The past decade has brought a sea change in the availability of data. Instead of a world in which we have small number of carefully curated data sources — instead we have a plethora of datasets, data versions, and data representations . Devices and data acquisition tools make it easy Read more…
Problems involving learning from ranking and choice data arise in several domains, ranging from sports rankings to elections and from recommender systems to marketing. Indeed, algorithms for these problems have been developed in many different fields, ranging from statistics to computer science and from social choice to marketing. Unfortunately, however, Read more…
The overarching goal of the DELPHI project is to simplify the development of correct and reliable software. Existing methodologies require significant human expertise, such as in the design of appropriate abstract domains and environment models, or suffer from significant barriers to adoption, such as limited scalability and flexibility. This project Read more…
The Internet of Things (IoT) promises to change the way we acquire and handle data — as Internet-connected devices (wearable or implantable medical sensors, environmental sensors, devices) “stream” data back to the cloud for processing, analysis, and real-time decision-making. Consider the example of heart arrhythmia detection below. A key component Read more…
Recent years have witnessed two major trends in the development of complex real-time systems. First, they are moving from physically isolated hosts towards common computing platforms shared by multiple systems. Using common platforms can bring significant benefits, including reduced cost and weight, as well as increased flexibility via dynamic resource Read more…
Critical care involves the use of two common types of medical devices: monitoring devices, such as bedside monitors; and actuation devices, such as infusion pumps. Traditionally, caregivers need to collect information from the monitoring devices and manually adjust the settings of actuating devices. Patient safety may be compromised if a Read more…
The project aims to develop control systems for ground vehicles that are resilient to a variety of external attacks. Our approach is to combine control-level techniques and code-level techniques, resulting in the approach comprises of two major stages: control law design and control task synthesis. Control-level defenses address attacks on Read more…
Operators of distributed systems often find themselves needing to answer a diagnostic or forensic question. Some part of the system is found to be in an unexpected state; for example, a suspicious routing table entry is discovered, or a proxy cache is found to contain an unusually large number of Read more…
Large, diverse datasets hold tremendous promise, if only we can derive statistical insights from them. But often, these datasets are siloed and withheld, because of privacy concerns. Differential privacy can mitigate these concerns — it provides a strong, attractive privacy guarantee that protects data owners from risks associated disclosure of Read more…
Important decisions such as police deployment, loan approvals, hiring, and parole from incarceration are beginning to be made by machine learning algorithms. This has lent urgency to the question of whether these algorithms are fair. A recent well-publicized analysis of recidivism risk predictions studied NorthPointe’s proprietary COMPAS algorithm used in Read more…