Posts tagged with "Artificial Intelligence":

Placeholder Alt Text

Los Angeles’s TECH+ Expo brought together innovations in project delivery

On February 6, The TECH+ Expo transformed the second floor of Los Angeles’s Line Hotel into a showcase of the latest innovations in architectural technology. But rather than exhibiting 3D printers, robot arms, and brick-laying drones, the conference highlighted products designed to streamline design research, project delivery, and the architect-to-client relationship. Chief creative officer of BQE Software, Steven Burns, FAIA, provided a demonstration of CORE, the company’s latest app designed to consolidate the billing, accounting, and reporting necessary to keep an architecture firm afloat. While CORE can distill a firm’s complex financial information in its easy-to-read web format, the company added artificial intelligence (AI) to enable its users to have human-like conversations with the app to simplify its interface even further. Using a fictional company as an example, Burns presented a 5-minute interaction with the app that elucidated everything from then-current financial progress to unpaid bills and variously categorized expense items. In her keynote lecture, Dr. Upali Nanda, director of research for global architecture firm HKS Inc., demonstrated how her firm has pioneered methods of design research that positively contributes to the mental health of its buildings’ occupants. Her presentation began with a valuable quote from Thomas Fisher, dean of the College of Design at the University of Minnesota: “Architects have been slow to champion the return on investment that their work can bring, but even a little data can convince clients that spending more can mean saving more.” Dr. Nanda then went on to explain how the data collected at HKS—focused on occupancy outcomes—has improved the performance of its projects while convincing clients to invest in operationally effective and energy-saving technology. “Design is a hypothesis,” Dr. Nanda concluded, “but what happens during occupancy is the outcome.” Several TECH+ participants shared their insights and techniques for improving construction efficiency. Chester Weir, design lead of global construction company Katerra, outlined how his company combined end-to-end integration with technological innovation to produce forward-thinking solutions to global construction issues. The company’s Materials & Supply Chain services, for instance, have aggregated demand across markets to supply building materials for itself and other market sectors. Rudy Armendariz, senior VDC/BIM Manager at Balfour Beatty, elaborated on the challenges his company faced in determining how to construct a people mover at the Los Angeles International Airport (LAX) without disrupting the airport’s automobile traffic. LAX Integrated Express Solutions (LINXS) developed a system for efficiently working on the project within the 5-hour period the construction team is allotted each night. The next TECH+ Expo will be held in New York City on July 16.
Placeholder Alt Text

AN talks to Eyal Weizman about tech in truth-telling ahead of Forensic Architecture’s first U.S. survey

Forensic Architecture has garnered a significant reputation within the field of architecture (they had a major showing at the most recent Chicago Architecture Biennial) and beyond for their work reconstructing violent events perpetrated by state actors and others using architectural tools and emerging technologies. The collective’s work has been displayed everywhere from the courthouse to major art exhibitions, including during this past year’s Whitney Biennial. The video One Building, One Bomb, co-produced with The New York Times, won an Emmy this past year, and in 2018 they were also nominated for the United Kingdom’s prestigious Turner Prize. This month, Forensic Architecture, which is based out of Goldsmiths, University of London, will have its first major U.S. survey; Forensic Architecture: True to Scale will open on February 20 at the Museum of Art and Design at Miami Dade College. Ahead of the Miami exhibition, AN spoke with Forensic Architecture founder Eyal Weizman to discuss the changes of the past decade, the power of technology, and the importance of forensics in a “post-truth” era. Drew Zeiba: Forensic architecture began a decade ago. How has the project changed and how have the tools you use evolved since then? Eyal Weizman: When we started around 2010, it was the beginning of the Arab Spring and the really heartbreaking civil war that came in its wake. Those particular sets of conflicts had a particular texture to them. They happened in an environment that had a lot of mobile phones and in the areas where there's internet connectivity, and where the government’s ability to shut down the internet was not always successful. We started being in an environment where increasingly you had more and more videos around incidents that we could map. It was also the early teens where at the time, in London great protest around tuition fees and then the big protest after the police killing of Mark Duggan in North London. This killing was during a period when police did not yet have dash cams. And ever since, we've seen the introduction of body cams and dash cams to police investigations. If you look today at the conflicts that are taking place, we have several thousand videos, hours long, broadcasting live as things are happening. The sheer media density requires us to use different technologies in order to bring accountability. We have recently developed machine vision and machine learning technologies that, working together with human researcher, can speed up the process of sieving through thousands and thousands and thousands of hours of content coming from confrontations with policing Hong Kong, for example. In relation to police violence, we have now concluded the investigation in Chicago [into the police killing of Harith Augustus] with full body cams available, several dash cams, a CCTV, etc. We are working in a much more media-saturated environment and need new tools like artificial intelligence to help us identify materials like our work on Warren Kanders that used machine learning. [Kanders is the ousted vice chairman of the board of the Whitney Museum, whose company Safariland sells tear gas used at the U.S.-Mexico border, in Gaza, and elsewhere, including in U.S. cities such as Baltimore and Ferguson, Missouri.] We're creating virtual reality sites for witnesses to walk through the scene of the together with the psychologists and lawyer and protection, but can also recall events. And we are trying to be at cutting edge of technologies that would help social movement and civil society to invert the balance of epistemological power, against the monopoly of knowledge that states have over information in battlefields and in crime scenes. The abundance of images also has to do with the increasing presence of surveillance—including by CCTV cameras and police body cams, as you mention. How can architectural and technological tools invert the power relationship embedded in some of these commonplace image-making tools? Forensics have to be in the hands of the people. Forensics was developed as a state tools, as a form of state power, as a police tool. But when the police is the agency that dispenses violence and the agency that's investigating it, we have a problem. We absolutely need to be able to have independent groups holding police to account. And what we have is our creativity and we can effectively mobilize and make more of much fewer bits of data and image, because we're working aesthetically and we work socially with those independent groups in producing evidence. We socialize the production of evidence, we make it a collective social practice that involves the communities that are experiencing state violence continuously. At the same time, Forensic Architecture often works in places where there is seemingly a limited amount of the evidence or data that investigators typically rely upon, or with evidence that is biased. Police body cams show the officer’s perspective only, for example. Your work is coming at a time that people are describing as “post-truth.” How does the work of Forensic Architecture fit in to this political context? The very nature of what we call investigative aesthetics is based on working with weak signals and with partial data. You need to fill that gap with a relation between those points you have, sort of like stars in a dark sky. You see very few dots and we need to actually see how they can support the probability of something to have occurred. And any investigative work that comes from the point of view of civil society is both about demolishing and building. So we need to use our training as critical scholars in deconstructing police statements, or military statements taken by secret services or the government—and we need to take those ruins, those scattered bits of media flotsam that exist and build something else with it. There’s always demolition and rebuilding that takes place. That is very structural to our work. Right now, the mistrust in public institutions in the political sphere, In the so-called post-truth era, that trust is not being replaced. Those that tell us not to believe anymore in science and in think tanks and in experts are not building a new epistemology in its stead. They're simply demolishing it. Rhetoric replaces verification. What we do similarly to them is we are questioning state given truths. We are attacking those temples of power and knowledge, but we attempt to replace them with a much more imminent form of evidence production that socializes the production of that evidence.
Placeholder Alt Text

Architects apply the latest in fabrication, design, and visualization to age-old timber

Every so often, the field of architecture is presented with what is hailed as the next “miracle building material.” Concrete enabled the expansion of the Roman Empire, steel densified cities to previously unthinkable heights, and plastic reconstituted the architectural interior and the building economy along with it.  But it would be reasonable to question why and how, in the 21st century, timber was accorded a miracle status on the tail-end of a timeline several millennia long. Though its rough-hewn surface and the puzzle-like assembly it engenders might seem antithetical to the current global demand for exponential building development, it is timber’s durability, renewability, and capacity for sequestering carbon—rather than release it—that inspires the building industry to heavily invest in its future.  Cross-laminated timber (CLT), a highly resilient form of engineered wood made by gluing layers of solid-sawn lumber together, was first developed in Europe in the early 1990s, yet the product was not commonly used until the 2000s and was only introduced into the International Building Code in 2015. While mid-to-large range firms around the world have been in competition to build the largest or the tallest timber structures to demonstrate its comparability to concrete and steel, a number of independent practitioners have been applying the latest methods of fabrication, computational design techniques, and visualization software to the primordial material. Here, AN exhibits a cross-section of the experimental work currently being pursued with the belief that timber can be for the future what concrete, steel, and plastic have been in the past. AnnaLisa Meyboom In the Fall of 2018, 15 of professor AnnaLisa Meyboom’s students at the University of British Columbia (UBC), along with David Correa at University of Waterloo, Oliver David Krieg of Intelligent City, and 22 industry participants designed and constructed the third annual Wander Wood Pavilion, a twisting, latticed timber structure made up entirely of non-identical components.  By taking advantage of the advanced fabrication resources available at the UBC Centre for Advanced Wood Processing, including a CNC mill and an multi-axis industrial robot, the project was both a learning opportunity for its design team and a demonstration to a broader public that timber is a more than viable material to which contemporary fabrication technologies can be applied. The pavilion forms a bench on one end that's large enough for two people, a public invitation test the structure's strength and durability for themselves. While the pavilion only required three days to fabricate and assemble on-site, a significant amount of time and energy was spent ensuring its quick assembly when the time came. A rigorous design workflow was established that balanced an iterative design process with rapid geometric output that accounted for logical assembly sequencing. Every piece of the pavilion was then milled to interlock into place and be further secured by metal rivets. The project was devised in part to teach students one strategy for narrowing the gap between digital design and physical fabrication while applying a novel material. In this vein, a standard industrial robot was used throughout the fabrication process that was then “set up with an integrator specifically to work on wood,” according to Meyboom. Gilles Retsin While Gilles Retsin, the London-based architect and professor at the Bartlett School of Architecture, has long experimented with both computational design and novel methods of fabrication, a recent focus on timber has propelled his practice into a bold new direction. A giant wooden structure installed at London’s Royal Academy in early 2019, for instance, was the architect’s first attempt at applying augmented reality to modular timber construction through the use of Microsoft’s Hololens. “We used AR to send instructions directly from the digital model to the team working on-site,” Retsin explained. “AR therefore helps us understand what a fully-automated construction process would look like, where a digital model communicates directly with people and robots on site.” In a recent international competition set in Nuremberg, Germany, Retsin set his sights on a much larger scale for what would have been the world’s first robotically prefabricated timber concert hall. Designed in collaboration with architect Stephan Markus Albrecht, engineering consultancy Bollinger-Grohmann, and climate engineers Transsolar and acoustic specialists Theatre Projects, the proposal takes advantage of the site’s location in a region with an abundance of timber while envisioning the material’s application to a uniquely challenging building type. The building’s form exhibits the material’s lightness using 30-foot sawtooth CLT prefabricated modules over the main lobby spaces, which are exposed from the exterior thanks to a seamless glass envelope.  “Designing in timber not only means a more sustainable future, but also has architects profoundly redesigning buildings from the ground up,” said Retsin. “It’s a challenging creative task, we’re really questioning the fundamental parts, the building blocks of architecture again.”  Casey Rehm For SCI-Arc professor Casey Rehm, working with timber has meant challenging many issues in the field of architecture at once. Timber is a rarely-considered building material in Los Angeles given the high time and material costs associated with its transportation and manufacturing. “Right now,” Rehm said, “the industry is manually laying up two-by-sixes into industrial presses, pressing them into panels, and then manually cutting window openings.” But if timber waste itself was adopted as a building material, he argued, the material could be far more globally cost-efficient.  While timber has been used in the construction of increasingly large structures around the world, such as multistory housing developments and office buildings, Rehm believes the material can be reasonably adapted to a smaller scale for quick deployment. In this vein, Rehm has been researching strategies with his students for producing inexpensive CLT panels for the construction of homeless housing and accessory dwelling units in Los Angeles, a city with a particularly conspicuous housing shortage.  But aside from its potential as a cost and material-efficient material, the architect has applied timber to even his most exploratory design work. NN_House 1, a sprawling single-floor home Rehm proposed in 2018 for the desert plains of Joshua Tree, California, was designed in part using a 3D neural network to develop ambiguous divisions between rooms, as well as to blur the divide between interior and exterior. The AI was trained on the work of modernist architects—while producing idiosyncrasies of its own—to develop a living space with multiple spatial readings. Kivi Sotamaa As an architect practicing in Finland, Kivi Sotamaa is certainly not unique in his community for his admiration of the far-reaching possibilities of timber construction. He is, however, producing novel research into its application at a domestic scale to reimagine how wood can be used as a primary material for home construction. The Meteorite, a three-story home the architect has designed near Helsinki constructed entirely of locally-grown CLT, was designed using an organizational strategy the architect has nicknamed ‘the misfit.’ This system, as Sotamaa defines it, creates two distinct formal systems to generate room-sized interstitial spaces that simultaneously act as insulation, storage space, and housing for the building’s technical systems. “Aesthetically,” Sotamaa elaborated, “the misfit strategy allows for the creation of a large scale monolithic form on the outside, which addresses the scale of the forest, and an intricate human-scale spatial arrangement on the interior.” Altogether, the architect estimates, the home’s CLT slabs have sequestered 59,488 kilograms, or roughly 65 tons, of carbon dioxide from the atmosphere. The Meteorite was developed and introduced to the client using virtual reality, and Sotamaa hopes to apply other visualization technologies to the design and production of timber architecture, including augmented reality that could allow builders to view assembly instructions in real-time on site. “When the pieces are in order on-site and [with clear] instructions,” Sotamaa explained, “the assembly of the three-dimensional puzzle can happen swiftly and efficiently, saving energy and resources when compared with conventional construction processes.” 
Placeholder Alt Text

Sidewalk Labs is using machine learning to make neighborhood design smoother

Sidewalk Labs, the Alphabet subsidiary focused on urban technology, has been working on a new software tool for generating optimized city layouts. In an effort to combat the disconnect between various stakeholders in the urban planning process—architects, planners, engineers, and real estate developers—and their software, product manager Violet Whitney and designer Brian Ho have created a new computational tool that analyzes a wide array of data to automatically create thousands, or millions, of neighborhood layouts from a baseline design.  Examples of inputs and considerations Sidewalk Labs listed include regulatory concerns, street layouts, block orientations, real estate, weather, building height, and more, which can then be considered against “quality of life” measures. Using machine learning, the technology should get “smarter” over time.  Design always takes compromise. Too much density can cause traffic or an abundance of building shadows, yet too little is also no good. A lot of open space can be great, until it gets in the way of easy movement. Designers and other involved parties can consider their goals and generate many new designs to see different possibilities, which would then inspire and instruct human designers (there’s no doing away with architects just yet). The Sidewalk Labs team also wants to diminish the disconnect between the different software different parties use, from developers' Excel sheets to the powerful modeling tools used by engineers, and make communication easier.  In a digital case study, the researchers presented a plan for a two-by-two-block neighborhood that aimed for at least 45 percent open space, 49 percent daylight access, and as a proxy for density, 1.5 million square feet of floor area. While the human-led design hit the required parameters, then using the new tool, researchers were able to generate thousands of variations of that initial design, around 400 of which outperformed the original. Sidewalk Labs also suggested that community feedback might be integrated into the technology and its holistic process in the future, likely important given the pushback its high-tech timber neighborhood—accused of having all sorts of ulterior motives like corporate surveillance—has been getting in Toronto. The tool is part of a broader trend to introducing automation into design, whether on the interior scale, such as WeWork’s proprietary space-laying algorithms, or at the city scale such as emerging “digital twin” projects. For more on the latest in AEC technology and for information about the upcoming TECH+ conference, visit https://techplusexpo.com/events/la/
Placeholder Alt Text

The Architectural Beast distorts architectural imagery at the FRAC Biennale

For the 2019 Fonds Régional d’Art Contemporain (FRAC) Biennale in Orléans, France, SCI-Arc director Hernan Diaz-Alonso curated The Architectural Beast, an installation featuring 17 contemporary artists and architects. Together with Diaz-Alonso, Los Angeles-based designer Casey Rehm co-produced the installation: 12 paired video screens that nod towards Marcel Duchamp's The Bride Stripped Bare by Her Bachelors, Even (or The Large Glass installation). The top panel exhibits printed images the artists have selected to represent their work, while the lower screens show that same imagery being transformed by artificial intelligence software developed by Rehm. Also named The Architectural Beast, the software was designed to independently alter the imagery presented over the course of the three-month installation. According to Rehm, the program's AI is "initially trained on curated datasets of images and texts of the artists representing an institutional understanding of architecture, to an understanding of architecture of populist valuation." The AI, in other words, spends each night conducting image searches for the day's most popular architectural images and then uses the results to manipulate the original imagery. "By the second month of its life," Rehm explains, "it should cross the 50 percent line of curated artist and internet images in its network."
"Through artificial intelligence," wrote Diaz-Alonso in the installation description, "the work featured will be exposed to a perpetual state of transformation and mutation. The exhibition gathers a key set of practices, primarily from architecture, but also from art and fashion, to reveal facets of the strange beast that the tumultuous paradigm shifts of recent decades have left behind." The AI also uploads the imagery as individual posts on Instagram daily under the username @thearchitecturalbeast, each of which is complemented by cryptic texts that are developed by a separate AI program. This writing, which at first glance read like heavy theoretical essays with the aid of predictive text, was initially trained on the written work of Rehm, Liam Young, and Damjan Jovanovic. The combination of text and imagery created by The Architectural Beast demonstrates one way architects can let go of the wheel and give artificial intelligence greater agency in the role of human-centered design. The installation is currently on view through January 19.
Placeholder Alt Text

ACADIA 2019 showcased the state of digital design

The presentations and activities at this year’s ACADIA (Association for Computer Aided Design in Architecture) conference gave attendees a glimpse of potentially disruptive technologies and workflows for computational architectural production. The conference was held this year in Austin from October 24 through 26 and was organized by The University of Texas School of Architecture faculty members Kory Bieg, Danelle Briscoe, and Clay Odom. The organizers collected papers, workshops, and projects addressing the theme of “Ubiquity and Autonomy” in computation. Contributors reflected on the state of architectural production, in which digital tools and methodologies developed in the boutique, specialized settings at the fringes of the profession a generation ago have now become commonplace in architectural offices—while at the same time, new forms of specialist computational practices are emerging which may themselves soon become mainstream. While each participant grappled to position themselves in the cyclical and ever-advancing framework of technological inheritance and transference, the most encouraging efforts can be described in three categories: Expansions, subversions, and wholesale disruptions of the computational status quo. The expansionists claimed new technological territories, enlisting emerging and peripheral technologies to their purposes. The subvertors sampled the work and scrambled the workflows of their predecessors, configuring novel material applications in the process. Disruptors actively sought to break the techno-positivist cycle, questioning the assumptions, ethics, and values of previous generations to leverage computational design and digital processes to advance pressing and prescient political, economic, and ecological agendas. Expansionists appropriated bleeding-edge technologies, or those newly introduced to the discipline, to stake new terrain in design and construction. The conference was the first of its kind to host a dedicated session on the use of Generative Adversarial Networks (GANs) in design. This machine-learning system pits two forms of artificial intelligence against each other—one AI acts as the creative “artist,” generating all the possible solutions to a given task, while the other acts as the “critic,” selectively editing and curating the most appropriate responses. After training the networks on archives of architectural imagery, panelists put the GANs to work on evaluative and generative design tasks, alternately generating passably authentic floor plans, building envelopes, and reconstructed streetscapes. The workshop sessions, hosted by a suite of computational research teams from several architectural offices, demonstrated possibilities for adopting emerging technologies with familiar platforms, adopting and adapting tools like Fologram and Hololens to more familiar software platforms and fabrication methods. The subvertors, familiar with the expected uses and applications of given tools, would offer intentionally contradictory alternatives, short-circuiting established workflows and celebrating the unintended consequences of digitally enhanced platforms. A project from MIT researchers Lavender Tessmer, Yijiang Huang, and Caitlin Mueller entitled “Additive Casting of Mass-Customizable Brick” is a good example of the subvertors’ approach to interrogating workflows, enlisting precision-equipment for low-fidelity effect. As the current state-of-the-art in custom concrete formwork employs costly and time-consuming workflows to task CNC routers or robotic arms with milling, the MIT project is a critical alternative. Instead of shaping the mold, the project mobilizes the mold, achieving a wide variety of sculptural concrete “bricks” using standard cylindrical forms wielded by a robotic arm, while leveraging the ability of liquid concrete to self-level. The molds are shifted to preset positions while the concrete sets, allowing the sequential states of self-leveled concrete to intersect in complex geometries. The process is surprisingly delightful to watch, as the robot controls seven molds simultaneously like a drummer with a drumkit. The unexpected combination of high- and low-tech recalibrates possibilities for the robotic craft. Other researchers swapped out expected materials to produce unexpected results. Vasily Sitnikov (KTH) and Peter Eigenraam (TU Delft) teamed with BuroHappold to produce IceFormwork, a project that uses milled blocks of ice as the unlikely forms for casting high-performance fiber-reinforced concrete. Ice, the team argued, is a preferred, environmentally neutral alternative to industry-standard EPS foam molds, which produce a vast amount of waste. Ice molds, the team demonstrated, are easy enough to make (with some help from a reliable water source and a repurposed refrigerated ISO container). Airborne particles suspended by the ice-milling process are harmless water vapor, unlike the dangerous foam dust requiring ventilation equipment and other protective measures. When it comes to de-molding, the ice can simply be left outside to melt. While these investigations showcased new ways to hack the assembly process of cast building elements, their choice of concrete as a material contradicted a growing consensus in the panels; that designers should actively seek alternatives to the glut of concrete in the building industry, given the high ecological cost and high carbon footprint of concrete manufacturing in the context of an accelerating global sand shortage. Daniela Mitterberger and Tiziano Derme (MAEID/University of Innsbruck) offered one of the more radical alternatives with their project “Soil 3D Printing.” The team is using hydrogels—non-toxic, biodegradable adhesives—as binding agents injected into loose soil, to form alien landscapes of networked, earthen structures that portend a near-future where biocompatible, organic additive manufacturing processes restructure geotechnical landscapes and planetary geology. The provocations of the disruptors—who radically repurpose computational tools beyond perceived disciplinary constraints—raised profound questions about the potential for design technologies to enable and enact larger societal transformations by lining up global supply chains, material economies, and non-human constituencies squarely in their sights. Jose Sanchez (Plethora Project/Bloom Games/USC), in the presentation he gave while accepting the Innovative Research Award, presented his work leveraging computation and game design to critically examine and transform economic and ecologic realities. Sanchez has developed a series of game environments which force players to navigate wicked problems in contemporary cities, to confront the complexities, contradictions, and paradoxes of urbanization, logistics, and manufacturing. Sanchez described the continued focus in his work on efforts to "optimize for the many"—as opposed to the few—in a period of increased economic inequality, re-assessing the predominant use of digital technologies over the past few decades to enable complex mass-customized assemblies. Sanchez, in his own work, and in projects like Bloom with Alisa Andrasek (Biothing/Bloom Games/RMIT), has been exploring the potential of digital technologies to disrupt mass-production models through high-volume production of serialized and standardized “discrete” architectural components. In a similar vein, Gilles Retsin (UCL/Bartlett) argued for a reconsideration of the labor practices and digital economies enmeshed in, and implicitly supported by,  a building industry that has not yet come to terms with automation. By focusing on the ability of digital tools to combat material waste, Retsin argued, a generation of digitally savvy architects have ignored the potential of automation to address wasted labor. Through speculative research and small projects, Retsin is hoping to disrupt the building industry, increasing the capacity of architects to design and implement new platforms for project delivery which can combat exploitative practices. As expansionists pointed out where to look for the next big advancement, subvertors demonstrated how existing tools could be used differently. Disruptors were some of the few to ask—and answer—why. Stephen Mueller is a founding partner of AGENCY and a Research Assistant Professor at Texas Tech University College of Architecture in El Paso.
Placeholder Alt Text

Roaming robot dogs could streamline jobsite documentation

Reality capture has revolutionized construction by increasing job site efficiency and safety and allowing for quick responses to design and building challenges. However, save for the use of drones, often operated by humans, on-the-ground monitoring has required the relatively traditional (and labor-intensive) task of walking around and taking photos and collecting data to feed into software. HoloBuilder, whose software helps builders document and analyze their underway projects, has partnered with the robotics firm Boston Dynamics to create a semi-autonomous solution to document under-construction projects. Using Boston Dynamics’ Spot, a dog-like robot that regularly goes semi-viral for its aerial acrobatics (and its more sinister uses, such as being put to work by the Massachusetts State Police), contractors can capture 360-degree overviews of their work and track changes throughout the build process. Controlled by the SpotWalk app, the robot is first semi-manually trained to walk its reality capture route via a user’s phone. Then, Spot learns to repeat the route on its own, avoiding obstacles and documenting the site consistently and regularly, creating documentation of the project over time. Contractor Hensel Phelps has been testing out Spot on the $1.2 billion San Francisco International Airport Terminal 1 project. A Spot unit walks through the site capturing imagery, which is then fed into HoloBuilder’s machine learning-powered SiteAI, which provides automated construction tracking and other data. Documenting construction sites currently is a tedious task that takes away time from project staff that could otherwise focus on other aspects of construction, safety, and design. It can only be done with relatively limited regularity because of the demands. With Spot, project managers predict that they could capture updates of their sites as frequently as twice a day with all the 360 imagery being automatically organized and analyzed. Because of Spot's greater consistency against humans, the photos are also more useful as tools and the collected data is more actionable due to its regularity.
Placeholder Alt Text

Jenny Sabin's installation for Microsoft responds to occupants' emotions

At Microsoft’s Redmond, Washington, campus, architect Jenny Sabin has helped realize a large-scale installation powered by artificial intelligence. Suspended from three points within an atrium, the two-story, 1,800-pound sculpture is a compressive mesh of 895 3D-printed nodes connected by fiberglass rods and arranged in hexagons along with fabric knit from photoluminescent yarn. Created as part of Microsoft’s artist-in-residence program, the project is named Ada, after Ada Lovelace, the English mathematician whose work on the analytical engine laid the groundwork for the invention of computer programming as we know it. Anonymized information is collected from microphones and cameras throughout the building. An AI platform designed by a team led by researcher Daniel McDuff processes this data to try to accurately sense people’s emotions based on visual and sonic cues, like facial movements and voice tone. This data is then synthesized and run through algorithms that create a shifting color gradient that Ada produces from an array of LEDs, fiber optics, and par (can) lights. “To my knowledge, this installation is the first architectural structure to be driven by artificial intelligence in real-time,” Sabin, Microsoft’s current artist in residence, told the company’s AI Blog. Microsoft touts Ada as an example of “embedded intelligence,” AI that’s built-in and responsive to our real-world environment. McDuff also hopes that his emotion tracking technology, as dystopian as it might sound, could have solutions in healthcare or other caregiving situations. (Microsoft employees are able to opt-out of individuated tracking and they assure that all identifying info is removed from the media collected).  Ada is part of a broader push to embed sensing and artificial intelligence into the built environment by Microsoft and many other companies, as well as artistic pavilions that grapple with the future of AI in our built world, like Refik Anadol's recent project at New York's ARTECHOUSE.
Placeholder Alt Text

Aesthetic of Prosthetics compares computer-enhanced design practices

How has contemporary architecture and culture been shaped by our access to digital tools, technologies, and computational devices? This was the central question of Aesthetics of Prosthetics, the Pratt Institute Department of Architecture’s first alumni-run exhibition curated by recent alumni and current students Ceren Arslan, Alican Taylan, Can Imamoglu and Irmak Ciftci. The exhibition, which closed last week, took place at Siegel Gallery in Brooklyn. The curatorial team, made up of current students and recent alumni, staged an open call for submissions that addressed the ubiquity of “prosthetic intelligence” in how we interact with and design the built environment. “We define prosthetic intelligence as any device or tool that enhances our mental environment as opposed to our physical environment," read the curatorial statement. "Here is the simplest everyday example: When at a restaurant with friends, you reach out to your smartphone to do an online search for a reference to further the conversation, you use prosthetic intelligence." As none of the works shown have actually been built, the pieces experimented with the possibilities for representation and fabrication that “prosthetic intelligence” allows. The selected submissions used a range of technologies and methods including photography, digital collage, AI technology, digital modeling, and virtual reality The abundant access to data and its role in shaping architecture and aesthetics was a pervasive theme among the show's participants. Ceren Arslan's Los Angeles, for instance, used photo collage and editing to compile internet-sourced images that create an imaginary, yet believable streetscape. Others speculated about data visualization when drawings are increasingly expected to be read by not only humans, but machines and AI intelligence, as in Brandon Wetzel's deep data drawing.

"The work shown at the exhibition, rather than serving as a speculative criticism pointing out towards a techno-fetishist paradigm, tries to act as recording device to capture a moment in architectural discourse. Both the excitement and skepticism around the presented methodologies are due to the fact that they are yet to come to fruition as built projects," said the curators in a statement. 

Placeholder Alt Text

URBAN-X 6 showcases new tech solutions at A/D/O

This past Thursday, URBAN-X hosted its sixth demo day in Brooklyn at A/D/O, where startups that were showing what Micah Kotch, the startup accelerator's managing director, called “novel solutions to urban life.” URBAN-X, which is organized by MINI, A/D/O’s founder, in partnership with the venture firm Urban Us, began incubating urban-focused startups back in 2016. Previous iterations have seen everything from electric vehicle companies to waste management startups, and for this session, the brief was intentionally broad, said Kotch. On display was everything from machine-learning solutions to building energy management to apps that let people buy leftover prepared food from fast-casual restaurants and cafes to prevent food waste and generate some extra revenue.  Pi-Lit showed off a networked solution to highway and infrastructural safety. Many lives are lost each year as people sit after accidents, or as construction workers operate in dangerous work zones. The California-based company has developed a smart solution of mesh-networked lighting that can be deployed by first responders or work on existing work zone infrastructure. In addition, they’ve developed an array of sensors that can be affixed to bridges, roads, and temporary barriers—which founder Jim Selevan says are prone to impact but without transportation departments being aware, leading to unknown compromises that can cause accidents later on. Sensors could also let relevant parties know if a bridge is vibrating too much, or when roads begin freezing and warnings need to be put out, providing users with “real-time ground truth.” 3AM also presented their plans for using mesh networks, with a focus on safety, as their program relies on drones and portable trackers to help support operational awareness for firefighters. More whimsically, Hubbster showcased their solution—already deployed in Paris and Copenhagen—to support urban play: basically an app-based rental system for basketballs, croquet set, and everything in between, which would deploy from small, battery-powered smart lockboxes. Less glamorously but quite critically, Varuna is trying to make a change in the old-fashioned U.S. water infrastructure system, which exposes as much as 63 percent of the country to unsafe water and largely relies on manual testing, even for federally mandated across-the-board chlorine monitoring. They hope that by introducing AI-equipped sensors to utility systems, U.S. water can be delivered more safely, efficiently, and cheaply, addressing "operational inefficiencies in water supply, outdated tools, and a lack of visibility.” Also working with utilities was the Houston-based Evolve Energy, whose AI behavioral classification solution, currently available in parts of Texas, allows electricity to be bought at wholesale prices at the times of day when it is cheapest, for the comfort and needs individual users value most. For example, a home can pre-cool with cheap electricity and then turn off when prices surge. Variable rates, a la airline tickets, were a common theme—for example, Food for All, an app that is designed to reduce food waste and create extra revenue for fast-casual restaurants, offers flexible pricing for customers to pick up food that might otherwise be tossed. Most relevant to architects, perhaps, were Cove.Tool’s recent updates. The startup reports that they’ve made big strides on their cloud-based app that helps architect’s create efficient buildings. Reportedly cutting down energy grading from tens of hours to mere minutes, the app can now simulate the effects of sunlight—through various types glass—on utility usage, among many other new micro-climatic simulation features.
Placeholder Alt Text

ARTECHOUSE's Chelsea Market space will let visitors experience architectural hallucinations

ARTECHOUSE, a technology-focused art exhibition platform conceived in 2015 by Sandro Kereselidze and Tati Pastukhova, has been presenting digitally inspired art in Washington D.C. and Miami. Now they’re coming to New York, “a clear next step for [their] mission,” with an inaugural exhibition by Refik Anadol. The Istanbul-born, Los Angeles-based Anadol is known for his light and projection installations that often have an architectural component, such as the recent animation projected on the facade of the Frank Gehry-designed Walt Disney Concert Hall. For ARTECHOUSE in New York (also Anadol’s first large exhibition in New York),  he’ll be presenting Machine Hallucination. The installation will create what he calls “architectural hallucinations” that are derived from millions of images processed by artificial intelligence and machine learning algorithms. “With Refik, it’s been a collaborative process for over a year and a half, bringing a new commission, Machine Hallucination to life,” explained Kereselidze and Pastukhova. “We have worked closely with Refik to develop the concept for this exciting new work, thinking carefully about how to most effectively utilize and explore our Chelsea Market space.” ARTECHOUSE is especially suited to visualizing Refik’s “data universe” with a floor-to-ceiling, room-wrapping 16K laser projector that the creators claim features “the largest seamless megapixel count in the world,” along with 32-channel sound from L-ISA. The more than 3 million photos, representing numerous architectural styles and movements, will be made to expose (or generate) latent connections between these representations of architectural history, generating “hallucinations” that challenge our notions of space and how we experience it—and providing insight into how machines might experience space themselves. It makes us consider what happens when architecture becomes information. Of the work, Anadol said, “By employing machine intelligence to help narrate the hybrid relationship between architecture and our perception of time and space, Machine Hallucination offers the audience a glimpse into the future of architecture itself.” Machine Hallucination will inhabit the new 6,000-square-foot ARTECHOUSE space in Chelsea Market, located in an over-century-old former boiler room which features exposed brick walls and a refurbished terracotta ceiling, which according to its creators, “supplies each artist with a unique canvas and the ability to drive narratives connecting the old and new.” ARTECHOUSE will be opening to the public early next month.
Placeholder Alt Text

How can new technologies make construction safer?

Construction remains one of the most dangerous careers in the United States. To stop accidents before they happen, construction companies are turning to emerging technologies to improve workplace safety—from virtual reality, drone photography, IoT-connected tools, and machine learning. That said, some solutions come with the looming specter of workplace surveillance in the name of safety, with all of the Black Mirror-esque possibilities. The Boston-based construction company Suffolk has turned to artificial intelligence to try and make construction safer. Suffolk has been collaborating with computer vision company Smartvid.io to create a digital watchdog of sorts that uses a deep-learning algorithm and workplace images to flag dangerous situations and workers engaging in hazardous behavior, like failing to wear safety equipment or working too close to machinery. Suffolk’s even managed to get some of their smaller competitors to join them in data sharing, a mutually beneficial arrangement since machine learning systems require so much example data; something that's harder for smaller operations to gather. Suffolk hopes to use this decade’s worth of aggregated information, as well as scheduling data, reports, and info from IoT sensors to create predictive algorithms that will help prevent injuries and accidents before they happen and increase productivity. Newer startups are also entering the AEC AI fray, including three supported by URBAN-X. The bi-coastal Versatile Natures is billing itself as the "world's first onsite data-provider," aiming to transform construction sites with sensors that allow managers to proactively make decisions. Buildstream is embedding equipment and construction machinery to make them communicative, and, by focusing on people instead, Contextere is claiming that their use of the IoT will connect different members of the workforce. At the Florida-based firm Haskell, instead of just using surveillance on the job site, they’re addressing the problem before construction workers even get into the field. While videos and quizzes are one way to train employees, Haskell saw the potential for interactive technologies to really boost employee training in a safe context, using virtual reality. In the search for VR systems that might suit their needs, Haskell discovered no extant solutions were well-suited to the particulars of construction. Along with their venture capital spinoff, Dysruptek, they partnered with software engineering and game design students at Kennesaw State University in Georgia to develop the Hazard Elimination/Risk Oversight program, or HERO, relying on software like Revit and Unity. The video game-like program places users into a job site, derived from images taken by drone and 360-degree cameras at a Florida wastewater treatment plant that Haskell built, and evaluates a trainee’s performance and ability to follow safety protocols in an ever-changing environment. At the Skanska USA, where 360-degree photography, laser scanning, drones, and even virtual reality are becoming increasingly commonplace, employees are realizing the potentials of these new technologies not just for improved efficiency and accuracy in design and construction, but for overall job site safety. Albert Zulps, Skanska’s Regional Director, Virtual Design and Construction, says that the tech goes beyond BIM and design uses, and actively helps avoid accidents. “Having models and being able to plan virtually and communicate is really important,” Zulps explained, noting that in AEC industries, BIM and models are now pretty much universally trusted, but the increased accuracy of capture technologies is making them even more accurate—adapting them to not just predictions, but the realities of the site. “For safety, you can use those models to really clearly plan your daily tasks. You build virtually before you actually build, and then foresee some of the things you might not have if you didn't have that luxury.” Like Suffolk, Skanska has partnered with Smartvid.io to help them process data. As technology continues to evolve, the ever-growing construction industry will hopefully be not just more cost-efficient, but safer overall.