This project is part of the AAMP-UP '22 program, which focuses on energetic material research.
AAMP-UP is separate but highly partnered with SURF.

The project is run by Dr. Marcial Gonzalez and his team. The goal is to enable discovery of the physical mechanisms driving impact-induced reactions in explosives (such as interparticle relative velocities and contact pressures) and their correlation with material properties, formulation, and processing. Description: Energetic materials used in the propulsion industry, and particle-binder composites in general, consist of (explosive) crystals in a binder matrix. The mechanical properties and performance of these materials are strongly correlated to their microstructure. Specifically, the mechanical response under weak impact loads is characterized by long-range microstructural correlations (typical of granular materials) that need to be resolved with high statistical significance in order to understand how complex phenomena (such as detonation) emerges from local simpler interactions. The interplay between these long-range interactions and local mechanisms (such as deformation, failure and, ultimately, hot spot formation) is responsible for the sensitivity of the material. A mesoscale modeling methodology, capable of modeling the dynamic response of large energetic composites (>50,000 crystals) by analytically upscaling complex crystal-binder deformation and failure mechanics into closed-form contact laws, will be used to build time-dependent multivariate probability distribution functions of microstructural features.

Big Data/Machine Learning, Deep Learning

• No Major Restriction

MATLAB knowledge. Must be a U.S. citizen, national, or permanent resident of the United States. Must have completed at least one academic semester of full-time study at associate's or bachelor's degree level from an accredited college or university.

Mechanical Engineering

Marcial Gonzalez

This project is part of the AAMP-UP '22 program, which focuses on energetic material research.
AAMP-UP is separate but highly partnered with SURF.

The project is run by Dr. Song Zhang and his team. The research aims at designing a 3D printer that can incorporate a high-end customized 3D vision sensor for close-loop controls. Undergrads will closely work with graduate students on both software and/or hardware development depending upon interest.

Computer Architecture, Deep Learning, Mobile Computing, Other

• No Major Restriction

Must be a U.S. citizen, national, or permanent resident of the United States. Must have completed at least one academic semester of full-time study at associate's or bachelor's degree level from an accredited college or university.

Mechanical Engineering

Song Zhang

This project is part of the AAMP-UP '22 program, which focuses on energetic material research.
AAMP-UP is separate but highly partnered with SURF.

The project is run by Dr. Steven Son and his team. Machine learning (ML) tools are playing an increasingly important role in science and engineering, revealing patterns and providing predictive capabilities not achievable otherwise. This research area explores the utility of machine learning algorithms in the design, development, and characterization of various energetic material systems. Particular emphasis is placed on bringing a data science formalism to the field, with an eye toward both future capability development and more intelligent (and appreciably faster) material formulation and system design. The REU student would work closely with a Research Scientist and graduate student to gather data, analyze it using ML tools, and share these results.

Some experience w/ coding, AI, or ML recommended.

Big Data/Machine Learning, Deep Learning, Material Modeling and Simulation

• No Major Restriction

Some experience w/ coding, AI, or ML recommended. Must be a U.S. citizen, national, or permanent resident of the United States. Must have completed at least one academic semester of full-time study at associate's or bachelor's degree level from an accredited college or university.

Mechanical Engineering

Steven Son

This project is part of the AAMP-UP '22 program, which focuses on energetic material research.
AAMP-UP is separate but highly partnered with SURF.

The project is run by Dr. Alejandro Strachan and his team. The project will use tools from data science and machine learning to develop predictive models for the performance of energetic materials. Students will learn about neural networks, deep learning, and the chemistry and physics of energetic materials.

Basic python knowledge is desirable for the project.

Big Data/Machine Learning, Deep Learning

• No Major Restriction

Basic python knowledge. Must be a U.S. citizen, national, or permanent resident of the United States. Must have completed at least one academic semester of full-time study at associate's or bachelor's degree level from an accredited college or university.

Materials Engineering

Alejandro Strachan

This project is part of the AAMP-UP '22 program, which focuses on energetic material research.
AAMP-UP is separate but highly partnered with SURF.

The project is run by Dr. Alejandro Strachan and his team. Computationally efficient models for accurately predicting the thermal and chemical properties of high explosive (HE) materials are of great interest for both military and civilian applications. In this project, students will work on developing a multiscale model for HE materials, namely HMX and TATB, via combination of dimensionality reduction through machine learning techniques and atomistic molecular dynamics (MD) simulations. Specifically, the decomposition behavior of HE will be investigated through reactive, atomistic MD simulations. The dimensionality of the thermal and chemical data collected from the MD will then be reduced through non-negative matrix factorization technique to develop reduced-order chemistry models for HMX and TATB. Utilizing these, continuum-level models that capture the decomposition behavior of these materials will be developed.

Students with programming experience would be a good fit for this project.

Big Data/Machine Learning, Deep Learning

• No Major Restriction

Programming experience a plus Must be a U.S. citizen, national, or permanent resident of the United States. Must have completed at least one academic semester of full-time study at associate's or bachelor's degree level from an accredited college or university.

Materials Engineering

Alejandro Strachan

The emergence of the Internet of Things (IoT) has transformed our world with billions of interconnected smart devices. It has received ubiquitous adoption in numerous industries such as healthcare transportation, manufacturing, and agriculture. Most IoT systems nowadays are empowered by AI technology to automate a lot of tasks with little human intervention. However, designing and customising IoT devices for individual needs still remains challenging for lots of end users with limited technical background , which greatly hinders the IoT’s mass adoption. In order to bridge this gap and lower the entry barrier, we plan to take full advantage of the immersive technology(AR/VR), which allows common users to create, author and debug IoT behaviours effortlessly inside a simulated virtual environment. Specifically, the undergrad will participate in design and fabrication of unique IoT devices for a variety of tasks and integrating them to an overarching virtual environment. The undergrad will also get involved in the user evaluation process. This work will eventually lead to the submission of a paper to a top-tier ACM conference.

Deep Learning, Human Factors, Internet of Things, Learning and Evaluation

Applicants should have a general interest in developing mixed-reality applications and designing electronic hardware. Applicants with experience in some (not all) of the following are preferred: PCB design, embedded programming, C#, Unity,3D-CAD Software, deep learning.

Electrical and Computer Engineering

Karthik Ramani

This project aims to design new deep learning/machine learning approaches to control two robots navigating through an environment. The robots aim to achieve a certain destination while avoiding obstacles and performing specific tasks, e.g., making a delivery along the way. The robots further perform information sharing and collaborate with each other to further increase their capability (the robot info:https://www.robotis.us/turtlebot-3-waffle-pi/). This setup aims to provide proof-of-concept for the efficacy of distributed deep learning models in a realistic multi-robot navigation task. The role of the undergraduate research would be to contribute to algorithm design, implementation, and benchmarking steps.

Big Data/Machine Learning, Deep Learning

• No Major Restriction

ECE

Abolfazl Hashemi

This project is part of the AAMP-UP '22 program, which focuses on energetic material research.
AAMP-UP is separate but highly partnered with SURF.

The project is run by Dr. Stephen Beaudoin and his team. The research will use tools from data science and machine learning to develop predictive models for the performance of energetic materials. Students will learn about neural networks, deep learning, and the chemistry and physics of energetic materials.

Big Data/Machine Learning, Deep Learning, Other

• No Major Restriction

Must be a U.S. citizen, national, or permanent resident of the United States. Must have completed at least one academic semester of full-time study at associate's or bachelor's degree level from an accredited college or university.

Chemical Engineering

Stephen Beaudoin

The student will work with PhD students on implementation of IoT technology on manufacturing machines and processes, database development, dashboard development, and machine learning for smart manufacturing.

Big Data/Machine Learning, Deep Learning, Fabrication and Robotics, Internet of Things

• Mechanical Engineering
• Computer Engineering
• Computer Science

Mechanical Engineering

Martin Jun

Machine learning, image processing, fluid-structure interaction, linear and nonlinear deformation, elasto-hydrodynamic lubrication... You may have learned or have heard some of those topics. But you may never see how those interdisciplinary techniques can be used together to solve a real-life engineering problem. This project offers you a unique experience to participate in my research group developing a first-in-kind machine learning-based simulation model for nonlinear contact problems in high-pressure hydrostatic machines. The objective of the SURF project is to create a machine learning algorithm capable of fast predicting the two-dimensional, nonlinear deformation distribution from the pressure distribution. The project will be rewarding and challenging, and your work will constitute a new modeling approach in the fluid power field. Therefore, it is also very possible to be published with you as a co-author.

You will be challenged to 1) learn to program a machine learning algorithm in TensorFlow, 2) generate a training dataset for machine learning models using a state-of-the-art numerical simulation tool, and 3) integrate the neural network into the existing modeling suite.

You will be supported by your graduate mentor, who specializes in these topic areas and will provide guidance throughout the project. You will also be supported by a group of 8 developers of the hydrostatic machine modeling toolset that are working on different aspects of the code.

Big Data/Machine Learning, Deep Learning, Fluid Modelling and Simulation, Material Modeling and Simulation

• No Major Restriction

• Intermediate knowledge of fluid mechanics • Basic programming knowledge (Python, MATLAB, C++, or similar languages)

Mechanical Engineering

Lizhi Shang

The ability to create 3D structures in the micro and nanoscale is important for many applications including electronics, microfluidics, and tissue engineering. This project deals with development and testing of a setup for building 3D structures using a femtosecond pulsed laser. A method known as two photon polymerization is typically used to fabricate such structures in which a polymer is exposed to a laser beam and at the point of the exposure the polymer changes its structure. Moving the laser in a predefined path helps in getting the desired shape, and the structures are then built in a layer by layer fashion. The setup incorporates all the steps from a designing a CAD model file to slicing the model in layers to generating the motion path of the laser needed for fabricating the structure. Like many other 3D printing processes, 3D printing at nanoscale is also slow. In order to make a 3D structure rapidly, many processes are currently being developed, including projecting 2D images and printing 3D structures in a rapid, layer-by-layer fashion. Other efforts include the use of machine learning to produce high quality 3D parts and printing materials other than polymers to achieve specific mechanical, electrical or optical properties. The undergraduate student will work with graduate student to learn the state-of-the-art 3D nanoprinting systems, help to develop rapid printing processes, and analyze printing results.

Big Data/Machine Learning, Deep Learning, Fabrication and Robotics, Material Processing and Characterization, Nanotechnology

• Mechanical Engineering
• Physics
• Industrial Engineering
• Computer Engineering

Junior or Senior standing, knowledge in CAD, knowledge in Python is a plus

Mechanical Engineering

Xianfan Xu

Atmospheric blocking events and 'Bomb Cyclones' are an important contributor to high impact extreme weather events. Both these weather extremes lead to heat waves, cold spells, droughts, and heavy precipitation episodes, which have dire consequences for the public health, economy, and ecosystem. For example, the blocking-induced heat waves of 2003 in Europe led to tens of thousands of human casualties and tens of billions of dollars of financial damage.

Traditionally, prediction of extreme weather events is based on direct numerical simulation of regional or global atmospheric models, which are expensive to conduct and involve a large number of tunable parameters. However, with the rapid rise of data science and machine learning in recent years, this proposed work will apply convolutional neural network to an idealized atmospheric model to conduct predictability analysis of extreme weather events within this model. With this proposed machine-learning algorithm, our project will provide a robust forecast of heat waves and atmospheric blocking with a lead-time of a few weeks. With more frequent record-breaking heat waves in the future, such a prediction will offer a crucial period of time (a few weeks) for our society to take proper preparedness steps to protect our vulnerable citizens.

This project is based on developing and verifying the machine learning algorithm for detecting extreme weather events in an idealized model. We will use Purdue’s supercomputer Bell to conduct the simulations. The undergraduate student will play an active and important role in running the idealized model, and participate in developing the algorithms. As an important component of climate preparedness, the proposed work aims to develop a physics-informed machine learning framework to improve predictability of extreme weather events.

Closely advised by Prof. Wang, the student will conduct numerical simulations of an idealized and very simple climate model, and use python-based machine learning tools to predict extreme weather events within the model. Prof. Wang will provide weekly tutorial sessions to teach key techniques along with interactive hands-on sessions. The students will get access to the big datasets on Purdue’s Data Depot, analyze and visualize data of an idealized atmospheric model. The student will use convolutional neural networks (CNNs) to train and assess a Machine-Learning model. The student will further use feature tracking algorithm to backward identify the physical structure in the atmosphere that is responsible for the onset of extreme weather events.

Big Data/Machine Learning, Deep Learning, Energy and Environment, Environmental Characterization, Fluid Modelling and Simulation

• Physics
• Planetary Sciences
• Atmospheric Science/Meteorology
• Computer Science
• Mathematics - Computer Science
• Mathematics
• Environmental Geosciences
• Mechanical Engineering
• Civil Engineering
• Aeronautical and Astronautical Engineering
• Computer Engineering
• Engineering (First Year)
• Multidisciplinary Engineering
• Natural Resources and Environmental Science (multiple concentrations)

Familiar with Machine Learning or prior knowledge of convolutional neural networks (CNNs); Have basic level training on PHYS172 Modern Mechanics or PHYS 15200 Mechanics or equivalent courses from other institutions; Familiar with Python scripting and visualization

Earth, Atmospheric, and Planetary Sciences

Lei Wang

The increasing worldwide demand for vaccines along with the intensifying economic pressure on health care systems underlines the need for further improvement of vaccine manufacturing. In addition, regulatory authorities are encouraging investment in continuous manufacturing process to ensure robust production, avoid shortages, and ultimately lower the cost of medications for patients. The limitations of in-line process analytical tools are a serious drawback of the efforts taken in place. In line analysis of viral particles are very limited, due to the large time required for the current techniques for detection, qualitative and quantitative analysis. Therefore, there is a need for new process analytical technology. This project has both experimental and computation components and two students will be recruited for perform different tasks. The student focusing on experiments will fabricate devices and test them. The student focusing on computations will focus on developing machine learning codes.

Big Data/Machine Learning, Biological Characterization and Imaging, Biological Simulation and Technology, Biotechnology Data Insights, Cellular Biology, Computer Architecture, Deep Learning

• No Major Restriction

For the experimental portion of the project: fabrication, cell culture, microfluidics, microscopy For the computational portion of the project: Coding, Python

Mechanical Engineering

Arezoo Ardekani

There is growing interest from Space agencies such as NASA and the European Space Agency in establishing permanent human settlements outside Earth. To advance knowledge in the field, the Resilient Extra-Terrestrial Habitat Institute (RETHi) is taking steps to develop technologies that will enable resilient habitats in deep space, that will adapt, absorb and rapidly recover from expected and unexpected disruptions without fundamental changes in function or sacrifices in safety.
To study, demonstrate, and evaluate the technologies developed in pursuit of this mission, a multi-physics cyber-physical testbed is being founded at the Ray W. Herrick Laboratories at Purdue University with collaboration from partners at three universities and two industrial partners. It allows to examine emergent behaviors in habitat systems and the interactions among its virtual (computational) and physical components. The testbed will consider a habitat system and will aim to emulate the extreme temperature fluctuations that happen in deep space. To achieve this goal, a thermal transfer system is being developed, consisting of a chiller, an array of glycol lines, in-line heaters, actuated valves, and a series of sensors. Operated under a tuned controller, the thermal transfer system can cool or heat a certain surface area of the structure of the habitat to maintain a given temperature. However, to fully control the thermal transfer system is not straightforward. One of the critical challenges is its deep uncertainty, which results from inaccurate or long-delay sensors, variant test setup, complex controller design, etc. Therefore, a systematic study is needed to quantify the uncertainties to facilitate the thermal transfer system development. Emulation of a particular scenario considering a meteoroid impact will be performed, with random variations in the location and size of the impact and resulting consequences.
We also aim to consider design trade-offs aimed toward the goals of resilience. Thus, we have also established a modeling platform to support rapid, stochastic simulations of habitat systems to quantify the space architectures that enhance resilience. These might consider the important features of the robots, the sensors, and the structure itself that make the habitat resilient. Physics-Infused modeling is a gray-box method to model physical parameters using low-fidelity/computationally-efficient models in conjunction with high-fidelity/computationally-expensive samples. We combine samples from the high-fidelity model framework with low-fidelity dynamic models and create a better combination for state prediction to achieve this goal. One of the critical problems here is the difference in state space of the models and finding the optimal method to sample a high-fidelity model.
We are looking for undergraduate students to play key roles in this project, under the guidance of a graduate student and faculty members. The students are also expected to prepare a poster presentation on the results, and author a research paper if the desired results are achieved.

Deep Learning, Human Factors, Material Modeling and Simulation, Thermal Technology

• No Major Restriction
• Mechanical Engineering
• Aeronautical and Astronautical Engineering
• Civil Engineering
• Computer Engineering
• Computer Science

Students interested in this project should be critical thinkers, and have good experimental skills. Some projects will require programming skills (Python), CAD skills, and experience in MATLAB/Simulink.

Mechanical Engineering, Civil Engineering, Aerospace Engineering (we will have multiple faculty advising)

Shirley Dyke

Possible industry involvement: Some of these projects are funded by Facebook research awards and J.P.Morgan AI research awards. 

We have three openings for different tasks including those listed below.
NOTE: We especially encourage applications from women, Aboriginal peoples, and other groups underrepresented in computing.

*** Subproject 1. Testing Deep Learning Systems 

We will build cool and novel techniques to make deep learning code such as TensorFlow and PyTorch reliable and secure. We will build it on top of our award-winning paper (ACM SIGSOFT Distinguished Paper Award)! 

Machine learning systems including deep learning (DL) systems demand reliability and security. DL systems consist of two key components: (1) models and algorithms that perform complex mathematical calculations, and (2) software that implements the algorithms and models. Here software includes DL infrastructure code (e.g., code that performs core neural network computations) and the application code (e.g., code that loads model weights). Thus, for the entire DL system to be reliable and secure, both the software implementation and models/algorithms must be reliable and secure. If software fails to faithfully implement a model (e.g., due to a bug in the software), the output from the software can be wrong even if the model is correct, and vice versa.  

This project aims to use novel approaches including differential testing to detect and localize bugs in DL software (including code and data) to address the testing oracle challenge. 

Early work and background can be found here: 
EAGLE: Creating Equivalent Graphs to Test Deep Learning Libraries (our ICSE 2022 paper, forthcoming, check my homepage)
https://www.cs.purdue.edu/homes/lintan/publications/fairness-neurips21.pdf
https://www.cs.purdue.edu/homes/lintan/publications/variance-ase20.pdf
https://www.cs.purdue.edu/homes/lintan/publications/cradle-icse19.pdf

*** Subproject 2. Inferring Specifications from Software Text for Finding Bugs and Vulnerabilities

A fundamental challenge of detecting or preventing software bugs and vulnerabilities is to know programmers’ intentions, formally called specifications. If we know the specification of a program (e.g., where a lock is needed, what input a deep learning model expects, etc.), a bug detection tool can check if the code matches the specification. 

Building upon our expertise on being the first to extract specifications from code comments to automatically detect software bugs and bad comments, in this project, we will analyze various new sources of software textual information (such as API documents and StackOverflow Posts) to extract specifications for bug detection. For example, the API documents of deep learning libraries such as TensorFlow and PyTorch contain a lot of input constraint information about tensors. 

Early work and background can be found here: 
https://www.cs.purdue.edu/homes/lintan/projects.html

*** Subproject 3. Leveraging Deep Learning to Detect and Fix Software Bugs and Vulnerabilities

In this project, we will develop cool machine learning approaches to automatically learn bug and vulnerability patterns and fix patterns from historical data to detect and fix software bugs and security vulnerabilities. 

Early work and background can be found here: 
https://www.cs.purdue.edu/homes/lintan/publications/cure-icse21.pdf
https://www.cs.purdue.edu/homes/lintan/publications/deeplearn-tse18.pdf

Big Data/Machine Learning, Cybersecurity, Deep Learning, Other

• Computer Science
• Computer Engineering
• Software Engineering

Strong coding skills and motivation in research are required. Background in security or machine learning is not required but a plus.

Computer Science

Lin Tan

Two projects are available. One involves the investigation of enhancing optical imaging resolution using single photon counting techniques. Conventional optical imaging has a hard limit on its spatial resolution, to about one half of the wavelength, and many situations can benefit from higher resolution. In addition, it is challenging to image through scattering media. By way of example, being able to sense with light deeper in the brain would be of enormous benefit in neuroscience. The statistics of photons emitted by or transmitted through an object contain valuable information about the object which could be used to enhance image resolution and possibly see through substantial background scatter. Experiments will be conducted using laser light and with a set of single photon avalanche detectors (SPADs) to measure photon correlations in time, over wavevector (direction), and between detectors in various imaging configurations. Results from these experiments will be used to assess the effectiveness of various techniques for enhancing spatial resolution in imaging applications. This work has a diverse set of potential applications including biological imaging, sensing defects in semiconductors, and imaging through fog. The other project relates to optical forces on structured membranes induced by a laser. The modeling of the mechanical motion of a thin membrane deflected by laser light will be used to determine the membrane properties from experimental and simulated data. This will allow extraction of the mechanical material properties and more generally the validation of a theory for optomechanics that can then be used in design. The nascent field of optomechanics offers enormous impact scope, including remote actuation and propulsion, of importance in fields as diverse and molecular biology, communication, and transport. This project relates to attaining the underpinnings to move along such paths in engineering, as well as the basic physics of optical forces in material at small length scales.

Biological Characterization and Imaging, Biological Simulation and Technology, Deep Learning, Material Modeling and Simulation, Nanotechnology, Other

• Electrical Engineering
• Physics

Students with an interest in experimental work and a strong background in electromagnetics would be a good fit for this project. The undergraduate student will work with graduate students to perform experiments in an optics laboratory, perform modeling and data analysis using MATLAB or python, and review relevant literature to develop a working understanding of single photon measurement techniques and their applications to super-resolution imaging. This project would be suitable for students majoring in electrical engineering, physics, or a related discipline.

Electrical and Computer Engineering

Kevin Webb

A stethoscope-based surface sound sensor has been developed at Purdue and this sensor is being used to monitor manufacturing equipment. Deep learning will be used to classify machine and process conditions using the sound sensor. Student will work with PhD student to implement sound sensors on machines in local manufacturing companies, collect data, build database and dashboard, and develop AI models.

Big Data/Machine Learning, Deep Learning, Fabrication and Robotics, Internet of Things

• Mechanical Engineering
• Computer Engineering
• Electrical Engineering
• Computer and Information Technology
• Computer Science

Mechanical Engineering

Martin Jun

Develop optical flow/depth estimation with dynamic vision sensor cameras as the only vision sensor.

Big Data/Machine Learning, Deep Learning

• Electrical Engineering

Deep learning

IN

Kaushik Roy