­wear­able scrawls and scrib­bles, while some re­ports on the Ama­zon Lab126 ini­tia­tive called the de­sign re­sults ‘crude’. But surely, th­ese are help­ing de­sign­ers make de­signs in a quick span.

With a proac­tive ap­proach on AI product de­sign­ing, even brands are try­ing to ex­plore the mul­ti­ple util­i­ties and pos­si­bil­i­ties. US- based Tommy Hil­figer has joined hands with IBM and the Fash­ion In­sti­tute of Tech­nol­ogy for a project that will use IBM AI. Th­ese AI tools will help in de­ci­pher­ing real-time fash­ion in­dus­try trends, cus­tomer sen­ti­ments around Tommy Hil­figer prod­ucts and run­way images and resur­fac­ing themes in trend­ing pat­terns, sil­hou­ettes, colours, and styles. This in­for­ma­tion can then be used by de­sign­ers to pro­duce the next col­lec­tion.

A start- up in USA, ‘Stitch Fix’ is gain­ing pop­u­lar­ity for its AI­driven per­son­alised en­sem­bles for its cus­tomers. The com­pany uses al­go­rithms to fetch out­fits based on the cus­tomers’ favourite colours, pat­terns, and tex­tiles. Th­ese al­go­rithms iden­tify trends and sug­gest new de­signs as well based on the in­for­ma­tion pro­vided by the cus­tomers. Af­ter the al­go­rithms, hu­man stylists take over to look at the final se­lec­tion and of­fer styling sug­ges­tions ac­cord­ingly.

Not only this, Stitch Fix is also us­ing al­go­rithms to ac­tu­ally de­sign new pieces. So far, the ini­tia­tive, called Hy­brid De­sign, has adopted ma­chine learn­ing to de­velop over 30 pieces us­ing

A start-up in USA, ‘Stitch Fix’ is gain­ing pop­u­lar­ity for its AI-driven per­son­alised en­sem­bles for its cus­tomers. The com­pany uses al­go­rithms to fetch out­fits based on the cus­tomers’ favourite colours, pat­terns, and tex­tiles.

this method­ol­ogy. “We’re uniquely suited to do this,” said Eric Col­son, Chief Al­go­rithms Of­fi­cer at Stitch Fix. “This didn’t ex­ist be­fore be­cause the nec­es­sary data didn’t ex­ist. A Nord­strom doesn’t have this type of data be­cause peo­ple try things on in the fit­ting room, and you don’t know what they didn’t buy or why. We now have this ac­cess to great data and we can do a lot with it.”

There are in­de­pen­dent de­sign plat­forms also avail­able that al­low peo­ple to make their own de­signs. One such platform is CLO that makes it easy to tweak de­signs on the fly. This means brands can al­ready use real-time AI in­sights to mod­ify fash­ion right up to the minute they hit pro­duc­tion.

The next era of fash­ion is all about per­son­al­i­sa­tion and pre­dic­tion. With more and more data, al­go­rithms will be­come trend hunters – pre­dict­ing ( and de­sign­ing) what’s next in ways that have never been pos­si­ble.

True Fit, a data- driven per­son­al­i­sa­tion platform for footwear and ap­parel re­tail­ers, uses big data and ma­chine learn­ing for ex­act-fit cloth­ing and shoe rec­om­men­da­tions. With over 56 mil­lion reg­is­tered users, the platform uses trans­ac­tion data to de­ter­mine cus­tomer pref­er­ences that ‘bet­ter per­son­alise all touch­points of the con­sumer jour­ney’ for brands. In­creas­ingly, con­sumer pref­er­ences will guide ev­ery as­pect of the de­sign and pro­duc­tion process.

The next era of fash­ion is all about per­son­al­i­sa­tion and pre­dic­tion. With more and more data, al­go­rithms will be­come trend hunters – pre­dict­ing (and de­sign­ing)

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.