Past CoE-funded Projects - Center of Excellence in Data Science and Artificial Intelligence

Past CoE Sponsored Projects

The CoE has provided many opportunities to further research of faculty and partner companies. The Center’s funding allowed data science researchers to produce innovative and visionary approaches in a variety of areas and industries. Their results created efficiencies, new approaches, and technological advances for the companies and their related fields. These partnerships also resulted in significant positive economic impact in New York State.

Assessment of Elasticity and Transmissivity of Crystalline Lens in Response to LIRIC Treatment

Headshot Photo of Jannick Rolland PI Researcher: Jannick Rolland and Kevin Parker

This project aims to develop a robust algorithm to accurately estimate shear wave speed (SWS) in elastography of the eye within a multi-frequency wave field. This will be employed to determine the effect of Clerio Vision’s laser treatment (“laser induced refractive index change”, LIRIC) upon the elasticity and transmissivity of the human crystalline lens. The motivation is to (1) advance clinical mapping of the biomechanical properties of the eye, and (2) restore the elasticity and transmissivity the human crystalline lens to reverse presbyopia. Computational models and animal lenses will be employed to determine accuracy and any impact on elasticity and transmissivity.

Chip – Fiber Alignment Assisted by Deep Learning

PI Researcher: Jaime Cardenas

We propose to develop a novel fiber array to chip alignment system using deep learning and computer vision for automating the alignment process for a fiber array.

Development of a Low-Cost, Low-Power Integrated Machine Health Monitoring Sensor

PI Researcher: Michael Heilemann

This project, in conjunction with ADVIS Inc. is focused on developing a device to cost-effectively bring machine health monitoring to a broad spectrum of Department of Defense (DoD) assets (vehicles, pumps, etc.), where the implementation of conventional monitoring systems is cost prohibitive. To meet size (~1 in.3) and power consumption (battery life of ~3 years) requirements, the device utilizes low-power embedded machine learning (ML) models trained on data acquired by a vibration sensor. Spectral features extracted from a recorded signal are used to train an embedded ML model to perform tasks such as the detection of anomalies and faults in mechanical systems.

Dynamical-System-Powered GL: When Hardware Dynamical System Meets Graph Learning

PI Researchers: Tong Geng

During the first year of CoE support, PI has demonstrated that Ising Machine with proper augmentation can be used to efficiently solve simple graph learning problems and outperforms GNNs in terms of inference speed and accuarcy. The resulting new graph learning paradigm is dubbed “IML”. This year, this project aims to improve IML through co-design of Hamiltonian, hardware architecture, and training algorithm and develop “DS-GL”, that effectively harnesses nature’s power within flexible, self-trainable, real-valued hardware Dynamical Systems to solve real-world Graph Learning problems. DS-GL aims to achieve fast training, high accuracy, and low inference response samultaneously.

Machine Learning to Produce Controllable Surface Nanopatterning

PI Researcher: Chunlei Guo

Surface nanopatterning has a range of applications, such as creating superhydrophobic glass without compromising its transparency. The Guo lab will work with AlchLight on developing machine learning for controllable laser surface nanopatterning. The project will focus on developing machine learning in searching

Multispectral Polarimetric Imaging of Nerves

PI Researcher: Wayne Knox

We have made significant progress in imaging nerves using polarized light sensing at different wavelengths. Starting in the visible spectral range and using raw or frozen chicken and now progressing to human cadaver nerve imaging, we have developed new methodologies of image processing to highlight the nerves to stand out from the surrounding tissues. We propose to expand our project to include a FIH (First in Human) surgical study which will present new data science challenges with image motion, and also to push into the near infrared for better tissue penetration, and to develop an endoscopic version of our technology.

Toward Speaker-Specific Voice Spoofing Countermeasures

Headshot Photo of Zhiyao Duan

PI Researcher: Zhiyao Duan

This is a continuation of last year’s CoE project which aimed to develop and deploy noise-resilient voice spoofing countermeasures to IngenID’s automatic speaker verification (ASV) engines. This new project aims to develop a new paradigm of spoofing countermeasures, specifically, speaker-specific countermeasures, to fit to each speaker’s characteristics more accurately. This is to respond to the highly realistic deepfakes generated by rapidly developing generative AI technology (e.g., voice cloning from ElevenLabs), which has been shown to successfully fool state-of-the-art speaker-generic voice spoofing countermeasures. By designing speaker-specific passive and active countermeasures, we will significantly improve the accuracy and robustness of countermeasures.

VUZIX – Multispectral Polarimetric Imaging of Nerves

PI Researcher: Wayne Knox

Development of a new Intraoperative Instrument Using AR/VR technology: We propose to continue and expand our recently started research project on using multispectral polarization technology for the detection of nerves and to learn how to incorporate it into a VR/AR/MR technology as a new surgical aid. This device could help to avoid accidental nerve cuts during routine surgery. The project will develop a research prototype for use in live human surgical clinical trials. We will test different modes of image processing such as artificial intelligence and methods of data science to determine the best modes of operation according to the surgeon’s requirements.

Photonect – Chip – Fiber Alignment Assisted by Deep Learning

PI Researcher: Jaime Cardenas

Developing a novel fiber array-to-chip alignment system using deep learning and computer vision to automate the alignment process for a fiber array.

AMD – Ising Machine Learning

PI Researchers:
Michael Huang (left)
Tong Geng (right)

Ising Machine Learning: When the Ising Model and the Ising Machine Meet Machine Learning: The drastically increasing computational cost of Machine Learning (ML) has brought concerns on environmental sustainability and challenges in its practical deployment. Inspired by the recent success of using the Ising model and the Ising machines to solve NP optimization problems with orders-of magnitude speedup and energy reduction, this project will investigate how to leverage the Ising model’s strong expressivity and the Ising machine’s extraordinary computational speed in learning problems to create a new ML paradigm, Ising Machine Learning (IML). IML has the potential to enable the sustainable development of ML by significantly reducing the computational cost of current ML modelities with increased accuracy.

Phlotonics – Big-Data Approaches to Wafer-Scale Analysis of Silicon Biosensor Yield and Prediction of Function

PI Researcher: Ben Miller

Digital health and healthcare analytics are rapidly growing fields in data science that are paving the way to personalized medicine. Democratization of healthcare analytics will be driven by the development of new point-of-care (POC) diagnostic tools that integrate seamlessly with electronic health record systems to facilitate data aggregation for analysis with machine learning methods thus improving patient outcomes. Critical to the availability and low-cost of diagnostic tools is the scale-up of manufacturing from hundreds to hundreds of thousands and beyond. However, fabrication at this scale requires fabrication quality and yield methods that scale as well. The aims of this proposal will advance the scale-up of POC diagnostic fabrication by 1) implementing new methods for data scrubbing as well as identifying opportunities to improve the quality of raw data, and 2) mapping the relationship between inline wafer-scale refractive index data and localized sensor performance. The proposed work will stimulate economic growth in Upstate New York by promoting technology transfer from the University of Rochester to Phlotonics, a U of R spin-out company.

Kitware – Towards Robust and Fair AI Algorithms against Multiple Shortcuts

PI Researcher: Chenliang Xu

Machine learning often achieves good average performance by exploiting unintended cues in the data. For instance, when backgrounds are spuriously correlated with objects, image classifiers learn background as a rule for object recognition. This phenomenon—called “shortcut learning”—undermines the robustness and fairness of AI algorithms. The objective of this proposal is to build robust and fair AI algorithms against multiple shortcuts. In contrast to prior arts holding the single-shortcut assumption, we propose a new framework to identify and mitigate multiple shortcuts. Our work can broadly benefit many research fields, e.g., algorithmic fairness and robustness, data-centric AI, and representation learning.

Immersitech – Objective Evaluation of 3D Spatial Audio System Performance

PI Researcher: Mark Bocko

Immersitech has developed an exciting portfolio of patented audio processing technologies, packaged as easy-to-integrate Software Development Toolkits (SDKs) that deliver advanced noise cancellation, voice clarity, and 3D spatial audio enhancement capabilities for business/event communications, social entertainment/gaming, and distance learning. Our SDKs are designed to provide global communication platforms with industry-leading audio quality capabilities, leading to higher user engagement and satisfaction.

IDEX Health and Science – Machine Learning for the Enhanced Design of Multilayer Optical Filters

PI Researcher: Pablo Postigo

Optical filters are one of the key products in optics. Nevertheless, there are still some limitations that affect the optical performance of optical filters, like angle of incidence or the operative spectral ranges. Pablo Postigo and IDEX Health Science, LLC conducted research to develop computational methods based in AI to design new optical filters with enhanced properties, like higher resistance to the angle of incidence, a wider transmission/ reflection band or a higher number of spectral bands. Furthermore, the use of a filter design procedure based on AI may open new paths for the development of filters with other enhanced properties like polarization selectivity.

LightTopTech – Microscope dual-mode design for Gabor-domain optical coherence microscopy and optical coherence tomography of the retina powered by automated layer detection and feature segmentation

Headshot Photo of Jannick Rolland PI Researcher: Jannick Rolland

Microscope dual-mode design for Gabor-domain optical coherence microscopy and optical coherence tomography of the retina powered by automated layer detection and feature segmentation: A microscope for dual-mode Gabor-Domain Optical Coherence Microscopy (GDOCM) and optical coherence tomography (OCT) imaging of the human retina, and related image processing tools will be developed in this project, enabling the commercialization of pre-clinical and clinical GD-OCM for retinal imaging.

Kitware – Domain Adaptation using Vision Transformers

Andreas Savakis Headshot PI Researcher: Andreas Savakis

Domain Adaptation using Vision Transformers: Domain adaptation aims to overcome the distribution shift between the target domain used for testing and the source domain used for training. We propose to utilize Vision Transformers as an alternative to convolutional features for fine-grained domain adaptation. Vision transformers, inspired by natural language processing, offer excellent performance due to their attention mechanism, and are promising for fine-grained classification. We propose to incorporate vision transformers in the source hypothesis transfer framework and apply our fine-grained adaptation method on the rare planes dataset, which includes synthetic data for training and measured data for testing.

IngenID – Developing and Deploying Spoofing Aware Speaker Verification System

Headshot Photo of Zhiyao Duan PI Researcher: Zhiyao Duan

Developing and Deploying Spoofing Aware Speaker Verification Systems: This is a continuation of our previous CoE project which aimed to develop a deep learning based Automatic Speaker Verification (ASV) system and deploy it in industrial settings. For the new project, we propose to systematically address the increasing challenge of spoofing attacks. Existing anti-spoofing research was separated from the overall ASV system design, and anti-spoofing models have not been widely deployed. We plan to develop Spoofing Aware Speaker Verification (SASV) systems and deploy them to in industrial settings. Objectives: 1) deep integration of anti-spoofing into ASV systems, 2) robust anti-spoofing techniques, and 3) open-source evaluation framework for SASV research.

Pfizer – Neural Network assisted Femtosecond Laser Fabrication of Anti-bacterial Surfaces

PI Researcher: Chunlei Guo

Neural Network assisted Femtosecond Laser Fabrication of Anti-bacterial Surfaces: The PI’s lab has been working with Pfizer in developing materials for biomedical applications. Pfizer plans to launch a new project in developing anti-bacterial surfaces to reduce the stringent requirements for equipment and vaccine container sterilization. Pfizer is a leading producer of COVID-19 vaccines. The economic values of solving these issues are immeasurable. We plan to incorporate our recently developed anti-bacterial surfaces to alleviate the aforementioned issues. The experimental procedure of producing anti-bacterial surfaces is complex and time-consuming. In this project, we will develop a neural network and genetic algorithm to optimize and speed up these fabrication processes.

ADVIS – Development of a Low-Cost, Low-Power Integrated Machine Health Monitoring Sensor

PI Researcher: Michael Heilemann

Development of a Low-Cost, Low-Power Integrated Machine Health Monitoring Sensor: We seek to develop a device to cost-effectively bring machine health monitoring (MHM) to a broad spectrum of DoD assets (vehicles, pumps, rotating machinery), where the implementation of conventional monitoring systems is cost prohibitive. The design objective is to provide a low-cost, general-purpose hardware platform to support a range of vibro-acoustic MHM and condition-based maintenance applications. The proposed platform will implement a novelty detection autoencoding neural network, but also will support other user-defined machine learning architectures and algorithms consistent with available memory. Availability of this inexpensive and flexible platform enables a wider adoption of MHM for both DoD and commercial applications.

Regeneron Logo Regeneron Pharmaceuticals – A transfer learning ensemble model for cell classification through phase images

headshot of Ling Wang PI Researcher:
Ling Wang

This project aims to develop a label-free and automated cell classification method for pharmaceutical research using phase imaging and deep learning techniques. The proposed approach leverages the strengths of multiple deep learning models and ensemble learning techniques to improve the accuracy and efficiency of cell classification. The project targets phase contrast images, which are label-free and provide detailed information about cell morphology and dynamics. The proposed method has the potential to benefit pharmaceutical research by providing an accurate and efficient tool for cell classification, aiding in drug discovery, toxicity testing, and other aspects of pharmaceutical research that require precise characterization of cell types and behaviors.

Parverio Logo Parverio – Second Generation Computer Vision Tools for the Study of Luekocyte Trafficking Across Vascular Barriers In Vitro

James Mcgrath Headshot PI Researcher:
James McGrath

Our laboratory has developed live microscopy experiments showing the transmigration of white cells (leukocytes) from the blood to tissue side of a “tissue chip” called the μSiM, and machine learning approaches for their analysis. These experiments are increasing popular around the world, creating a commercial opportunity for Parverio, a small NYS company developing AI solutions for the μSiM. This project, Part 1 of a two part project we seek to fund through CoE, will create a comprehensive dataset of experiments needed to train next generation machine learning tools and support a robust commercial solution through Parverio.

Advanced Atomization Technologies – Establishing a Multi-modal Data Fusion Framework for Acquiring Tacit Knowledge from Machinists

PI Researcher: Yunbo Zhang

This project aims to extract tacit knowledge from experienced machinists and transfer it to the next generation of machinists. We will collect multimodal data from experienced machinists through an experimental set for tool wear determination in machining; develop machining Learning methods for tacit knowledge extraction consist of: behavior data fusion based on an integrated Bayesian hierarchical and infinite hidden Markov model, and identifying useful tacit knowledge patterns; and integrate tacit knowledge into the XR system. The project’s ultimate goal is to transfer the developed XR training system to our industry partner, AA Tech, for their training of new machinists.

Immersitech – Development of a Framework for the Evaluation of Spatial Audio System Performance

Mark Bocko, Distinguished Professor, Electrical and Computer Engineering Professor, Physics and Astronomy
Director, Center for Emerging and Innovative Sciences (CEIS) at University of Rochester, is working with Rochester based Immersitech, Inc. to develop a framework to evaluate the performance of spatial sound reproduction. This research combines representations of human peripheral auditory system response and low-level processing of binaural data in the brainstem with learning networks to infer auditory scenes from given acoustic stimuli. The framework provides a tool to evaluate and guide the development of spatial sound reproduction systems.

Velan Studios – Causal Modeling Methods for Big Data: Optimizing Gaming Design Decisions Using Data Science

Causal modeling represents an active area of research for determining causal relationships when a randomized controlled experiment is not available. The RPI team has developed extensive experience building causal models using micro-level data. Jason Kuruzovich (left) and Kristin Bennett (right) will be worked with Velan Studios to apply methods to the context of answering business relevant questions on engagement outcomes in gaming that are also relevant for theory development. Questions may involve training, skill matching, in a skill-based collaborative / competitive game Knockout City. The analytics capabilities will be further incorporated as infrastructure for future games launched by Velan Studios and available to startups funded through Velan Ventures.

LightTopTech – Microscope design for Gabor-domain optical coherence microscopy of the brain and organoids powered by automated image processing and feature extraction

Headshot Photo of Jannick Rolland Jannick Rolland will be working with LighTopTech Corp. to develop a microscope for Gabor-Domain Optical Coherence Microscopy (GDOCM) and related image processing tools will be developed in this project, enabling commercialization of pre-clinical and clinical GD-OCM for brain tissue and organoid characterization.

Kitware – Identify Hidden Biases of AI Algorithms via Human-Machine Collaboration

AI algorithms, especially deep learning, learn biases from data; thus, it is urgent and vital to identify and mitigate biases in these algorithms. However, previous bias identification pipelines rely on human experts to conjecture potential biases (e.g., gender, color, etc.). This approach might work for limited domains and problems but have no chance to scale when considering general AI deployment; humans alone cannot realize all the biased attributes, e.g., identifying biases for a bridge classifier trained on ImageNet. Kitware, Inc. has contacted Chenliang Xu, Associate Professor in the Department of Computer Science and Wilmot Distingshied Professor 2021-2023 at the University of Rochester, to develop novel, scalable human-machine collaborative systems to assist humans in discovering hidden biases in any image classifiers.

Trendly – Few-Shot learning for Fine-Grained Object Recognition

Trendly, Inc. worked with Jiebo Luo on issues with fine-grained object recognition. Prof. Luo will be exploring few-shot learning for fine-grained object recognition by utilizing contrastive learning to extract a discriminative representation of objects to facilitate learning from few examples. The concepts will be verified by a high-precision model for luxury bag authentication and generalized to other fine-grained object recognition tasks such as fine art, artifacts, and jewelry.

L3Harris – Scalable Architectures and Increased Capacity for Bi-stable Resistively-coupled Ising Machine – BRIM

Dr. Michael Huang (right) and Dr. Zeljko Ignjatovic (left) worked with L3Harris to explore two scalable architectures (a microchip and a multi-processor types) based on a previously proposed CMOS-compatible Ising machine will be explored and characterize to address issues using an Ising machine.

IBM – Learning to Localize Sources of Network Diffusion

PI Researcher: Gonzalo Mateos Buckstein

Learning to Localize Sources of Network Diffusion: We propose a deep learning solution to the inverse problem of localizing sources of network diffusion. Invoking graph signal processing (GSP) fundamentals, the problem boils down to blind estimation of a diffusion filter and its sparse input signal encoding thesource locations. While the observations are bilinear functions of the unknowns, a mild requirement on invertibility of the filter enables a convex reformulation that we solve via the alternating-direction method of multipliers (ADMM). We unroll and truncate the novel ADMM iterations, to arrive at a parameterized neural network architecture for Source Localization on Graphs (SLoG-Net), that we train in an end-to-end fashion using labeled data. This way we leverage inductive biases of a GSP model-based solution in a data-driven trainable parametric architecture, which is interpretable, parameter efficient, and offers controllable complexity during inference. By advancing innovative machine learning technologies to tackle data science problems encountered with sensor, information, social, and brain networks, this university-industry collaboration is primed to generate economic and broader societal impacts.

Flaum Eye Institute – 3D eye imaging and machine learning strategies to improve cataract surgery

Cataract surgery is the most often performed surgery in any hospital of the world (28 million/year). However, the process by which the intraocular lens to replace crystalline lens is selected relies on limited anatomical information and rudimentary formulas. Susana Marcos worked with the Flaum Eye Institute at the University of Rochester will attempt to propose the use of 3-D quantitative optical coherence tomography images and machine learning approaches to obtain an accurate expression of the estimated lens position based on the pre-operative anterior segment anatomy and full crystalline lens shape. This method will improve the refractive outcomes of cataract surgery, increasing patient satisfaction and reducing the burden of refractive error correction.

Monro Inc. – Developing Analytics for Optimizing the Customized Discount Provided by Salesforce for Automobile Service

Yufeng Huang (left) and Mitchell Lovett (right) worked with Monro Inc. will study how to help salesforce provide better customized discount to consumers.

Leveraging data from a large automobile service company, this project will use business analytics methods and develop a structural econometric model to characterize consumer demand and bargaining with salesperson. They will use the model to predict the market outcome under different interventions to salespersons, including setting a monthly discount quota, or completed forbidding discount.

Pfizer – Using Neural Network and Genetic Algorithm to Optimize Laser Surface Functionalization for Biomedical Applications

Chunlei Guo, worked with Pfizer in developing advanced materials for biomedical applications, including preserving fluidic drug delivery and increasing delivery accuracy. Pfizer is a leading producer of COVID-19 vaccines and the economic values of solving these issues are immeasurable. We plan to incorporate our pioneered superhydrophobic surfaces to Pfizer applications to alleviate the aforementioned issues. The experimental procedure of producing superhydrophobic surfaces is complex and time-consuming. In this project, we will develop a neural network and genetic algorithm to optimize these fabrication parameters to speed up the process and achieve the optimized surface property for biomedical applications.

Immersitech – Development of a Framework for the Evaluation of Spatial Audio System Performance

Mark Bocko, Distinguished Professor, Electrical and Computer Engineering
Professor, Physics and Astronomy
Director, Center for Emerging and Innovative Sciences (CEIS) at University of Rochester, is working with Rochester based Immersitech, Inc. to develop a framework to evaluate the performance of spatial sound reproduction. This research combines representations of human peripheral auditory system response and low-level processing of binaural data in the brainstem with learning networks to infer auditory scenes from given acoustic stimuli. The framework provides a tool to evaluate and guide the development of spatial sound reproduction systems.

IngenID – Developing and Deploying Spoofing Aware Speaker Verification Systems

Headshot Photo of Zhiyao Duan Zhiyao Duan is working with IngenID

LightTopTech – Microscope design for Gabor-domain optical coherence microscopy of the brain and organoids powered by automated image processing and

Pfizer – Neural Network assisted Femtosecond Laser Fabrication of Anti-bacterial Surfaces

Chunlei Guo, Professor of Optics
Senior Scientist in the Laboratory for Laser Energetics at the University of Rochester is working with Pfizer on Neural Network assisted Femtosecond Laser Fabrication of Anti-bacterial Surfaces.

AMD – Architectural Support to Increase Ising Machine Capabilities

Dr. Huang (left) and Dr. Ignjatovic (right) from the University of Rochester are working with AMD researching Architectural Support to Increase Ising Machine Capabilities.The goal of the project is to develop advances in Ising machine technologies that will bring them closer to realizing their envisioned potential of orders-of-magnitude improvements in speed and energy efficiency when used to solve hard optimization problems (such as SAT). A concrete technical goal is to move from usual quadratic interactions to also account for qubic interactions that can facilitate mapping (and then solving) a broader class of problems more efficiently to the proposed Ising machine substrate.

ACV Auctions – Auto Auction Data as a Leading Indicator of Economic Activity and Vehicle Valuation

Jason Kuruzovich, Associate Professor and Academic Director, Severino Center for Technological Entrepreneurship at RPI is working with the ACV Auctions.

IBM – Learning to Localize Sources of Network Diffusion

Gonzalo Mateos, Associate Professor with the Dept. of Electrical and Computer Engineering, University of Rochester, is working with IBM on Learning to Localize Sources of Network Diffusion.

Kitware – Domain Adaptation using Vision Transformers

Andreas Savakis Headshot Andreas Savakis, Professor Department of Computer Engineering in the
Kate Gleason College of Engineering at Rochester Institute of Technology is working with Kitware, Inc. on Domain Adaptation using Vision Transformers.