The Auton Lab works on theory, algorithms, and applications. The lab is unique in that identified needs in real-world application domains inform design choices in theory and algorithms. There are four broad application areas that motivate much of our research at the lab, and many of our projects fit into one or more of these groups.

Health Care & Public Health

Auton Lab performs research at the intersection of artificial intelligence and healthcare. Research efforts range from mining new biomedical imaging modalities for learnable structure, to forecasting cardio-respiratory instabilities from continuous vital sign measurements, to public health efforts aimed at modeling outbreak detection and testing policy interventions to determine a course of action.

Highlighted Work

  • Public Health Surveillance
    - The Auton Lab develops statistical inference model in a long-term joint research collaboration with epidemiologists and infection prevention experts from the University of Pittsburgh. Our algorithms detect systematic outbreaks and identify root causes by joining disparate sources of information such as genetic tests, patient electronic health records (EHRs), and other epidemiological information. Leveraging multiple data sources, our algorithms establish corroborating evidence to support or dismiss hypothetical outbreak scenarios, both increasing detectability and speed of analysis while maintaining low false alert rates. We also perform analytics to detect and forecast new positive cases of COVID-19 using microbiological testing of wastewater and develop a systematic analysis capability that the Allegheny County Health Department will use in daily practice for public health surveillance.
  • Critical Care Medicine
    - Critical Care Medicine, from initial incident and response to emergency room discharge or hospital admission in urban, rural, and field settings, is a challenging domain for machine learning. These incidents arise in routine emergency medicine operations as well as in large scale crises, inlcluding natural disasters, mass casualty incidents, emerging pandemics, and terrorist events, each putting unique demands and stresses on provision of the necessary care. The nature of the emergency involves unpredictability of demand for services, varying severity of cases, need for rapid assessment, and required persistent availability of stand-by resources. It puts a physical, cognitive and emotional strain on performers, exacerbating risk of human errors.
  • Field Medicine
    - Using cutting edge research and technologies, CMU develops techniques and systems intended to rapidly triage and stabilize injuries sustained in high-risk areas, particularly the battlefield or during military operations. Automated diagnosis and detection of injuries can quickly identify problems, with the eventual goal of automatic medical kits to stabilize soldiers and victims of disasters until qualified help can arrive.
  • Biomedical Imaging
    - We develop algorithms to leverage structure in images and videos. This intelligent featurization enables interpretable, downstream modeling.
Health Care & Public Health

AI x Natural Sciences & Knowledge Discovery

AI x Natural Sciences & Knowledge Discovery

Using AI to augment the ability of subject matter experts (SMEs) to pour through trememdous amounts of data and exract insights which further human knowledge and understanding of the universe.

Highlighted Work

  • Cosmology
    - Nature may resist a simple description, and the most important discoveries of the next century may be complex theories with countless variables and parameters. The era of big data opens up a promising new approach to scientific discovery. We develop statistical and machine learning methods for using observed and simulated data to advance machine learning with applications to cosmology. Bayesian Optimization based active-learning methods accelearte both the execution of the cosmological simulations and the search for best-fitting parameters. Many cosmology and other science applications require ML methods that can operate on more complex objects such as functions, distributions, or set and point clouds. The goal of this work is to make fundamental contributions in machine learning, statistics, and cosmology.
  • Physics
    - Physics-informed machine learning builds good models faster. The lab works with physicists to make fundamental contributions to physical science, ranging from reinforcement learning for plasma control policies for nuclear fusion to high fidelity synthetic data generation governed by the laws of physics.
  • Materials Discovery
    - Jeff's project

Robotics & Autonomous Vehicles

Auton Lab research on robotics systems spans perception, cognition, and actuaion. Current application domains for robotics research include autonomous trama care robots as well as autonomous vehicles.

Highlighted Work

  • Autonomous Vehicles
    - Auton lab makes significant contributions to autonomous vehicle technology ranging from reinforcement learning for motion planning to developing systems to deploy the technology in the real world.
  • Automated Medicine
    - Systems that operate in clinical theaters take many forms. Our work ranges from technology that supports physical robots that perform needle insertion to devices that measure vitals of patients in intensive care and forecast future episodes that will require emergency care. We develop systems that perform closed loop control of medial procedures such as fluid resuscitation, and monitoring quality of data collected at the bedside and reliability of healthcare equipment.
  • Automated Manufacturing
    - [blank]
Robotics & Autonomous Vehicles

Safety & Security

Safety & Security

Projects in this application area represent a variety of critical contexts for artificial intelligence and machine learning. Trust and understanding of the inner workings of models is an important step toward widespread deployment. Testing models to provide assurances of efficacy and safety encompasses both statistical and logical analyses. The Auton Lab not only makes an impact in Safety & Security by building state of the art AI, but also by informing policy decisions on how to ready complex, critical systems for a future where more insights and promise can be unlocked with AI.

Highlighted Work

  • Predictive Maintenance of Equipment
    - Predictive Maintenance in the Auton Lab focuses on applying machine learning to complex, critical systems such as aircraft, spacecraft, automobiles, and biological systems. With over 15 years of experience in predictive maintenance, we focus on reducing risks of unforeseen issues, reducing false positives in detection systems, and forecasting future failures of individual components. Common challenges in our predictive maintenance work include multi-modal data sources (e.g. text data describing failures combined with time series data from sensor), severe class imbalances due to existing maintenance processes, censored labels, and a general sparsity of labeled data. To overcome these challenges, work closely with end users and subject matter experts and use weak learning and weak supervision to rapidly construct large datasets, we use time series analysis to predict component failures, and use anomaly detection methods to discover outbreaks in the fleet.
  • Radiation Safety
    - We develop algorithms for both detection and decision support in nuclear threat identification. Using our flagship Bayesian Aggregation method for source detection and characterization we are developing fast and efficient tools for situational awareness and safety applications. Our work focuses on robust methods, multi-sensor and multi-modal data fusion, and decision support infrastructure for rapidly processing alerts.
  • Traffic & Public Safety
    - Identifying anomalous trajectories of pedestrians and highlighting when people appear to be where they should not be.

Time-series Intelligence

Anomaly detection, simultaneous anomaly detection and classification. Structure anomaly detections

Highlighted Work

Time-series Intelligence