Mahadev Satyanarayanan, CSD - Elijah-Cloudlet
Ever thought of how mobile systems are able to do more every year, despite their limited size and power? Satya’s group has been shaping the research in Mobile & Distributed computing over the past many years. However, he believes that true mobility still eludes us, as we still rely on desktops/laptops for larger amounts of work. At best, today’s mobile systems use public Cloud technology for larger work. However, due to limitations of this architecture, Mobile systems do not perform comparably to a laptop/desktop. The Elijah-Cloudlet research group aims to make mobile systems more capable and fast, with applications to the cars, surveillance systems, and campuses of the future.
David Garlan, ISR - Task-oriented Computing (ToC)
To make the Internet of Things a tractable goal, users need natural ways to interact with an increasingly vast array of devices and to deal with the information they produce. ToC explores ways to represent user tasks in machine-readable formats, so that systems may self-configure to user needs rather than users having to figure out and adjust to system features. Research questions include: What task models are flexible enough for everyday activities such as do-it-yourself projects and cooking? How to link task models with knowledge about context and about user skills? Which software architectures support the deployment of ToC onto open, decentralized systems for the Internet of Things?
Abhinav Gupta, RI - Computer Vision and Machine Learning
Representation and reasoning for understanding a scene; reasoning based on physical, functional and causal relationships among different elements in the scene. What role does language play in vision and vice versa? Exploring how declarative information and linguistic information can be harnessed to learn how the world works. How can we mine huge visual data from the web to learn common sense relationships and how can we improve Computer Vision itself by exploiting such relationships. Studying how do humans interact with their environment and how does their perception of visual world depends on these interactions and their abilities.
Attila Yavuz, Bosch IoTS - Security and Privacy
Security and privacy are vital requirements for the success of smart infrastructures and big data technologies. We will discuss some of the challenges in these domains: Smart infrastructures require secure and reliable operation of distributed and heterogeneous systems under heavy resource constraints (e.g., bandwidth, battery, computation). Big data technologies involve storage and analysis of large scale data. To reduce financial and operational overhead, data/analysis is outsourced to external parties, which creates security and privacy problems. We will discuss practical and efficient privacy enhancing technologies such as searchable encryption and multi-party computation.
Anthony Rowe, ECE - Smart Campus
The vision of the Internet of Things and Services (IoTS) is one of a fully connected and programmable world where software applications seamlessly interact with the physical environment. Advances over the last decade have brought us closer to realizing this vision, and yet often even basic cyber-physical systems (CPS) can be difficult to develop, prototype, test and validate. With performance, scalability and security as first-class concerns, we have created the Sensor Andrew framework, which is enabling CMU to become the „World’s Most Sensed Campus.” We will present details of the Sensor Andrew architecture, key features such as data warehousing and visualization, and initial results from the first live deployment in Scaife Hall.
Niki Kittur, HCII - Social Computing
People are being increasingly overwhelmed by the complexity of products and enormous amount of information available today. We aim to build computational systems that harness the power of many human intelligences working together to turn overwhelming information into useful knowledge. The goal of these systems is to collect, annotate and synthesize information to meet people’s information needs (e.g., “which camera should I buy”, “why is my hot water not working” ). We will present a set of studies investigating how to harness crowds to identify relevant and useful information, annotate it, and turn it into a cohesive, compelling, and personalized answer.
Bosch also offers: PhD Fellowships, summer internships, research opportunities