Statistical Signal Processing and Machine Learning

Nick Roseveare, Phd
Research Scientist
Featured research: High-dimensional Statistical Modeling, e.g., model-order selection

I focus on development and application of statistical signal processing and machine learning algorithms for a variety of applications. In particular, I am interested in theoretical under-pinning and hidden behaviors of ML algorithms and what can be understood better about them in terms of classical statistics.

Interests and Philosophy of Research

I am always attempting to abide by the following life (and career) principles:

• Learning: I am always trying to become a better engineer and researcher. I am always absorbing lessons from my own and others’ experience, as well as reading and keeping up with trends in my field. I also very much enjoy the teaching process, it is gratifying to help other come to an understanding of a topic and to strengthen and question my own knowledge of the field.
• Leadership: I enjoy and work at understanding the big picture in order to stay on target and to help others do the same. I try to be intentional and to take ownership of my deliverables. I also recognize that sometimes others are more suited to a particular task. A good leader creates environments and teams that don’t need them.
• Teamwork: Excellent research and products require teamwork, collaboration, and respect for one another. Being right is feels nice, but it always comes in second-place behind what is good for the project and the team.
• Hard Work: A successful project is always the result of much striving and frustration. I push hard on a project and catalog all paths until a solution is found. I find the learning along the way and the hard-won results to be rewarding twin goals.

My technical strengths include:

• Algorithm development, including initial prototyping from research papers, progressive improvements and novels approaches, all the way through practical implementation and infrastructure choices (database, message-queues, containerization, etc.).
• Statistical signal processing, where I have worked on several challenging problems in tracking and filtering, time-series smoothing, phase reconstruction, high-dimensional model-order selection, etc.
• Optimization, where I have worked on a variety of convex-relaxations of NP-hard problems (TSP, cost assignment, constrained resource allocation, etc.).
• Machine learning, with experience in anomaly detection, natural language processing, and computer vision utilizing a variety of existing and in-house developed solutions including ensemble methods, deep-learning, and high-dimensional correlation analysis.

View summary profile here.
View résumé here.