The velocity with which new applied sciences hit the market is nothing in comparison with the velocity with which gifted researchers discover artistic methods to make use of them, prepare them, even flip them into issues we are able to’t stay with out. One such researcher is MIT MAD Fellow Alexander Htet Kyaw, a graduate pupil pursuing twin grasp’s levels in architectural research in computation and in electrical engineering and laptop science.
Kyaw takes applied sciences like synthetic intelligence, augmented actuality, and robotics, and combines them with gesture, speech, and object recognition to create human-AI workflows which have the potential to work together with our constructed setting, change how we store, design advanced constructions, and make bodily issues.
Considered one of his newest improvements is Curator AI, for which he and his MIT graduate pupil companions took first prize — $26,000 in OpenAI merchandise and money — on the MIT AI Convention’s AI Construct: Generative Voice AI Options, a weeklong hackathon at MIT with ultimate displays held final fall in New York Metropolis. Working with Kyaw had been Richa Gupta (structure) and Bradley Bunch, Nidhish Sagar, and Michael Received — all from the MIT Division of Electrical Engineering and Laptop Science (EECS).
Curator AI is designed to streamline on-line furnishings buying by offering context-aware product suggestions utilizing AI and AR. The platform makes use of AR to take the size of a room with places of home windows, doorways, and current furnishings. Customers can then communicate to the software program to explain what new furnishings they need, and the system will use a vision-language AI mannequin to seek for and show numerous choices that match each the consumer’s prompts and the room’s visible traits.
“Customers can select from the urged choices, visualize merchandise in AR, and use pure language to ask for modifications to the search, making the furnishings choice course of extra intuitive, environment friendly, and personalised,” Kyaw says. “The issue we’re making an attempt to unravel is that most individuals don’t know the place to begin when furnishing a room, so we developed Curator AI to offer good, contextual suggestions primarily based on what your room appears to be like like.” Though Curator AI was developed for furnishings buying, it might be expanded to be used in different markets.
One other instance of Kyaw’s work is Estimate, a product that he and three different graduate college students created throughout the MIT Sloan Product Tech Convention’s hackathon in March 2024. The main focus of that competitors was to assist small companies; Kyaw and crew determined to base their work on a portray firm in Cambridge that employs 10 individuals. Estimate makes use of AR and an object-recognition AI know-how to take the precise measurements of a room and generate an in depth price estimate for a renovation and/or paint job. It additionally leverages generative AI to show photographs of the room or rooms as they may appear like after portray or renovating, and generates an bill as soon as the undertaking is full.
The crew gained that hackathon and $5,000 in money. Kyaw’s teammates had been Guillaume Allegre, Might Khine, and Anna Mathy, all of whom graduated from MIT in 2024 with grasp’s levels in enterprise analytics.
In April, Kyaw will give a TedX speak at his alma mater, Cornell College, by which he’ll describe Curator AI, Estimate, and different initiatives that use AI, AR, and robotics to design and construct issues.
Considered one of these initiatives is Unlog, for which Kyaw related AR with gesture recognition to construct a software program that takes enter from the contact of a fingertip on the floor of a fabric, and even within the air, to map the size of constructing parts. That’s how Unlog — a towering artwork sculpture constructed from ash logs that stands on the Cornell campus — took place.
Gesture Recognition for Suggestions-Based mostly Blended Actuality and Robotic Fabrication of the Unlog Tower
Video: Alexander Htet Kyaw
Unlog represents the chance that constructions will be constructed straight from an entire log, reasonably than having the log journey to a lumber mill to be became planks or two-by-fours, then shipped to a wholesaler or retailer. It’s a superb illustration of Kyaw’s need to make use of constructing supplies in a extra sustainable means. A paper on this work, “Gestural Recognition for Suggestions-Based mostly Blended Actuality Fabrication a Case Examine of the UnLog Tower,” was revealed by Kyaw, Leslie Lok, Lawson Spencer, and Sasa Zivkovic within the Proceedings of the fifth Worldwide Convention on Computational Design and Robotic Fabrication, January 2024.
One other system Kyaw developed integrates physics simulation, gesture recognition, and AR to design energetic bending constructions constructed with bamboo poles. Gesture recognition permits customers to govern digital bamboo modules in AR, and the physics simulation is built-in to visualise how the bamboo bends and the place to connect the bamboo poles in ways in which create a secure construction. This work appeared within the Proceedings of the forty first Schooling and Analysis in Laptop Aided Architectural Design in Europe, August 2023, as “Lively Bending in Physics-Based mostly Blended Actuality: The Design and Fabrication of a Reconfigurable Modular Bamboo System.”
Kyaw pitched an analogous thought utilizing bamboo modules to create deployable constructions final 12 months to MITdesignX, an MIT MAD program that selects promising startups and offers teaching and funding to launch them. Kyaw has since based BendShelters to construct the prefabricated, modular bamboo shelters and neighborhood areas for refugees and displaced individuals in Myanmar, his house nation.
“The place I grew up, in Myanmar, I’ve seen a number of day-to-day results of local weather change and excessive poverty,” Kyaw says. “There’s an enormous refugee disaster within the nation, and I wish to take into consideration how I can contribute again to my neighborhood.”
His work with BendShelters has been acknowledged by MIT Sandbox, PKG Social Innovation Problem, and the Amazon Robotics’ Prize for Social Good.
At MIT, Kyaw is collaborating with Professor Neil Gershenfeld, director of the Middle for Bits and Atoms, and PhD pupil Miana Smith to make use of speech recognition, 3D generative AI, and robotic arms to create a workflow that may construct objects in an accessible, on-demand, and sustainable means. Kyaw holds bachelor’s levels in structure and laptop science from Cornell. Final 12 months, he was awarded an SJA Fellowship from the Steve Jobs Archive, which offers funding for initiatives on the intersection of know-how and the humanities.
“I take pleasure in exploring totally different sorts of applied sciences to design and make issues,” Kyaw says. “Being a part of MAD has made me take into consideration how all my work connects, and helped make clear my intentions. My analysis imaginative and prescient is to design and develop programs and merchandise that allow pure interactions between people, machines, and the world round us.”
