Now, artificial intelligence powers fashion ecommerce in India

如今，人工智能助力印度時尚電子商務

You see a dress you like on a fashionportal and want to compare it with similar ones to make a choice. It seemssimple. But do you realize how smart your brain has to be to decipher that onedress is similar to another? Color, light, dimensions, depth, gender – yourbrain processes a multitude of inputs to make the call.

當你在一個時尚門戶網站上看到一個喜歡的裙子時，并想與其他相似的裙子比較來做出選擇。這看似很簡單，但是你是否意識到你的大腦需要多么智能，才能夠辨識一個裙子與另一個裙子很相似呢？顏色，光亮，尺寸，深度，性別，你的大腦需要處理大量的信息來做出回應。

Now, imagine trying to get a machineto do that. To come up with intelligent suggestions for dresses that match theone you like, a computer has to convert images into numerical and symbolicdata. This is a complex task involving geometry, probabilities, learningtheory, and so on. Of course, once it learns to do that, it can scan a lot moredresses much quicker than we humans can, so we get better choices.

現在，幻想一下嘗試用機器來做這些事情。利用智能建議來搭配你喜歡的裙子，計算機需要將圖像轉換成數字和符號數據。這是一項包含幾何，概率以及學習理論等的復雜工作。當然，一旦機器學會了這些，與我們人類相比，它就能以更快的速度掃描更多的裙子，這樣我們就能得到更好的選擇。

Taggingto seeing

通過標簽來觀察

Indian fashion portalsVoonikandCraftsvillaare rolling this out today: if you goto these sites and click on a Bollywood saree or terracotta jewelry you like,artificial intelligence will kick in and recommend similar sarees and jewelryto you instantly. (For a couple of weeks, the feature will be in A/B testingmode, which means only half the users can access it.)

印度時尚門戶Voonik以及Craftsvilla今天推出這項服務：如果你訪問這些網站并點擊你喜歡的寶萊塢紗麗或赤土首飾，人工智能就會彈出并立刻為你推薦相似的紗麗或珠寶。（在未來的幾周，這一功能將會處于A/B測試模式，這意味著只有一半的用戶可以進入）

We’ve got used to seeingrecommendations for products when we click on books, apparel, or anything elseonecommercesites. But those are different from visual comparison.“Earlier, we only had recommendations based on tags. We look at the physicalattributes of items (color, sleeve, fabric, etc.) and suggest items which havesimilar attributes. With the new feature, we show items that are visuallysimilar to the item the user is viewing,” explains Sujayath Ali, co-founder andCEO of Voonik.

我們已經習慣于在電子商務網站上點擊書籍，器械或其他東西時看到一些產品的推薦。但是這與視覺比較還有很大的不同。“早期，我們只有基于標簽的推薦。我們看到條目的相關物理屬性（顏色，袖套，面料等），并建議有相似屬性的條目。有了新功能后，我們就能推薦與用戶瀏覽的外觀相似的條目，”Voonik的聯合創始人以及CEO，Sujayath Ali解釋道。

Tags are text-based descriptors whichmay even include visual attributes like color. But they don’t tell you theexact shade of color, pattern, or shape of the dress like computer vision can.That’s why you often find oddball recommended products. After all, describingan object with tags cannot mimic how we see an object in its entirety. Butcomputer vision comes close to seeing an object as the human brain does.

標簽是基于文本的描述符，其中還可能會包括可視化的特征，比如顏色。但是他們并不能像計算機視覺那樣告訴你確切的色度，圖案或者裙子的形狀。這就是為什么你經常會看到一些古怪的推薦的原因。總而言之，用標簽來描述對象并不能完全的模擬我們所看到的物體。但是計算機視覺可以類似于人類大腦一樣觀察物體。

Apart from tags, recommender systemson ecommerce sites rely on user data – such as, what you’ve liked earlier orwhat’s popular among other users. But for fashion portals like Voonik andCraftsvilla, in particular, it makes sense for the recommendations of similarproducts to be based on the visual attributes of something you click.

除了標簽之外，電子商務網站的推薦系統也依賴于用戶數據，比如說，你之前喜歡什么以及在其他用戶之中比較流行什么。但特別是像Voonik以及Craftsvilla這類時尚門戶網站，主要是基于用戶點擊物品的可視化特征，從而做出有效的產品推薦。

Behind this visual recommendationengine is a cool product called MADstack – a cloud-based suite of computervision and AI modules that an ecommerce or any other site can plug into with anAPI(applicationprogram interface) to create new experiences for customers.

在可視化推薦引擎背后一個很酷的產品叫做MADstack，這是一種基于云技術的計算機視覺以及人工智能模塊套裝，電子商務網站或其他網站可以接入API（應用程序接口），為客戶創造新的體驗。

MAD stands for Mind-Abled Devices,and MADstack is a creation ofMad Street Den, an Indianstartupthat’s barely two years old. Its founders, however, areseasoned pros from the hallowed domains of AI, user interface, and neuroscienceinSilicon Valley.

MAD代表思維可用設備，而MADstack也是Mad Street Den公司的產品，這是一個僅僅成立兩年的創業公司。然而它的創始人是來自于硅谷神圣領域的人工智能，用戶界面，以及神經科學方面的資深專家。

Ashwini Asokan led the mobileinnovation team at Intel’s Interaction and Experience Research Lab. Herhusband, and Mad Street Den co-founder, Anand Chandrasekaran was part of an AIteam atStanfordUniversitythat built Neurogrid – adevice that simulates the neurons and synapses of the brain to control bioniclimbs.

Well, now they’re here to help youpick a Bollywood saree.

Ashwini Asokan在英特爾交互以及體驗研究實驗室領導移動創新團隊。她的丈夫Anand Chandrasekaran，也是Mad Street Den的聯合創始人，Anand Chandrasekaran是斯坦福大學人工智能團隊的成員之一，他們打造了可以模擬大腦神經元和突觸來控制仿生肢的設備—Neurogrid。

Morethan visual search

不僅僅是視覺搜索

WhenTech in Asiafirst

profiled the couple last year, right after they moved to theirhometown of Chennai inIndia,they were experimenting with fun ways to use computer vision in everyday life,games, and business.

剛好在他們搬到在印度的家鄉Chennai后，Tech in Asia網站去年第一次介紹了這對夫婦，他們試驗用有趣的方式在日常生活，游戲以及商務中使用計算機視覺技術。

Amazon’s Firefly, billed as “visualsearch on steroids,” had come out around then. But, of course, the Fireflywould only work with Amazon. What the ‘Mad’ couple decided to do was to putthat capability and more on the cloud, for any ecommerce site to use.

標榜為“視覺搜索激素”的亞馬遜螢火蟲當時已經出現。但是螢火蟲只能在亞馬遜網站上使用。這對“瘋狂”的夫婦決定將此功能擴展到云技術，這樣任何電子商務網站都可以使用。

Othervisual search?playershave entered the ecommercescene inIndiasince then.Wazzat Labs, which recently graduatedfrom the TargetacceleratorinBangalore,has gone live on the US-based retailer’s website.iLenze, incubated at TLabs, is a new entrant. And last week,?Flipkartannounced a?visual search feature from Singaporean startup Visenze, even though it’s stillat a beta stage of development.

從那以后其他視覺搜索的玩家就開始進軍印度電子商務行業。最近從位于班加羅爾Targetaccelerator試驗成功的Wazzat Labs，已經在TLabs孵化的美國零售商網站行業新軍.iLenze上使用。上周Flipkart從新加坡創業公司Visenze發布了新的視覺搜索功能，盡管它仍處于開發測試階段。

So, visual search has suddenly becomethe buzzword in Indian ecommerce.

Mad Street Den co-founder AshwiniAsokan points out, however, that visual search-based recommendation is only oneof the tricks in MADstack’s bag. While object recognition powers the visualsearch, computer vision offers other possibilities such as recognizing thegestures and emotions of customers for businesses to understand user experienceand respond accordingly. Mad Street Den is thus positioned as a holisticartificial intelligence company.

因此，視覺搜索忽然成為了印度電子商務屆的流行語。

Mad Street Den的聯合創始人Ashwini Asokan指出，雖然基于視覺搜索的推薦僅僅是MADstack服務包的招數之一。然而對象識別功能支援了視覺搜索，計算機視覺提供了其他的功能，比如說識別客戶的手勢以及情感，這樣企業就會理解客戶的體驗并相應地做出回應。Mad Street Den也因此被定為為全面的人工智能公司。

Newlevel of personalization

定制化新水平

“Our platform is architected in sucha way that we can scale up any kind of computer vision feature,” Asokan tellsTech in Asia. “We start with visual recommendation, but as we goalong our technology is such that we move towards visual personalization.”

“我們以這種方式打造平臺，這樣我們就可以擴展任何類型的計算機視覺功能，”Asokan告訴Tech in Asia，“我們從視覺推薦開始，但是當推廣我們的技術時，就開始轉向視覺定制化。”

For example, if I click on a pinkBollywood saree, MADstack can see that it’s a fuchsia shade I prefer and notbright pink, as well as the patterns that appeal to me. This lends itself tomore granular personalization of my shopping experience than anything atext-based search can provide.

例如，如果我點擊一個粉紅色的寶萊塢紗麗，MADstack就會明白我喜歡的是櫻紅色，而不是亮粉色，以及吸引我的圖案。這可以為我提供更加完美的定制化購物體驗，這是其他任何基于文本的搜索不能提供的。

Visual search is still a work inprogress, even though the technology has been around for years.Google?acquired a visual search company –Like.com – way back in 2010. And yet, five years later the technology is stillnot widespread in ecommerce or other use cases.

視覺搜索仍然處于試運行階段，盡管這項技術已經存在多年。谷歌在2010年收購了一家視覺搜索公司——Like.com。然而五年后這項技術仍然沒有在電子商務或其他應用案例中普及開來。

“People still have to work out how tomake it usable,” says Asokan. “Visual search is a fundamental change in the waysomebody searches for something as people are used to text search. Nobody takespictures of things and searches for them.”

“人們仍然在努力研究如何使其可用，”Asokan說。“視覺搜索是對人們習慣使用的文本搜索方式的根本性轉變。沒有人去留意并搜索他們。”

‘Clicks’take on a new meaning

‘點擊’有新的含義

Mad Street Den itself started outwith recognition of gestures and facial expressions captured on smartphonecameras, when it built the MADstack last year. It didn’t take long to extendthat to visual search and recommendation, which had more takers in the market.But then to get it market-ready took five months afterraising?US$1.5 million in funding.

在去年打造MADstack時，Mad Street Den本身從智能手機攝像頭識別手勢以及面部表情開始。它并沒有花費很長時間就將其延伸到視覺搜素以及推薦服務中，這在市場上有更多的接受者。然而在籌集150萬資金后，花費了5個月的時間讓其充分投入市場。

Now the company is all set to rollout not only visual recommendation on ecommerce sites,but other possibilities of artificial intelligence as well.

現在公司不僅在電子商務網站上推出了可視化推薦功能，還可能會推出其他人工智能服務。

For now, Craftsvilla and Voonik arethe first of several ecommerce sites which will be plugging into the MADstack,Asokan tellsTech in Asia. After starting withvisual recommendations, several of these sites will roll out visual search –which will let you click a picture of somebody wearing a dress you like andsearch for it online, as demonstrated in this video.

Asokan告訴Tech in Asia，目前Craftsvilla和Voonik是最先嵌入MADstack的網站之一。在推出可視化推薦后，很多網站還將會推出可視化搜索，正如視頻中表明的那樣，這可以讓你點擊一張穿著你喜歡的裙子的人物圖片，并在網上進行搜索。

聲明：本文為個人翻譯最新商務類新聞文章，均有中英文對照，轉載請與作者聯系，尊重版權！