schedule

Abstract: The last few years, robots have moved from the pages of science fiction books into our everyday reality. Currently, robots are utilized in entertainment, scientific exploration, manufacturing, and household maintenance. While the above advances were made possible by recent improvements in sensors, actuators, and computing elements, the research of today is focused on the computational aspects of robotics. In particular, methodologies for utilizing the vast volumes of data that can be generated by a robotic mission, together with techniques that would allow a robot to respond adequately in unforeseeable circumstances are the challenges of tomorrow. This talk presents an overview of algorithmic problems related to marine robotics, with the particular focus on increasing the autonomy of robotic systems in challenging environments. I will talk about vision based state estimation and mapping of underwater caves. Motion planning algorithms for covering an aquatic environment will be discussed with a focus on rivers and lakes. In addition I will talk about several vehicles used at the University of South Carolina such as drifters, underwater, and surface vehicles. In addition a short overview of current projects will be discussed. The work that I will present has a strong algorithmic flavour, while it is validated in real hardware. Experimental results from several testing campaigns will be presented.

Abstract: Combination puzzles, such as the Rubik’s cube, pose unique challenges for artificial intelligence. Furthermore, solutions to such puzzles are directly linked to problems in the natural sciences. In this talk, I will present DeepCubeA, a deep reinforcement learning and search algorithm that can solve the Rubik’s cube, and six other puzzles, without domain specific knowledge. Next, I will discuss how solving combination puzzles opens up new possibilities for solving problems in the natural sciences. In particular, I will describe how we are using DeepCubeA to tackle problems in chemistry. Finally, I will show how problems we encounter in the natural sciences motivate future research directions. A demonstration of our work can be seen at http://deepcube.igb.uci.edu/.

Abstract: Intelligent decision making is at the heart of Artificial Intelligence (AI). A large number of real-world domains, such as autonomous vehicles, delivery robots, cyber security, and so many others, involve multiple AI decision makers, or agents, that cooperate with collective efforts in a distributed manner, where each agent’s decisions are based on its local information, with often limited communication with others. This distributed nature makes it challenging to design efficient and reliable multiagency. Issues like failure to coordinate, unsafe interactions, and resource misallocation can easily arise. In part one of the talk, I will be introducing our work on utilizing the notion of social commitments to achieve reliable and trustworthy multiagent coordination. Intuitively, a commitment regularizes an agent’s behavior so that it can be well anticipated and exploited by another. I will build up a formalism of this intuition, and discuss how multiagent commitments can be efficiently identified and faithfully fulfilled. In part two of the talk, I will be describing our ongoing research on automatically discovering multiagent learning algorithms, with a focus on learning to communicate.

Abstract: Processor technology has reached an impasse. While Moore’s Law remains alive for the moment, traditional general-purpose computer architectures have reached efficiency limits that have effectively stalled their performance growth. In response, processor designers have shifted their focus to designing specialized processors to deliver real-time performance of increasingly intensive workloads and achieving new levels of energy efficiency for low power mobile platforms. In this talk, I will describe how our research group is contributing to this effort by highlighting two of our current projects involving domain-specific processor architectures. The first is a reconfigurable overlay architecture for high-performance pattern-matching, which outperforms the state-of-the-art CPU- and GPU-based implementations for benchmark datasets. The second is a run-time optimization framework for computer vision applications designed for an embedded Digital Signal Processor architecture.

Abstract: Wireless communication has truly transformed the world. It has enabled us to connect the entire globe, and made it simple to reach people separated by thousands of miles. However, what receives less attention are other interesting properties of these wireless signals. The fact that wireless signal spread out in all directions and bounce off objects, make them a power lens to look at our world through. This facilitates sensing of the world through wireless signals. In this talk I present two main ideas: Wireless localization which measures the time wireless signals take to travel between two devices, and wireless material identification which analyzes the effect of a liquid on wireless signals, in order to identify the liquid.

Abstract: The pursuing of safer self-driving cars, smarter robots, and cell phone apps is accelerating, in which deep learning techniques are playing a major role. However, one of the major bottlenecks of deep learning in these edge computing devices is their limited computing power and severe energy constraints. As alternatives to von Neumann architectures, neuromorphic systems have big potential to address these issues by its avoidance of the processor-memory bottleneck, reduced energy consumption, and area-sparing computation. In this talk, we will introduce the “Neuro-Edge” project which proposes leveraging neuromorphic computing algorithms and hardware to address four important needs in machine learning (ML) at edge devices: (1) Energy-Efficient Computing: One of the major bottlenecks of ML in edge devices is their severe energy constraints. (2) Incremental Learning: in most of the edge devices data arrives in form of a stream rather than batches, therefore an ML model is required to learn continuously and incrementally upon the arrival of each sample data. (3) Capture Temporal Information: data streams produced by edge devices often exhibit temporal patterns and dependencies, which are important to be captured to distinguish the relationship between input features over time. (4) Drift Tolerance: edge devices normally operate in non-stationary environments, and are prone to concept drifts that are induced by the appearance and/or disappearance of features and/or classes in the incoming data stream.

Abstract: Millimeter-wave (mmWave) technology plays a central role in next-generation wireless networking, sensing, and imaging. In this talk, we will present an overview of our work on pushing the performance boundaries of millimeter-wave systems. We will present results on fast beamforming and alignment algorithms as well as medium access protocols for enabling dense spatial reuse in mmWave networks. We will also highlight our recent work on millimeter wave wireless networks on-chip. Finally, we will discuss our work on enabling through fog high-resolution mmWave imaging for self-driving cars using generative adversarial networks.

Abstract: Businesses large and small face a common challenge around attracting new customers and retaining existing ones, with marketing as a core component in tackling this challenge. As customers and businesses move online, the amount of data available to inform and improve marketing decisions has grown significantly. In this talk we’ll look at a high level overview of some of the technical challenges involved in making use of this growing set of data to improve marketing decisions and optimize toward business goals, as well as a sample of solutions explored.

Abstract: For more than two decades since the rise of the World Wide Web, the “Notice and Choice” framework has been the governing practice for the disclosure of online privacy practices. The emergence of new forms of user interactions, such as voice, and the enforcement of new regulations, such as the EU’s recent General Data Protection Regulation (GDPR) promises to change this privacy landscape drastically. In this talk, I will discuss the challenges towards providing the privacy stakeholders with privacy awareness and control in this changing landscape. I will also present our recent research on utilizing AI to analyze privacy policies and settings.

Abstract: This talk presents the challenges and opportunities of building a city-scale low-power wireless Internet-of-Things. We build upon low-power wide-area networking (LP-WAN), a technology that enables low-cost devices with a 10-year battery to communicate at few kbps to a base station, kilometers away. We address the challenges in deploying LP-WANs in large urban environments, given the power limits of the clients and attenuation from buildings that limit signal range. We further show how LP-WANs at shorter ranges can eliminate the need for a battery altogether. Beyond communication, the talk also discusses novel applications and sensing opportunities of an omnipresent low-power Internet.

Abstract: Secure pairing is key to trustworthy deployment and application of Internet of Things (IoT) devices. However, IoT devices lack conventional user interfaces, such as keyboards and displays, which makes many traditional pairing approaches inapplicable. Proximity-based pairing approaches are very usable, but can be exploited by co- located malicious devices. Approaches based on a user’s physical operations on IoT devices are more secure, but typically require inertial sensors, while many devices do not satisfy this requirement. A secure and usable pairing approach that can be applied to heterogeneous IoT devices still does not exist. We develop a technique, Universal Operation Sensing, which allows an IoT device to sense the user’s physical operations on it without requiring inertial sensors. With this technique, a user holding a smartphone or wearing a wristband can finish pairing in seconds through some very simple operations, e.g., pressing a button or twisting a knob. Mor over, we reveal an inaccuracy issue in original fuzzy commitment and propose faithful fuzzy commitment to resolve it. We design a pairing protocol using faithful fuzzy commitment, and build a prototype system named Touch-to-Pair (T2Pair, for short). The comprehensive evaluation shows that it is secure and usable.

Abstract: Motion tracking and RF sensing is a broad area with classical problems that dates back many decades. While significant advances have come from the areas of robotics, control systems, and signal processing, the emergence of mobile and IoT devices is ushering a new age of embedded, human-centric applications. Fitbit is a simple example that has rapidly mobilized proactive healthcare; medical rehabilitation centers are utilizing wearable devices towards injury diagnosis and prediction. In this talk, I will discuss a variety of (new and old) IoT applications that present unique challenges at the intersection of mobility, multi-modal sensing, and indirect inference. For instance, I will discuss how inertial sensors embedded in balls, racquets, and shoes can be harnessed to deliver real-time sports analytics on your phone. In a separate application, I will show how GPS signals can be utilized to track the 3D orientation of an aggressively flying drone, ultimately delivering the much needed reliability against crashes. Finally, I will discuss sensing liquid materials by passing WiFi-like signals through containers holding liquids. In general, I hope to show that information fusion across wireless signals, sensors, and physical models can together deliver motion-related insights, useful to a range of applications in IoT, healthcare, and cyber physical systems.

Abstract: The recent proliferation of acoustic devices, ranging from voice assistants to wearable health monitors, is leading to a sensing ecosystem around us – referred to as the Internet of Acoustic Things or IoAT. My research focuses on developing hardware-software building blocks that enable new capabilities for this emerging future. In this talk, I will sample some of my projects. For instance, (1) I will demonstrate carefully designed sounds that are completely inaudible to humans but recordable by all microphones. (2) I will discuss our work with physical vibrations from mobile devices, and how they conduct through finger bones to enable new modalities of short range, human-centric communication. (3) Finally, I will draw attention to various acoustic leakages and threats that arrive with sensor-rich environments. I will conclude this talk with a glimpse of my ongoing and future projects targeting a stronger convergence of sensing, computing, and communications in tomorrow’s IoT, cyber-physical systems, and healthcare technologies.