Location: 46-3310Zoom: https://mit.zoom.us/j/92495348437Speaker: Tamar RegevAffiliation: Fedorenko lab​​​​​​​Title: Characterizing the relationship between language and prosody using neural and computational approachesAbstract: Although much past work on human communication has focused on the verbal component of language, communication also relies on a variety of non-verbal cues, including visual cues, such as facial expressions and gestures, and auditory cues, such as non-verbal vocalizations and prosody. In the recent surge of interest in multimodal communication, these different cues have often been discussed on par with one another. However, I will here argue and provide empirical evidence that prosody occupies a privileged position among non-verbal communication signals, showing a uniquely tight relationship with language processing. First, I will present an fMRI investigation of brain areas that process prosody, revealing that prosody brain areas are similar in topography to, and show partial overlap with, language brain areas. This overlap is in sharp contrast to past work that has established that facial expressions, gestures, and non-verbal vocalizations are supported by brain areas distinct from the language areas. Second, I will describe a set of computational studies using information theory and large language models to quantify the overlap between prosodic and linguistic information in natural spoken language. The findings suggest that prosodic information is largely redundant with linguistic content. The overlap between prosodic and linguistic processing in the brain and in the information they carry suggests that prosodic and linguistic representations are tightly coupled within the cognitive architecture of human communication. I will speculate on why this might be, including the distributional properties of prosody vs. other cues during communication, the role of prosody in linguistic prediction, and the scaffolding that prosody may provide during early language learning.

Please join us for the CSAIL Forum with Alison GopnikCSAIL Forum hosted by Daniela Rus Title: Empowerment Gain as Causal Learning, Causal Learning as Empowerment Gain: A bridge between Bayesian causal hypothesis testing and reinforcement learning Speaker: Alison Gopnik Dept. of Psychology, UC Berkeley Date/time: Tuesday 12:00-1:00 EDT, November 4, 2025 Venue: Live stream via Zoom: Registration required Bio: https://psychology.berkeley.edu/people/alison-gopnik Abstract: Learning about the causal structure of the world is a fundamental problem for human cognition, and causal knowledge is central to both intuitive and scientific world models. However, causal models and especially causal learning have proved to be difficult for standard Large Models using standard techniques of deep learning. In contrast, cognitive scientists have applied advances in our formal understanding of causation in computer science, particularly within the Causal Bayes Net formalism, to understand human causal learning. These approaches also face challenges when it comes to learning however. In parallel, in the very different tradition of reinforcement learning, researchers have developed the idea of an intrinsic reward signal called “empowerment”. An agent is rewarded for maximizing the mutual information between its actions and their outcomes, regardless of the external reward value of those outcomes. In other words, the agent is rewarded if variation in an action systematically leads to parallel variation in an outcome so that variation in the action predicts variation in the outcome. Empowerment, then has two dimensions , it involves both controllability and variability. The result is an agent that has maximal control over the maximal part of its environment. “Empowerment” may be an important bridge between classical Bayesian causal learning and reinforcement learning and may help to characterize causal learning in humans and enable it in machines. If an agent learns an accurate causal model of the world they will necessarily increase their empowerment, and, vice versa, increasing empowerment will lead to a more accurate (if implicit) causal model of the world. Empowerment may also explain distinctive empirical features of children’s causal learning, as well as providing a more tractable computational account of how that learning is possible.

Speakers: Robin Pemantle (University of Pennsylvania)Title: Course development and active learning: a retrospective spanning nine coursesAbstract: The long term success of a new course depends on a lot of factors: personnel to teach the course, student incentives to take the course and learning goals (careful, different interested parties are likely to see these differently).I will tell some stories of course development spanning courses for pre-service teachers, calculus courses, and graduate level applied mathematics. Questions that will arise include whether to tweak an existing course versus re-invent from scratch and what pedagogies to build into the course. For clarity on the issue of pedagogy, one needs a solid understanding of the factors above, particularly the learning goals.Zoom link: https://cornell.zoom.us/j/92415199317Zoom Link Password: olsumeFor more information on OLSUME: https://olsume.org/

Navigate the future of technology with confidenceFor over 25 years, EmTech MIT has been the trusted destination for established senior executives and emerging leaders, researchers, and entrepreneurs to stay ahead of change. Curated by the expert editors of MIT Technology Review, our flagship technology event delivers the clarity and insight you need to navigate uncertainty and lead with conviction.Join us on November 4-6 at the MIT Media Lab for EmTech MIT 2025, MIT Technology Review’s flagship event on transformative technology for business leaders.Learn more and register: emtechmit.com.Contact MIT Technology Review with any questions and discount opportunities.**Discounts are available to the MIT community. Register here with your MIT email address and save 40%.

"Leadership, Gender, and Workplace Climate in Large Corporations" | Sule Alan (Cornell)

Speaker: Leonid Berlyand (Penn State University)Title: Stability and Bifurcations in Free Boundary PDE Models of Cell Motility.Abstract:We begin with a brief overview of the rapidly developing research area of active matter, a.k.a. active materials. These materials are intrinsically out of equilibrium, resulting in novel physical properties whose modeling requires the development of new mathematical tools. We present a free boundary PDE model of the cytoskeleton of a moving cell. The key mathematical features of our model are the nonlocal boundary conditions, nonlinear diffusion, and the Keller-Segel cross-diffusion term. We present an overview of three recent works on that model. We begin from the 2D model with linear diffusion, in which we derive an explicit formula for the stability-determining eigenvalue for the linearized non-self-adjoint operator. Next, we present a recent result on the nonlinear stability of stationary and traveling wave solutions in a 1D model. Here, we focus on the non-self-adjointness of the linearized problem, which plays a key role in the spectral stability analysis. Finally, we consider a 2D model with nonlinear diffusion and prove that this nonlinearity results in the change of the bifurcation from supercritical to subcritical, leading to two drastically different scenarios of the onset of the cell motion. Here, we derive an explicit formula that governs the change of the bifurcation type in terms of measurable physical parameters and therefore can be used for both qualitative and quantitative biological predictions. Finally, we discuss how our results lead to an open question of bistability. This work resulted in two published papers with A. Safsten and V. Rybalko ( Phys. Rev . E, 2022, Transactions of AMS, 2023), as well as recent papers with A. Safsten and L. Truskinovsky (ARMA, 2025) and with O. Krupchytski and T. Laux (Nonlinear Science, 2025, accepted subject to revision)Biography:Leonid Berlyand is a Professor of Mathematics and a member of the Materials Research Institute and the Huck Institute of Life Sciences at Penn State University. His research is at the interface between mathematics and other disciplines: physics, life sciences, and computer science. In 2021 he received Humboldt Research Award. Leonid is a founding co-director of PSU Center for Mathematics of Living and Mimetic Matter and PSU Center for Interdisciplinary Mathematics. He received his Ph.D. degree from Kharkiv State University (Ukraine) and co-authored three books and over 100 research papers.