AI Applications
Point AI at real problems and define what "good enough" and "fast enough" mean for each one.
- Agentic AI
- Computer vision
- Reinforcement learning
- Spectrum sensing
- Trustworthy AI
NSF REU Site · Award #2548190 · Summer 2027
Build AI that runs faster, leaner, and smarter from app to silicon.
ARISE is a paid 10-week summer research program in the School of Electrical, Computing & Software Engineering (ECSE) at the University of Arizona in Tucson. Work with two faculty mentors on a real project that spans AI applications, efficient algorithms, and hardware.
A smaller model doesn't always run faster. An algorithm tuned for a data-center GPU may not run at all on a low-power edge chip. A medical imaging tool can wait minutes for an answer; a self-driving car has a fraction of a second. Making AI efficient in the real world means working across the application, the algorithm, and the hardware at the same time.
At ARISE you will do exactly that. Each project is co-led by two faculty mentors from different research areas, so you learn how specialists across the stack think and work together. You'll start with a week-long AI bootcamp, then spend the summer on hands-on research with graduate students in your mentors' labs, and finish by presenting a poster at the University of Arizona summer research colloquium.
Every ARISE project sits where two of these layers meet.
Point AI at real problems and define what "good enough" and "fast enough" mean for each one.
Shrink and speed up models without losing the accuracy that matters.
Design and program the chips and runtimes that AI actually runs on.
Co-mentoring by design. You're matched with two faculty mentors from different thrusts and meet them every week. A CS major and an ME major could both work on a robot: one on the algorithm and application, the other on the application and hardware. Students from ECE, CS, and related majors all fit.
Every ARISE project pairs two layers of the AI stack, with one mentor from each. New projects are designed for each cohort from our mentors' active research. Here is what each combination tends to involve.
Take a real AI application and redesign the model so it gives the same quality answers with less memory, time, and energy.
Co-design a model and the hardware that runs it, so an efficient algorithm turns into real savings on a device.
Put AI inside a physical system with tight limits on power and response time, then measure how it performs on the device.
You'll be matched to a project during selection based on your interests and experience, and you can still switch in the first week.
All ARISE mentors are faculty in Electrical, Computing & Software Engineering at the University of Arizona, with track records of mentoring undergraduate researchers to publications and graduate school.
Efficient AI algorithms and systems for vision, generative models, and language. Co-created deep learning courses at Duke and UA.
Faculty profileMachine learning and distributed computing for edge intelligence. Former Executive Director of Rice University's D2K Lab.
Faculty profileEfficiency, security, privacy, and trustworthy machine learning systems. NSF CAREER awardee.
Faculty profileScalable, efficient, and trustworthy AI, including large language models and model compression.
Faculty profileDecision-making, reinforcement learning, and neural-symbolic AI. Mentor in several past NSF REU programs.
Faculty profileMachine vision and computational imaging. Fellow of IEEE, SPIE, Optica, and the National Academy of Inventors.
Faculty profileLeads the ANNIE group on efficient, interpretable, human-aligned multi-agent systems.
Faculty profileHeterogeneous, reconfigurable, and neuromorphic computing. Has mentored 22 undergraduate researchers in the past six years.
Faculty profileSecure and intelligent computing, AI accelerators, and privacy-preserving edge computing. NSF CAREER awardee.
Faculty profileHardware and IoT security and ML for secure hardware design. Founder of the AI Hardware Design League.
Faculty profileOrientation, research ethics training, and a 4-day AI bootcamp with hands-on PyTorch and HPC labs.
Learn your lab's tools, read the literature, and finalize a research plan with your mentors.
Supervised research Monday to Thursday. Fridays for research skills (LaTeX/Overleaf, GitHub, poster design, responsible use of LLMs), professional development workshops, all-hands meetings, lab tours, and career and grad school panels.
Write your report, document your code, and present a poster at the UA summer research colloquium.
| Time | Mon | Tue | Wed | Thu | Fri |
|---|---|---|---|---|---|
| 08:30 | 1-on-1 with mentors | Research | Research | Research | Research skills training |
| 10:30 | Research | ||||
| 12:00 | Lunch on campus, often with mentors and lab members | Professional development workshop (lunch included) | |||
| 13:30 | Research | All-hands (weeks 3, 5, 7, 9) or lab tours and panels | |||
| 17:00 | Dinner on campus | ||||
| 18:00 | Sports, socials, and events with other summer researchers | ||||
Mentors are also available outside the Monday meeting whenever you need help.
Exact dates will be posted here when the application opens.
No prior research experience required. We especially encourage students from colleges with limited research opportunities, student veterans, and students from across the Southwest to apply to ARISE.
No. Most students in REU programs like ARISE have never done research before. The first two weeks are dedicated to a bootcamp and lab onboarding, and you'll receive self-study materials from our Introduction to Deep Learning course before the program starts. Programming experience (especially Python) is helpful.
Yes, if you study a related field such as mathematics, physics, or mechanical or biomedical engineering and want to work on AI systems. The ARISE co-mentoring model is designed so students with different backgrounds can each own a different part of a project.
Yes. Most participants will come from other institutions, but UA students are welcome to apply.
Participants receive a stipend, on-campus housing, and a campus meal plan for the 10 weeks.
We match you with two mentors from different research thrusts based on your application and interview. If the fit isn't right, you can change projects during the first week.
Many projects lead to conference papers. Mentors continue working with you through the fall to submit your results, and stay in touch to support graduate school and career plans.
huanruiyang@arizona.edu520-621-1118