February 26, 2025 - July 17, 2025Hidden within MIT’s Distinctive Collections, many architectural elements from the earliest days of the Institute’s architecture program still survive as part of the Rotch Art Collection. Among the artworks that conservators salvaged was a set of striking windows of gypsum and stained-glass, dating to the late 18th- to 19th c. Ottoman Empire. This exhibition illuminates the life of these historic windows, tracing their refracted histories from Egypt to MIT, their ongoing conservation, and the cutting-edge research they still prompt.The Maihaugen Gallery (14N-130) is open Monday through Thursday, 10am - 4pm, excluding Institute holidays.

Speaker: Mo ChenTitle: New Regimes for Topology Optimization in PhotonicsZoom: https://mit.zoom.us/j/91981902244

Zoom link: https://mit.zoom.us/j/92562397534----- Speaker: Gasser ElbannaAffiliation: McDermott LabTitle: Modeling Continuous Speech Recognition to Understand Contextual Effects in Human Speech PerceptionAbstract: Humans excel at transforming acoustic waveforms into meaningful linguistic representations, despite the inherent variability of speech signals. The perceptual and neural mechanisms that enable such robust perception remain unclear. Progress has been limited by the lack of 1) stimulus-computable models that replicate human behavior and 2) large-scale behavioral benchmarks for comparing model and human speech perception. I will present our work on developing candidate models of continuous speech perception along with new behavioral experiments to compare phonemic judgments in humans and models. Our models reproduce patterns of human responses and confusions, and by manipulating the model’s access to past and future speech input, we are testing the role of context in shaping human speech perception.Bio: Gasser is a second-year PhD student in the Speech and Hearing, Bioscience and Technology (SHBT) program at Harvard University. He works with Josh McDermott at the Laboratory for Computational Audition at MIT. His research aims to understand how the brain dynamically perceives, encodes, and integrates speech information over time, thereby unraveling the perceptual and neural foundations of auditory intelligence.

Tuesday 12:00-1:00 EDT, April 22, 2025 live stream via Zoom: Registration requiredAbstract:Transformers are the dominant architecture for language modeling (and generative AI more broadly). The attention mechanism in Transformers is considered core to the architecture and enables accurate sequence modeling at scale. However, the complexity of attention is quadratic in input length, which makes it difficult to apply Transformers to model long sequences. Moreover, Transformers have theoretical limitations when it comes to the class of problems it can solve, which prevents their being able to model certain kinds of phenomena such as state tracking. This talk will describe some recent work on efficient alternatives to Transformers which can overcome these limitations.Bio:Yoon Kim is an assistant professor at MIT EECS and a principal investigator at CSAIL, where he works on natural language processing and machine learning. He obtained his Ph.D. in computer science from Harvard University.

"The Right Problem to Solve" with Christine Taylor-ButlerAbout the presentation: A recent survey shows the US ranks 36th in global literacy. While serving on MIT's Educational Council, Christine saw a growing number of urban and rural students entering 12th grade without the appropriate skills. She wondered: what if you took literacy and STEAM concepts and embedded them in an epic adventure? And what if the characters were different than those seen in traditional literature? Ten years later, the Lost Tribes was published. The first books were used during an MIT summer middle school program. In addition, one librarian reported a student whose reading score doubled. Come learn how an MIT engineer switched gears mid-career to help grow a new generation of independent readers.Christine is completing her second year as an MLK Scholar sponsored by the Department of Anthropology.This event is hybrid. Please choose your ticket accordingly.We are committed to making this event fully accessible to everyone who wants to attend. Please let us know if there is anything you need to participate fully in this event by e-mailing vulfp@mit.edu.Photographs and/or videos may be taken at this event.By entering and attending this event, you acknowledge and agree that your likeness and/or voice may be included in photos and videos of the event and used by MIT in connection with communications about the Institute Community and Equity Office or in other MIT communications.If you do not agree to this usage, please notify the event organizer or do not enter the event.

Have some fun(nel) on a tunnel or sun(nel) walk! Join us for a 30-minute volunteer-led walk either through MIT’s famous tunnel system or around Killian Court. As the weather gets warmer, walk leaders may choose to take the group outside. Is the weather warm and you missed the start? Find the group on Killian Court and join in!Sun(nel) Walk Leaders will identify themselves by holding a white flag at the meeting location.Location details: Meet in the atrium by the staircase. [See image below]Prize Drawing: Attend a walk and scan a QR code from the walk leaders to be entered into a drawing for a getfit canvas boat tote bag at the end of the getfit challenge. The more walks you attend, the more entries you get. Winner will be drawn and notified at the end of April. Winner does not need to be a getfit participant.Disclaimer: Tunnel walks are led by volunteers. In the rare occasion when a volunteer isn’t able to make it, we will do our best to notify participants. In the event we are unable to notify participants and a walk leader does not show up, we encourage you to walk as much as you feel comfortable doing so. We recommend checking this calendar just before you head out.Getfit is a 12-week fitness challenge for the entire MIT community. These tunnel walks are open to the entire MIT community and you do not need to be a current getfit participant to join.