Building on last year’s collaboration with Tommy Hilfiger, in a project that used artificial intelligence to identify fashion trends in advance, IBM and the Fashion Institute of Technology’s Infor Design and Tech Lab are looking to expand the effort with a new AI partnership to train today’s burgeoning crop of fashion designers.
Where the Hilfiger project used IBM’s AI to target trends for design and creation, the current deal’s scope includes those things as well as product development, planning, allocation, merchandising, customer sales and service. The goal is to cover more areas of fashion retail, as “opportunities across the value chain,” Karl Haller, global leader for consumer center of competence at IBM Global Business Services, told WWD.
“[Think of those] capabilities, combined with deep learning and natural language processing and natural language understanding,” Haller continued. “When you get into the product design and product development phases of the production calendar or the overall supply chain, there’s a lot of visual work that takes place early on — with designers going out and shooting images, sometimes fashion images, architectural images, home design or inspirations from people on the street.”
A designer could spot a tile in Morocco, he continued, and envision a print on a garment. With IBM’s tools, someone could look at that pattern or particular color combination, and explore whether someone had done something on that recently. “And if so, who? Have we done something like that in our past? Does it exist in one of our existing fabric or print vendors?” he added. “As a way of being able to understand images and create context around those images, we believe this is going to help save time in the design process and also provide new sources of inspiration for design teams.”
In a lab environment, students and faculty will get a front-row view of IBM’s AI application programming interfaces. APIs are software tools that let developers work with or tie into a company’s platforms. Here, the goal is to give students hands-on experience while appealing to educators, so that they integrate the tech into the curriculum.
Whereas a dozen or so students worked on the Hilfiger effort, the partners hope more will avail themselves of the expanded resources. “One of the exciting opportunities from FIT’s perspective is that we’re going to be working off a base platform, and really looking at bringing the kind of fresh perspectives and insights of creative and talented students to bear on these APIs,” FIT’s Michael Ferraro, director of Infor DTech Lab, told WWD. “[These students can] begin to tune and adjust and train and develop user experiences that are really informed by their knowledge and their insights.”
According to Ferraro, the students looked at how the tools draw insights from things such as deep social media listening and how that can inform supply chain decisions. “I’m envisioning an interactive experience which utilizes a whole series of tools in concert with each other,” he explained. “The mirror recognizes you, you have a conversation back and forth with it. You say that you’re going to go see a concert and you’re going to go see XTC, and the band members are wearing lots from Gigi’s newest collection.”
From there, a brand’s AI could say it doesn’t carry that exact item in the store, “but we have these which are very similar,” he added. “So you get this whole cascade of interaction that’s informed by machine learning and visual recognition. And when you combine the tools in that way, you actually start to get a much more fluid and interactive retail experience.”
For large brands, the IBM and DTech Lab aims to help “de-risk” innovation, by offering a fairly siloed testing environment to explore new areas.
These are opportunities and challenges facing modern retailers today, making any effort to boost fashion’s creative and technical education seem rather timely. But education is only one facet. Turns out, it may also be a good business decision for IBM — a possibly savvy play to appeal to the next generation of designers early on.
“The benefit of working directly with students and faculty at FIT is that these students are all going to end up being clients of IBM’s or other companies at some point,” Haller added. “We’re getting input from [future] business users, and many of the faculty from FIT also come out of the industry. So we’re getting a real practical, industry perspective.”
That can lead to a number of different use cases. One, for example, is style iteration: Once there’s a human-led design, the machines could automate different variations according to trends identified in the data. Think shorter hems, high-rise versus low-rise alterations and other tweaks. Machines could do the grunt work, freeing people from such cumbersome tasks, so they can ideate and create new works.
It’s no secret that various organizations, including Amazon, are exploring areas like AI-driven fashion design. Still, the practical application of such innovation still seems far off in the distance. Iteration, however, as a factor of automation, may be within the industry’s grasp in the nearer term, and fashion’s rising stars and sage educators could be the ones to blaze that trail.
Here’s more on the IBM relationship, courtesy of a recent video from FIT.