CMS Upcoming Seminars CMS Upcoming Seminar Feed en Inaugural IDEAS Lecture: Point of Care, Everywhere! () Inaugural IDEAS Lecture <strong>Speaker(s):</strong> Axel Scherer (Caltech)<br><strong>Location:</strong> Online Event<br><p><b>Abstract:</b></p><p>Health monitoring today is based on observing symptoms or extracting samples from patients. After samples are taken, these are tested in complex and expensive instruments, usually located in centralized laboratories, by skilled technicians. This system has evolved into an industry of highly specialized and heavily regulated laboratories, but suffers from delays and logistical challenges resulting from the need to transport perishable samples. The problems with our centralized diagnostic infrastructure has recently become apparent, as it was overburdened by the Covid-19 pandemic. We will outline an alternative, decentralized, point of care vision, and describe the technology required for field detection of pathogens by qPCR tests. Ultimately, our goal is to provide continuous patient monitoring for patients with chronic diseases, and immediate test results for patients with infectious diseases. Through a combination of automation and miniaturization, individual health monitoring is within our grasp, and holds the promise of early and even pre-symptomatic detection of both infectious and chronic diseases.</p><p><b>Biosketch:</b></p><p>Axel Scherer is the Bernard A. Neches Professor of Electrical Engineering, Medical Engineering, Applied Physics and Physics at Caltech. He received his PhD in 1985, and worked in the Microstructures Research Group at Bellcore on optoelectronic miniaturization. He joined the Electrical Engineering faculty at Caltech in 1993, and established a group that develops micro- and nanofabricated optical, electronic and fluidic devices and integrates these into microsystems. He has co-authored over 350 publications and holds over 120 patents in the fields of optoelectronics, microfluidics, and nanofabrication techniques. Professor Scherer has co-founded several high-technology companies pioneering silicon photonics and medical diagnostics. He helped establish a state of the art cleanroom at Caltech, and pioneered microcavity lasers, such as vertical cavity surface emitting lasers, and integrated photonic systems. His group developed silicon nanophotonics and surface plasmon enhanced light emitters, and has perfected the fabrication and characterization of ultra-small structures with sizes down to 2nm. His group is currently focused on integrating nanoscale vacuum tubes for high frequency electronics, and works on the application of wireless chips with electrochemical and metabolic sensors for the purpose of building injectable health monitors. Over the past 10 years, the goal of this group has been to build inexpensive medical diagnostic tools that can provide immediate feedback for patients in point of care settings. Professor Scherer's group has developed automated instruments for clinical pathology as well as wireless solutions for chronic diseases and metabolic health monitoring.</p><p>Please register to join: <a ></a></p> Mon, 07 Dec 2020 17:00:00 -0800 CMX Lunch Seminar: Learning Sparse Non-Gaussian Graphical Models (Jolene Brink) CMX Lunch Seminar <strong>Speaker(s):</strong> Rebecca Morrison (University of Colorado Boulder)<br><strong>Location:</strong> Online Event<br><p> Identification and exploitation of a sparse undirected graphical model (UGM) can simplify inference and prediction processes, illuminate previously unknown variable relationships, and even decouple multi-domain computational models. In the continuous realm, the UGM corresponding to a Gaussian data set is equivalent to the non-zero entries of the inverse covariance matrix. However, this correspondence no longer holds when the data is non-Gaussian. In this talk, we explore a recently developed algorithm called SING (Sparsity Identification of Non-Gaussian distributions), which identifies edges using Hessian information of the log density. Various data sets are examined, with sometimes surprising results about the nature of non-Gaussianity. </p> Wed, 09 Dec 2020 12:00:00 -0800 Ulric B. and Evelyn L. Bray Social Sciences Seminar: Topic to be announced (Letty Diaz) Ulric B. and Evelyn L. Bray Social Sciences Seminar <strong>Speaker(s):</strong> Konstantin Sonin (University of Chicago)<br><strong>Location:</strong> Online Event<br><p>Please check later for additional details</p><p><b>How to view the seminar:</b><br>Sign up for a free <a ></a> account, and tune in on Wednesdays at noon pacific time on <a ></a>. You will be able to ask questions on the twitch chat.</p> Wed, 09 Dec 2020 12:00:00 -0800 Mechanical and Civil Engineering Seminar: Microstructure-Enabled Plasticity in Nano-to-Microscale Materials (Holly Golcher) Mechanical and Civil Engineering Seminar <strong>Speaker(s):</strong> Haolu (Jane) Zhang (Caltech)<br><strong>Location:</strong> Online Event<br><p>Haolu (Jane) Zhang, Mechanical Engineering Graduate Student<br>PhD Thesis Defense<br><br>Abstract:<br>Microstructure-governed damage resistance in materials enables a variety of functional applications, such as durable biomedical implants and robust product packaging. As industrial applications continue to shrink in size, it became increasingly apparent that when a material's external structural size and internal microstructural size become comparable, its mechanical behavior starts to deviate from that of bulk, such is the smaller-is-stronger size-effect in metals. This elucidation necessitates the characterization of materials at lengthscales relevant to their internal microstructure to guarantee accuracy in the design of real-world applications.</p><p>Our work aims at deciphering the microstructure-mechanics relationship for materials at lengthscales bridging the gap between the atomic level and macroscale, with shape memory ceramics, scorpion shells, and jellyfish biogel as sample systems. We use electron and x-ray diffraction to characterize microstructures such as twinning, defects, and fiber organization, while revealing strength, toughness, and other deformation mechanisms through <i>in-situ</i> nanomechanical experiments. We show improved shape recovery in an otherwise brittle ceramic by tuning its phase compatibility at the nanoscale and reveal unprecedented smaller-is-stronger size-dependence for its twinning-induced plasticity. We then unveil competing fiber orientations in Scorpion shells that follow fiber-mechanics principles and demonstrate combined poroelasticity/viscoelasticity constitutive relation in Jellyfish that explains their self-healing behavior. The correlation between microstructure and mechanical behavior unveil unique damage mitigation and energy dissipation techniques in both brittle ceramics and natural biomaterials at each order of lengthscale, paving the road to designing macroscopic materials with hierarchical mechanical behavior and improved plasticity.</p><p>Please attend this Defense virtually:</p><p><a ></a><br>Meeting ID: 828 2304 4981<br>Passcode: 064028</p> Thu, 10 Dec 2020 15:00:00 -0800 Mechanical and Civil Engineering Seminar: "Multiscale, data-driven and nonlocal modeling of granular materials" (Jenni Campbell) Mechanical and Civil Engineering Seminar <strong>Speaker(s):</strong> Konstantinos Karapiperis (Caltech)<br><strong>Location:</strong> Online Event<br><p></p><p><b>PhD Thesis Defense</b></p><p><b>Abstract:</b> Granular materials are ubiquitous in both nature and technology. They play a key role in applications ranging from storing food and energy, to building reusable habitats and soft robots. Yet, predicting the continuum mechanical response of granular materials continues to present extraordinary challenges, despite the apparently simple laws that govern particle-scale interactions. This is largely due to the complex history dependence arising from the continuous rearrangement of their internal structure, and the nonlocality emerging from their self-organization. There is an urge to develop methods that adequately address these two aspects, while bridging the long-standing divide between the grain- and the continuum scale.</p><p>This dissertation introduces novel theoretical and computational approaches for behavior prediction in granular solids. To begin with, we develop a framework for investigating their incremental behavior from the perspective of plasticity theory. It relies on systematically probing, through level-set discrete element calculations, the response of granular assemblies from the same initial state to multiple directions is stress space. We then extract the state- and history-dependent elasticity and plastic flow, and investigate the evolution of pertinent internal variables. Next, inspired by the abundance of generated high-fidelity micromechanical data, we develop an alternative data-driven approach for behavior prediction. This new multiscale modeling paradigm completely bypasses the need to define a constitutive law. Instead, the problem is directly formulated on a material data set, generated by grain-scale calculations, while pertinent constraints and conservation laws are enforced. We particularly focus on the sampling of the mechanical phase space, and develop two methods for parametrizing material history, one thermodynamically motivated and one statistically inspired. In the remainder of the thesis, we direct our attention to the understanding and modeling of nonlocality. In particular, we first delve into a complex network analysis of a granular assembly undergoing shear banding, and study the self-organized and cooperative evolution of topology, kinematics and kinetics within the band. We characterize the evolution of fundamental topological structures called force cycles, and propose a novel order parameter for the system, the minimal cycle coefficient. We also analyze the statistics of nonaffine kinematics, which involve rotational and vortical particle motion. Finally, inspired by these findings, we extend the previously introduced data-driven paradigm to include nonaffine kinematics within a weakly nonlocal micropolar continuum description. By formulating the problem on a phase space augmented by higher-order kinematics and their conjugate kinetics, we bypass for the first time the need to define an internal length scale, which is instead discovered from the data. By carrying out a data-driven prediction of shear banding, we find that this nonlocal extension of the framework resolves the ill-posedness inherent to the classical continuum description. Our theoretical developments are finally validated by comparing with available experimental data.</p><p></p><p>Please virtually attend this thesis defense:</p><p>Zoom Link: <a ></a></p> Tue, 15 Dec 2020 10:00:00 -0800 Mechanical and Civil Engineering Seminar: "Assuring Safety under Uncertainty in Learning-Based Control Systems" (Sonya Lincoln) Mechanical and Civil Engineering Seminar <strong>Speaker(s):</strong> Richard Cheng (Caltech)<br><strong>Location:</strong> Online Event<br><p></p><p><b>PhD Thesis Defense</b></p><p><b>Abstract:</b> Learning-based controllers have recently shown impressive results in well-defined environments, promising to enable a host of new capabilities for complex robotic systems. However, such controllers cannot be widely deployed in highly uncertain environments due to significant issues relating to learning reliability and safety.</p><p>The first half of the talk will focus on reinforcement learning, and discuss why the integration of model information into the reinforcement learning framework is crucial to ensure reliability and safety. I will discuss how such model information can be leveraged to constrain/guide exploration and provide explicit safety guarantees for uncertain systems.</p><p>The second half of the talk will discuss fundamental limitations that arise when utilizing machine learning to derive safety guarantees. In particular, I will show that widely used uncertainty models can be highly inaccurate when predicting rare events, and examine the implications of this for safe learning. To overcome some of these limitations, I propose a novel approach based on assume-guarantee contracts to ensure safety in human environments.</p><p></p><p>Please virtually attend this thesis defense:</p><p>Zoom Link: <a ></a></p> Tue, 15 Dec 2020 15:00:00 -0800 Diverse Minds Seminar: The Hierarchy of Knowledge in Machine Learning & Related Fields and Its Consequences (Jamie Meighen-Sei) Diverse Minds Seminar <strong>Speaker(s):</strong> Dr. Timnit Gebru (Google)<br><strong>Location:</strong> Online Event<br><p>Feminist and race &amp; gender scholars have long critiqued "the view from nowhere" that assumes science is "objective" and studied from no particular standpoint. In this talk, I discuss how this view has resulted in a hierarchy of knowledge in machine learning and related fields, devaluing some types of work and knowledge (e.g. those related to data production, annotation and collection practices) and mostly amplifying specific types of contributions. This hierarchy also results in valuing contributions from some disciplines (e.g. Physics) more than others (e.g. race and gender studies). With examples from my own life, education and current work, I discuss how this knowledge hierarchy limits the field and potential ways forward.</p><p><i>Due to the pandemic this workshop will be available via Zoom Webinar. The workshop will be open to the Caltech community. URL and registration forthcoming.</i></p><a >The Diverse Minds Seminar Series</a> is a campus-wide program that aims to provide a platform for distinguished speakers from a diversity of backgrounds to share their science and personal journey into science with the Caltech Community. These seminars aim to educate, inspire, and initiate purposeful discussion. Caltech recognizes that diverse perspectives and experiences benefit everyone and by intentionally inviting speakers from underrepresented and underserved backgrounds Caltech can continue to cultivate a community that values equity, inclusion, and embraces the power of diversity in STEM. Wed, 06 Jan 2021 16:00:00 -0800 Control Meets Learning Seminar: The Provable Effectiveness of Policy Gradient Methods in Reinforcement Learning and Controls (Jolene Brink) Control Meets Learning Seminar <strong>Speaker(s):</strong> Sham Kakade (University of Washington)<br><strong>Location:</strong> Online Event<br><p> </p><p>Reinforcement learning is the dominant paradigm for how an agent learns to interact with the world in order to achieve some long term objectives. Here, policy gradient methods are among the most effective methods in challenging reinforcement learning problems, due to that they: are applicable to any differentiable policy parameterization; admit easy extensions to function approximation; easily incorporate structured state and action spaces; are easy to implement in a simulation based, model-free manner.</p><p>However, little is known about even their most basic theoretical convergence properties, including:</p><p>- do they converge to a globally optimal solution, say with a sufficiently rich policy class?</p><p>- how well do they cope with approximation error, say due to using a class of neural policies?</p><p>- what is their finite sample complexity?</p><p>This talk will cover a number of recent results on these basic questions and also provide the first approximation results which do have not worst case dependencies on the size of the state space. We will highlight the interplay of theory, algorithm design, and practice.</p><p>Joint work with: Alekh Agarwal, Jason Lee, Gaurav Mahajan</p><p></p> Wed, 13 Jan 2021 09:00:00 -0800 Ulric B. and Evelyn L. Bray Social Sciences Seminar: Topic to be announced (Letty Diaz) Ulric B. and Evelyn L. Bray Social Sciences Seminar <strong>Speaker(s):</strong> Luciano Pomatto (Caltech)<br><strong>Location:</strong> Online Event<br><p>Please check later for additional details</p><p><b>How to view the seminar:</b><br>Sign up for a free <a ></a> account, and tune in on Wednesdays at noon pacific time on <a ></a>. You will be able to ask questions on the twitch chat.</p> Wed, 13 Jan 2021 12:00:00 -0800 Watson Lecture - Artificial Intelligence: How it Works and What it Means for the Future: TBD (The Caltech Ticket Office) Watson Lecture - Artificial Intelligence: How it Works and What it Means for the Future <strong>Speaker(s):</strong> Yisong Yue (Caltech)<br><strong>Location:</strong> Online Event<br><p>Over the past decade, artificial intelligence and the massive amounts of data powering such systems have dramatically changed our world. And as both the technology and the way in which scientists and engineers handle it becomes more refined, the impact of AI in society will become more profound. In this lecture, Yue will explore the key principles powering the current revolution in AI, consider how cutting-edge AI techniques are transforming the way research is done across science and engineering at Caltech, and what it means for the future of material design, robotics, and big data seismology, among other areas of investigation. Yue will show how, where human intuition breaks down, AI can guide scientists in finding data-driven solutions to complex problems.<br><br> <i>This Zoom webinar is free and open to the public. Advance registration is required.</i></p><p><a ><b>? REGISTER NOW</b></a><i><br><br>Each Watson Lecture will begin at 5:00 p.m. Pacific Time as a Zoom Webinar</i> <i>with live audience Q&amp;A at the end. Please note the new start time. At 8 p.m. Pacific Time the recorded lecture (without Q&amp;A) will be posted on</i> <a s YouTube channel</i></a><i>.</i><br></p><p><b>About the Speaker</b></p><p><a >Yisong Yue</a> is Professor of Computing and Mathematical Sciences at Caltech.<br><br> <b>About the Series</b></p><p>Since 1922, The Earnest C. Watson Lecture Series has brought Caltech's most innovative scientific research to the public. The series is named for Earnest C. Watson, a professor of physics at Caltech from 1919 until 1959. Spotlighting a small selection of the pioneering research Caltech's faculty is currently conducting, the Watson Lectures are geared toward a general audience, as part of the Institute's ongoing commitment to benefiting the local community through education and outreach. Through a gift from the estate of Richard C. Biedebach, the lecture series is able to highlight assistant professors' research each season.</p><p>Many past Watson Lectures are available in a <a >playlist on YouTube</a>.</p> Wed, 13 Jan 2021 17:00:00 -0800 Mechanical and Civil Engineering Seminar: TBA (Mikaela Laite) Mechanical and Civil Engineering Seminar <strong>Speaker(s):</strong> Nathalie Vriend (University of Cambridge)<br><strong>Location:</strong> Online Event<br><p>TBA</p> Thu, 14 Jan 2021 11:00:00 -0800 CMX Lunch Seminar: TBA (Jolene Brink) CMX Lunch Seminar <strong>Speaker(s):</strong> William Shadwick (Omega Analysis Limited)<br><strong>Location:</strong> Online Event<br><p></p> Wed, 20 Jan 2021 12:00:00 -0800 Mechanical and Civil Engineering Seminar: TBA (Mikaela Laite) Mechanical and Civil Engineering Seminar <strong>Speaker(s):</strong> Laura De Lorenzis (Eidgen?ssische Technische Hochschule Zürich)<br><strong>Location:</strong> Online Event<br><p>TBA</p> Thu, 21 Jan 2021 11:00:00 -0800 CMX Lunch Seminar: Modeling and optimizing set functions via RKHS embeddings (Jolene Brink) CMX Lunch Seminar <strong>Speaker(s):</strong> David Ginsbourger (University of Bern)<br><strong>Location:</strong> Online Event<br><p> We consider the issue of modeling and optimizing set functions, with a main focus on kernel methods for expensive objective functions taking finite sets as inputs. Based on recent developments on embeddings of probability distributions in Reproducing Kernel Hilbert Spaces, we explore adaptations of Gaussian Process modeling and Bayesian Optimization to the framework of interest. In particular, combining RKHS embeddings and positive definite kernels on Hilbert spaces delivers a promising class of kernels, as illustrated in particular on two test cases from mechanical engineering and contaminant source localization, respectively. Based on several collaborations and notably on the paper "Kernels over sets of finite sets using RKHS embeddings, with application to Bayesian (combinatorial) optimization" with Poompol Buathong and Tipaluck Krityakierne (AISTATS 2020). ?? </p> Wed, 27 Jan 2021 12:00:00 -0800 Ulric B. and Evelyn L. Bray Social Sciences Seminar: Topic to be announced (Letty Diaz) Ulric B. and Evelyn L. Bray Social Sciences Seminar <strong>Speaker(s):</strong> Teddy Mekonnen (Brown University)<br><strong>Location:</strong> Online Event<br><p>Please check later for additional details</p><p><b>How to view the seminar:</b><br>Sign up for a free <a ></a> account, and tune in on Wednesdays at noon pacific time on <a ></a>. You will be able to ask questions on the twitch chat.</p> Wed, 03 Feb 2021 12:00:00 -0800 CMX Lunch Seminar: TBA (Jolene Brink) CMX Lunch Seminar <strong>Speaker(s):</strong> TBD ()<br><strong>Location:</strong> Online Event<br><p></p> Wed, 17 Feb 2021 12:00:00 -0800 CMX Lunch Seminar: TBA (Jolene Brink) CMX Lunch Seminar <strong>Speaker(s):</strong> TBD ()<br><strong>Location:</strong> Online Event<br><p></p> Wed, 24 Feb 2021 12:00:00 -0800 H.B. Keller Colloquium: TBA (Diana Bohler) H.B. Keller Colloquium <strong>Speaker(s):</strong> Amy Ko (University of Washington, Seattle)<br><strong>Location:</strong> Online Event<br><p>TBA</p> Mon, 01 Mar 2021 16:00:00 -0800 CMX Lunch Seminar: TBA (Jolene Brink) CMX Lunch Seminar <strong>Speaker(s):</strong> TBD ()<br><strong>Location:</strong> Online Event<br><p></p> Wed, 21 Apr 2021 12:00:00 -0700 CMX Lunch Seminar: TBA (Jolene Brink) CMX Lunch Seminar <strong>Speaker(s):</strong> TBD ()<br><strong>Location:</strong> Online Event<br><p></p> Wed, 28 Apr 2021 12:00:00 -0700 十大老品牌网赌信誉平台