One would think that Bollywood fashion and artificial intelligence (AI) have nothing in common. Well, looks like technology can play a key role in redefining future fashion trends as leading designers Falguni and Shane Peacock use IBM Watson for their new collection and make cognitive fashion a reality. IBM Watson, running on IBM Cloud, analyzed multiple sources of fashion and Bollywood data for this collaboration. Tuhina Anand, Editor, Paul Writer caught up with Falguni and Shane Peacock to get more details on how they used IBM Watson.
Paul Writer: When it comes to a creative process, we still don’t think of technology, what made you choose IBM Watson for your new collection?
Falguni: Falguni Shane Peacock (FSP) as a brand is known for being edgy and futuristic and we have always been curious about using technology in our collection. It has always inspired and intrigued us on how technology can help in creating something better and creative and Watson allows you to do that.
Shane: We have used technology in the past and when IBM showcased Watson to us it seemed like a natural progression. And in this process we got to push our creative zone and broaden our outlook. FSP DNA signifies edginess, glamour and futuristic designs. Watson makes it simple to access huge data from various sources and fastens the process to make it relevant. I think that it would have been humanly impossible to sift through so much data so it definitely is a time saver. We then use our creative sensibility to conceptualize a collection that is unique yet has the stamp of our signature style. I would say that the end result is FSP style and Watson combined together.
Paul Writer: Can you elaborate on what results did you receive when you used IBM Watson?
Falguni: Watson is basically like a higher intelligence working for you and giving information on trends, silhouette that will work and does all the research that you would have to do and put in the man hours which is impossible. It is very convenient and helps in exploring limits which otherwise one would not be able to achieve. Watson gave us a lot of leads in terms of what are the colors, sort of prints, silhouettes, but eventually with all this data its best to use creativity and that’s what we did. We used the Watson derived prints and used data to create our own new silhouette. As creative people, Watson helped us to do a collection which is high on trend, has the right colors and very accurate.
Shane: We used large amount of data from run way shows from London, New York, Paris and Milan spanning autumn winter, spring summer and couture shows and what we got from Watson was very comprehensive data which saved us a lot of time. The output we got would have been physically impossible to achieve especially in such a short time. We got insights on silhouette, color, trend and pattern. It made our work so simplified that if I wanted to know the silhouette that worked in the last 20 years or the colour predicted for 2018, Watson gave us the results in mere seconds and that was the most amazing part of working with it. Typically, when we start a collection, we do use inspiration but make do with using limited information that we have access to. Our new collection is truly about collaboration and amalgamation of FSP creativity with IBM Watson.
Paul Writer: What was the coolest part of using Watson?
Falguni: The coolest part of using IBM Watson is that it allows you to travel back and forth in time in mere seconds.
Shane: The coolest part of working with Watson is that it works with mountains of data which is unstructured and what it derives from it is relevant data and that’s really wow.
Sriram Raghavan, Director, India Research Lab, IBM, shares the technology aspect of how IBM Watson was used by Falguni and Shane Peacock.
Paul Writer: What was the data you used for this AI inspired fashion?
Sriram: For understanding the global fashion trends, IBM Watson analyzed around 600,000 publicly available historical fashion runway images for the past decade (2006-2017) spanning the ‘Big Four’ fashion weeks (London, Paris, Milan, New York). In order to understand Bollywood fashion trends, IBM Watson analyzed close to 5,000 Bollywood celebrity images from various social media sites, besides 3,000 fashion-related images from the history of Bollywood in the form of Bollywood movie posters across 4 decades starting from 70s. The first data source is representative of high-end couture and the second one is more indicative of Bollywood-centric fashion across the years. Finally, in order to generate novel prints, the team gathered close to 100,000 print swatches.
Paul Writer: What were the outputs from Watson?
Sriram: Watson was able to analyze the collection of fashion images to ascertain popular and trending colours for every season for a decade, and then predict the trending colours for the next season. We were also able to analyze the dominant prints and silhouettes for each season, which the designers can use as a starting point to explore further. The insights provided by this were used as inspiration for the collection designed by Falguni Shane Peacock. The analysis was presented to them as an interactive web application hosted on IBM Cloud where they could explore the colour trends for the next season, analyze the key prints and silhouettes for each season and derive inspiration for their new collection.
Paul Writer: How does Watson leverage unstructured data to its advantage?
Sriram: 80 percent of all data currently available is dark and unstructured, and current computing systems cannot read or use it. By 2020, that number will be percent. Cognitive computing has the ability to read and make sense of this gigantic amount of data that was being wasted so far. Watson is IBM’s cognitive computing platform that understands the world in the way that humans do through senses, learning, and experience. It learns at scale, reasons with purpose and interacts with humans naturally. Using machine learning, deep learning, and natural language processing Watson is able to consume vast amounts of information to identify patterns, make hypotheses and derive deep insights in a short turnaround time.
