Electrical and Systems Engineering
Wednesday, October 2nd, 2019 Raisler Lounge
Meta-Learning through Coupled Optimization in Reproducing Kernel Hilbert Spaces
Bio: Juan received received the B.Sc. degree in electrical engineering from the Universidad de la República Oriental del Uruguay, Montevideo, in 2018. He is now a PhD student in the Department of Electrical and Systems Engineering at the University of Pennsylvania, supervised by Professor Alejandro Ribeiro. Juan’s current research interests are in machine learning, optimization and control.
Abstract: In this talk we consider the problem of meta-learning, consisting of building policies that achieve good generalization performance and adapt quickly to different tasks. We introduce meta-learning through the coupled optimization of a set of rewards that are defined for different tasks. This coupling is effected by a projection step that brings the task-specific policies close to a central one which combines the information collected across tasks. While our initial meta-learning formulation is widely general, and connects with state-of-the art strategies, we will focus on the case of reinforcement learning.
Wednesday, October 16th, 2019 Raisler Lounge
Photonic Doping of ENZ Metastructures & Potential Applications
Bio: Ehsan received his bachelor’s degree in electrical engineering from Sharif University of Technology, Iran in 2016. He is now a fourth year PhD student in the Engheta Lab, in the ESE department at the University of Pennsylvania. His core research projects focus on applications of doped ENZ metastructures, as well as extremely tunable Smith-Purcell radiation from an electron beam moving adjacent to spatio-temporally modulated substrates.
Abstract: Epsilon-near-zero (ENZ) hosts doped with non-magnetic dielectric inclusions have been shown to exhibit a highly tunable magnetic permeability. In this talk, we discuss how such a peculiar magnetic response may be exploited to realize a broad range of novel functionalities, such as impedance matching, enhanced magnetic nonlinearity, nonlinear absorbers, optical bistability, tunable electric field enhancement.
Wednesday, October 30th, 2019 Raisler Lounge
Graph Convolutional Neural Networks
Bio: Fernando Gama received the electronic engineer degree from the University of Buenos Aires, Argentina, in 2013, and the M.A. degree in statistics from the Wharton School, University of Pennsylvania, Philadelphia, PA, USA, in 2017. He is currently working towards the Ph.D. degree with the Department of Electrical and Systems Engineering, at the University of Pennsylvania. He has been a visiting researcher at TU Delft, the Netherlands, in 2017 and a research intern at Facebook Artificial Intelligence Research, Montreal, Canada, in 2018. His research interests are in the field of information processing and machine learning over network data. He has been awarded a Fulbright scholarship for international students.
Abstract: Convolutional neural networks (CNNs) restrict the linear operation of neural networks to be a convolution with a bank of learned filters. This makes them suitable for learning tasks based on data that exhibit regular structure. The use of convolutions, however, makes them unsuitable for processing data that do not exhibit such a regular structure. Graph signal processing (GSP) has emerged as a powerful alternative to process signals whose irregular structure can be described by a graph. Central to GSP is the notion of graph convolutions which can be used to define convolutional graph neural networks (GNNs). In this paper, we show that the graph convolution can be interpreted as either a diffusion or aggregation operation. When combined with nonlinear processing, these different interpretations lead to different generalizations which we term selection and aggregation GNNs. The selection GNN relies on linear combinations of signal diffusions at different resolutions combined with node-wise nonlinearities. The aggregation GNN relies on linear combinations of neighborhood averages of different depth. Both of these models particularize to regular CNNs when applied to time signals but are different when applied to arbitrary graphs. Numerical evaluations show different levels of performance for selection and aggregation GNNs.
Wednesday, November 6th, 2019 Raisler Lounge
Finite Sample Analysis of System Identification
Bio: Anastasios Tsiamis received the Diploma degree in electrical and computer engineering from the National Technical University of Athens, Greece, in 2014. He is now a fifth year PhD student in the Department of Electrical and Systems Engineering at the University of Pennsylvania. His research interests include learning in control, system identification, security and privacy in networked control systems.
Abstract: System identification methods often return parameters which deviate from the true ones, especially when the number of input/output data is finite. Previously, asymptotic methods have been employed to compute bounds for the system identification parameter errors. However, such methods are valid when the number of data goes to infinity. In this work, we analyze the finite sample complexity of stochastic system identification using modern tools from machine learning and statistics. An unknown discrete-time linear system evolves over time under Gaussian noise without external inputs. The objective is to recover the system parameters as well as the Kalman filter gain, given a single trajectory of output measurements over a finite horizon of length N. Based on a subspace identification algorithm and a finite number N of output samples, we provide non-asymptotic high-probability confidence intervals for the parameter estimation errors. Our analysis uses recent results from random matrix theory, self-normalized martingales and SVD robustness in order to show that with high probability the estimation errors decrease with a rate of inverse of the square root of N. Our non-asymptotic bounds not only agree with classical asymptotic results, but are also valid even when the system is mildly non-stationary (marginally stable and non-explosive).
Wednesday, November 13th, 2019 Raisler Lounge
Contact-Aware Controller Design for Complementarity Systems
Bio: Alp Aydinoglu received B.S. degrees in control engineering and communications engineering from Istanbul Technical University in 2016 and 2017 respectively. He is now a PhD student in the Department of Electrical and Systems Engineering at the University of Pennsylvania. His current research interests include hybrid dynamical systems, complementarity systems and contact mechanics.
Abstract: While many robotic tasks, like manipulation and locomotion, are fundamentally based in making and breaking contact with the environment, state-of-the-art control policies struggle to deal with the hybrid nature of multi-contact motion. Such controllers often rely heavily upon heuristics or, due to the combinatoric structure in the dynamics, are unsuitable for real-time control. Principled deployment of tactile sensors offers a promising mechanism for stable and robust control, but modern approaches often use this data in an ad hoc manner, for instance to guide guarded moves. In this work, by exploiting the complementarity structure of contact dynamics, we propose a control framework which can close the loop on rich, tactile sensors. Critically, this framework is non-combinatoric, enabling optimization algorithms to automatically synthesize provably stable control policies.
Wednesday, November 20th, 2019 Raisler Lounge
Safety Verification of Nonlinear Polynomial System via Occupation Measures
Bio: Shaoru Chen received his Bachelor of Engineering degree in Automation from Zhejiang University, China in 2017. He is now a third-year PhD student in the Department of Electrical and Systems Engineering at the University of Pennsylvania. His research interest is in robust and nonlinear control.
Abstract: In this talk, we introduce a flexible notion of safety verification for nonlinear autonomous systems by measuring how much time the system spends in given unsafe regions. We consider this problem in the particular case of nonlinear systems with a polynomial dynamics and unsafe regions described by a collection of polynomial inequalities. In this context, we can quantify the amount of time spent in the unsafe regions as the solution to an infinite-dimensional linear program (LP). We approximate the solution to the infinite-dimensional LP using a hierarchy of finite-dimensional semidefinite programs (SDPs). The solutions to the SDPs in this hierarchy provide monotonically converging upper bounds on the optimal solution to the infinite-dimensional LP.
Optimal reduced-order modeling of bipedal locomotion
Bio: Yu-Ming Chen is a PhD student in the Department of Electrical and Systems Engineering at the University of Pennsylvania, supervised by Professor Michael Posa. He has a Bachelor’s degree in Physics from National Taiwan University and a Master’s degree in Robotics from the University of Michigan, Ann Arbor. His interests include legged locomotion, optimization and control.
Abstract: State-of-the-art approaches to legged robots are widely dependent on the use of models like the linear inverted pendulum (LIP) and the spring-loaded inverted pendulum (SLIP), popular because their simplicity enables a wide array of tools for planning, control, and analysis. However, they inevitably limit the ability to execute complex tasks or agile maneuvers. In this work, we aim to automatically synthesize models that remain low-dimensional but retain the capabilities of the high-dimensional system. For example, if one were to restore a small degree of complexity to LIP, SLIP, or a similar model, our approach discovers the form of that additional complexity which optimizes performance. We define a class of reduced-order models and provide an algorithm for optimization within this class. To demonstrate our method, we optimize models for walking at a range of speeds and ground inclines, for both a five-link model and the Cassie bipedal robot.
Wednesday, December 4th, 2019 Raisler Lounge
Reducing FPGA Compile Time with Separate Compilation for FPGA Building Blocks
Bio: Yuanlong Xiao receives his Master Degree of Microelectronics from Fudan University, China in 2017 and Bachelor degree of Microelectronics from Sun Yat-sen University, China in 2014. His research interest is quick computing mapping, reconfigurable computing and FPGA chip design.
Abstract: In this talk, we are going to talk about how to realized quick FPGA computing mapping. Today’s FPGA compilation is slow because it compiles and co-optimizes the entire design in one monolithic mapping flow. This achieves high quality results but also means a long edit-compile-debug loop that slows development and limits the scope of design-space exploration. We introduce PRflow that uses partial reconfiguration and an overlay packet-switched network to separate the HLS-to-bitstream compilation problem for individual components of the FPGA design. This separation allows both the incremental compilation, where a single component can be recompiled without recompiling the entire design, and parallel compilation, where all the components are compiled in parallel. Both uses reduce the compilation time. Mapping the Rosetta Benchmarks to a Xilinx XCZU9EG, we show compilation times reduce from 42 minutes to 12 minutes (one case from 160 minutes to 18 minutes) when running on top of commercial tools from Xilinx. Using Symbiflow (Project X-Ray/Yosys/VPR), we show preliminary evidence we can further reduce most compile times under 5 minutes, with some components mapping in less than 2 minutes.
Wednesday, December 11th, 2019 Raisler Lounge
