Research
The Department of Computer Science has a strong emphasis on research, both within the department and collaboratively with other departments, universities, government organizations and industry partners. The departmental research areas include intelligence, systems, and algorithms.
Intelligence
Algorithmic Robotics
An interdisciplinary research area drawing from traditional computer science, artificial intelligence, cognitive science, philosophy, and engineering. Our laboratories leverage theories, methods, and techniques from these fields to perform research on computer vision and perception, learning and adaptation, planning and manipulation, natural language understanding and generation, and decision making in the context of unified and/or networked robot systems, as well as into the social, cognitive, and theoretical implications of algorithmic design choices.
Active Groups and Centers
- Pervasive Computing Systems Group (PeCS)
- Dynamic Automata Lab (DyaLab)
- Mines Interactive Robotics Research Lab (MIRRORLab)
- The Mines Robotics Graduate Program
Affiliated Faculty
Sponsors
- National Science Foundation (NSF)
- National Aeronautics and Space Administration (NASA)
- Air Force Office of Scientific Research (AFOSR)
- Office of Naval Research
- Army Research Laboratory (ARL)
- Department of Energy (DOE)
- Department of Transportation (DOT)
- Toyota Motor North America, Inc.
- Metcalf Archaeological Consultants, Inc.
- Alpha Foundation
Recent collaborators
- Army Research Laboratory
- Brown University
- Colorado State University
- Fine Art Miracles, Inc.
- George Mason University
- IBM
- Michigan State University
- National Renewable Energy (NREL)
- Rice University
- Royal Institute of Technology (KTH)
- University of Colorado
- University of Texas Rio Grande Valley
- US Air Force Academy (USAFA)
- Xcel Energy Inc.
Augmented Reality
Augmented reality is the process of augmenting the user’s view of the real world with computer-generated sensory information in the form of graphics (although sound can also be used). It is different from virtual reality, which completely replaces the user’s view of the real world with a simulated one. In AR, the user still sees the real world but the world is enhanced (or augmented) by virtual objects. For displaying the information, a head-mounted display can be used. More commonly, a hand-held device such as a smart phone or tablet is used. Virtual objects are overlaid on the live video images coming from the device’s camera.
A critical component in augmented reality is sensing the real world, in terms of recognizing objects and estimating the position and orientation (pose) of the user’s display relative to the real world. This information is critical in accurately rendering the virtual objects such that they are aligned with the real world. Another critical component is to automatically learn a model of the task being performed (such as a maintenance or assembly task) and recognize the state of the user’s progress through the task, in order to give guidance.
Affiliated Faculty
Machine Learning
Mines machine learning group is directed by Dr. Hua Wang, and consists of multiple PhD, master, and undergraduate research assistants.
We focus on developing mathematical foundations and algorithms needed for computers to learn. Our research spans the areas of machine learning and data mining, as well as their applications in a number of practical areas, such as cheminformatics, bioinformatics, medical image analysis, and computer vision. The goal of our research is to bridge the gap between computational solutions and real-world problems. Several of our current research topics include:
- Robust learning models: Data collected from real-world problems are inevitably compromised by noises and outliers, which makes conventional learning models hard to achieve satisfactory performance. As a result, designing learning models that are insensitive to noises, particularly to outlying data points, is of great value for practice. We utilized state-of-the-art numerical methods and redesigned a number of most broadly used learning models to improve their robustness against noises.
- Data fusion models: In many real-world applications, data is usually collected from different sources. For example, in disease diagnosis physicians can collect different types of data to evaluate different aspects of a patient, such as the physical data, imaging data, genetic data, and so on. How to effectively integrate this data is playing a critical role in diagnosis. We have built many mixed-norm induced learning models, which can elegantly extract the most relevant information from massive raw data collected by various instruments.
- Learning models for big data: With the recent development in technologies, we have accumulated vast amount of data, such as those in biology, information technology, to name a few. This data provides a wealth of information that is valuable for our lives. However, how to efficiently process the data with either a big number of samples or a big number features is a very challenging problem, which has aroused a lot of interests in both academia and industries. We are actively engaged in this area and have designed learning models with low computational complexity to deal with big data problems extracted from a variety of real-world domains.
Active Groups and Centers
Affiliated Faculty
Sponsors
Recent collaborators
Systems
cyberSecurity
Mines CSP (Cyber Security and Privacy) research group is directed by Dr. Chuan Yue, and it consists of multiple PhD, master, and undergraduate research assistants.
We focus on investigating the cyber security and privacy problems related to the web, mobile, cloud, cyber-physical, and IoT systems as well as their users. We take four main approaches in our research:
- design novel systems and use techniques such as machine learning and program analysis to investigate security and privacy vulnerabilities
- design and perform novel user studies towards achieving usable security and privacy
- design novel mechanisms and software features to effectively strengthen the security and privacy protection capabilities of systems
- design and conduct novel security and privacy educational research
We actively collaborate with researchers in other areas and other disciplines, as well as with industry and government partners.
Some Examples of Research Projects in Dr. Yue’s CSP Group:
Affiliated Faculty
Sponsors
- National Science Foundation
- National Security Agency
- Army Research Office
- Amazon
Recent collaborators
- IBM Thomas J. Watson Research Center
- Facebook Inc.
- NEC Labs America
Data Management Systems
Dr. Fierro’s research focuses on the design, implementation, and development of novel database systems that address real-world cyber-physical data management issues. This includes investigating the foundations of graph, linked-data and timeseries databases in both cloud-native and edge-computing/IoT contexts. In collaboration with the National Renewable Energy Laboratory, Dr. Fierro is researching the application of semantic web technologies to smart buildings, smart grids and other critical infrastructure in order to facilitate data discovery and automated configuration of IoT ensembles. This work intends to enable the application of modern data science techniques to solve built environment issue such as digitization and decarbonization.
Affiliated Faculty
Sponsors
- US Department of Energy
Recent collaborators
- National Renewable Energy Laboratory
- UC Berkeley
- Lawrence Berkeley National Laboratory
- Pacific Northwest National Laboratory
High Performance Computing and Programming Languages
Mines High Performance Systems and Software (HypeSYs) research group is co-directed by Dr. Bo Wu, Dr. Jedidiah McClurg, and Dr. Mehmet Belviranli. The interests of this group lie within the broad fields of high performance computing (HPC), programming languages, compilers, and heterogeneous computer architectures.
Dr. Bo Wu’s research aims at building high-performance software systems for deep learning and graph applications. His focus is on leveraging domain knowledge to create novel compiler and runtime techniques to systematically optimize parallel computing efficiency, maximize memory bandwidth utilization, and reduce computation redundancy.
Dr. Mehmet Belviranli’s research is centered on improving the utilization of diversely heterogenous architectures. He develops runtimes, analytical models and programming solutions to increase the computing and energy efficiency of autonomous and embedded systems. The target application areas of Dr. Belviranli’s research include but are not limited to machine learning acceleration, object detection and tracking, and motion planning & kinematics computing for self-driving cars, autonomous drones and collaborative robots.
Affiliated Faculty
Sponsors
Recent collaborators
- Katana Graph
- Microsoft
- North Carolina State University
- Northeastern University
- Oak Ridge National Laboratory
- Pacific Northwest National Lab
- Rochester University
- Sandia National Laboratory
- University of California
- University of Massachusetts
- University of Oregon
Networked systems
Continuing advances in the computational power, radio components, and memory elements have led to the proliferation of portable devices (e.g., intelligent sensors, actuators, RFID readers, PDAs, smartphone) with substantial processing capabilities and various networking interfaces. These devices are rapidly permeating a variety of application domains such as monitoring and remediation of an oil spill, underground mine safety, earth dam failure detection, and climate forecasting. These application areas align very well with Colorado School of Mines’ strategic areas: earth, energy, and environment.
We conduct research and development within the realm of wireless networking and mobile computing, addressing challenges raised by the emerging pervasive computing environments and Internet of Things (IoT). Our overall research style involves the design of algorithms and protocols, the evaluation of concepts and ideas via both simulations and actual system building, and the development of applications using the algorithms and systems built. In addition to focusing on basic computer science research, we also actively conduct interdisciplinary research. Example contributions include using wireless robotic networks for oil refinery inspection and using wireless sensor networks for intelligent geosystems, subsurface contaminant monitoring, and building monitoring and control.
Active Groups and Centers
Affiliated Faculty
Sponsors
- National Science Foundation
Algorithms
Applied Algorithms
Our research in Applied Algorithms and Data Structures combines classical algorithms research (characterized by the development of elegant algorithms and data structures accompanied by theory that provides mathematical guarantees about performance) and applications research (consisting of the actual development of software accompanied by empirical evaluations on appropriate benchmarks). Applications include cheminformatics and material science, blockchain, data analytics, edge computing, networking, Internet of Things, and VLSI design automation.
Active Groups and Centers
Affiliated Faculty
Sponsors
Computer Science Education
CS Education
This area encompasses research on STEM recruitment and diversity, K-12 computing education, and computing/engineering education at the university level. Current projects include an on-campus computing outreach program tailored for girls across a broad age range; professional development opportunities for CS high school teachers; and incorporating ethics into core and elective computing courses.