This museum is using AI to remind us of all the threads we have in common

man working on textile

Open your closet. If you find something with a paisley print you’ve also found a connection to the art of the ancient world. The curvy design is one of the most ubiquitous motifs in modern global fashion, but its roots go back a few millennia.

The pattern probably came out of Persia, now Iran, and made it to Babylonia, in what is now Iraq. Scribes there likened it to an uncurling date palm shoot. It also emerged in various parts of India where it came to be called as badamkairi, or amba, among others—to describe its teardrop or mango-like shape.

It travelled to the West from the subcontinent a couple of hundred years ago and became all the rage. Eventually it took on the name of a town in Scotland where factories once specialized in imitating Kashmiri shawls with the pattern for global export. ‘Paisley’ has been mainly in—and occasionally out of—style ever since.

Paisley’s journey, across time and continents, has been unflagging. And it is just of one many histories of textiles brought to life by INTERWOVEN, an artificial intelligence (AI) powered virtual experience from the Museum of Art and Photography (MAP) in Bengaluru, one of India’s main tech hubs.

On the INTERWOVEN website, a single click on the image of a mid-19th-early 20th century Kashmiri shawl fragment with the paisley motif leads one on a cultural odyssey that takes you to dozens of places where the motif appears. This includes a church in Britain, a chasuble from Italy, and a cloth panel from ancient Persia among others.

INTERWOVEN is an innovative tool developed in collaboration between MAP, the MAP Academy, and Microsoft’s AI for Cultural Heritage initiative. It uses machine learning and AI to reveal the threads connecting MAP’s textile artworks and those from the collections of other museums from across the world.

Microsoft’s AI for Cultural Heritage leverages the power of AI to empower people and organizations dedicated to the preservation and enrichment of cultural heritage. The platform demonstrates how technology can serve as a useful tool for helping people preserve and connect to cultural heritage on a global scale, and across generations.

As part of the initiative, Microsoft supports individuals and organizations, in this case MAP, through collaboration, partnership, and investment in AI technology and resources.

“Using technology to enhance human ingenuity, celebrate human creativity, and enable human connection is at the heart of Microsoft’s work,” says Brad Smith, president and vice chair, Microsoft. “India has been a melting pot of cultures and an epicenter of global trade for centuries. It’s exciting to see how INTERWOVEN, our first AI for Cultural Heritage project to come out of India, reminds us both of the vibrancy of different cultures and how these traditions are shared in a conversation across time and place.”

INTERWOVEN uses Azure Custom Vision and AI Text Analytics services to help find common threads and shared histories in artistic traditions between MAP’s textile artifacts and the artistic treasures of its partner museums. So, you can see the journey of textiles not just through motifs and designs but also through filters of color, occasions, geographies, and eras, among others.

South Asia’s centuries-old traditions and their trade make textiles a great entry point to study the spread of culture.

“Textiles from the subcontinent were central to the trade, with some of the best cotton and silk being exported by land and maritime routes. As a result, textiles have influenced and shaped global politics for centuries,” points out Shrey Maurya, managing editor at MAP Academy, INTERWOVEN’s knowledge partner.

“They were the commodity to purchase spices that were traded around the world,” explains Kate Irwin, curator of costume and textiles at Rhode Island School of Design’s Museum of Art. “Cotton was a global industry for a thousand years and that industry was centered in India till the 19th century; we have long span of history where the world—from East Africa and Indonesia to the Arab world and Europe—was craving for textiles from South Asia. Hence it makes sense to start with South Asian textiles to understand pattern and artistic language around the world; there’s simply so much solid history there.”

INTERWOVEN can help curators, academics, and art historians find common threads between artefacts and enable them to tell new stories or unravel new connections between different civilizations. They can now access vast collections sitting in their homes, and AI can help find obscure threads of connectivity between objects.

But by virtue of it being discovery-based, the platform also has the ability to draw in someone with even just a passing interest in the subject. Visitors to the experience can either go down a curated journey or explore MAP’s collection and click on any artwork to explore the various connections on their own.

“Here you’ll never look at an object in isolation. The moment you select anything on the platform—even a stray motif on a textile patch—the algorithm will automatically show you all the connections it has around the world,” Maurya says. “In that, INTERWOVEN helps foster and sustain interest in textiles and art.”

Making art and the museum experience accessible is one of MAP’s primary objectives and technology is at its heart.

“We’ve always wanted to be a museum of the future. India’s youth are native technology users and MAP wants to harness it to showcase India’s cultural heritage and show that art can be fun and not elitist,” says Abhishek Poddar, the museum’s founder.

So, when the COVID-19 pandemic derailed its plans of opening its doors to the public in 2020, MAP embraced digital solutions just as seamlessly as its prospective young audiences have technology.

Among the efforts the team took to engage with their online communities, is Museums Without Borders, a digital collaboration with its partner museums that eventually led to the creation of INTERWOVEN. MAP also launched the MAP Academy during this time, which is a major online platform consisting of an art encyclopedia and online courses, including one on the history of textiles in South Asia, which is also featured on INTERWOVEN for users who want to learn more.

“While INTERWOVEN began with an idea to explore MAP’s entire collection, we realized one of the artforms that spoke so eloquently to global connections was textiles,” says Kamini Sawhney, director of MAP.

MAP partnered with 16 other museums including the Metropolitan Museum of Art in New York, the Cleveland Museum of Art, the Victoria and Albert Museum in London, and the Smithsonian in Washington, DC to connect their collections with the INTERWOVEN experience.

“We are thrilled to be a part of the INTERWOVEN project, led by Microsoft’s AI for Cultural Heritage program and the Museum of Art and Photography in India. It serves as a window into our magnificent collection, allowing people to connect with Cleaveland Museum of Art’s artwork visually, emotionally, or at a scholarly level. This project is a wonderful implementation of our Open Access collection” said Jane Alexander, chief digital information officer, the Cleveland Museum of Art. “We recognize that people may never visit Cleveland to experience our collection in person, this interactive tool allows visitors around the world to explore our remarkable collection in a way that offers deeper context and cultural insights, ultimately making art more relatable to them.”

Creative technologists from A_da, an experience innovation and design agency, helped devise the AI component and its deployment on INTERWOVEN. They deployed a mix of several Microsoft AI services to deliver rich connection results between MAP’s digital collection and open access collection pieces of its partner museums.

MAP’s digital collection has been codified using precise keywords, meta tags, and high-resolution photographs. The algorithm uses Azure AI Text Analytics to smartly parse through these keywords and Azure AI Custom Vision to help make visual and model comparisons.

Even though the algorithm is live, the creative team is simultaneously working on improving the visual detection by training the models to be able to identify visual elements like floral motifs to animal patterns to geometric shapes found in South Asian art.

Mandara Vishwanath, from the MAP Academy, continues to remain fascinated as she encounters various similarities in the style and aesthetics of textiles in different regions.

“For instance, we have many kalamkari textiles within India that feature stories and episodes from the Ramayana and Mahabharata. These stories are also presented on block-printed cotton fabrics in South-East Asia. INTERWOVEN helped us observe the style differences between the textile paintings from India and the block printed fabrics found in countries like Indonesia and Malaysia,” she says.

For someone interested in the subject, all this information, categorized and contextualized, is packaged in a manner that is easily understood. For others, it’s an open canvas to curate their own journeys.

And even as you click from one artwork to another, crossing geographical boundaries and time periods, the thing that INTERWOVEN reminds you is that no matter what, beyond our differences, we’re all on a common journey connected by threads that you never knew existed.

Top Image: A Kashmiri artisan working on a shawl with the ‘paisley’ pattern (Isam Wani for Microsoft)