Data Analytics | Office of Technology Licensing

Available Technologies by Category
Predictive Maintenance System for Large-Scale Operations Offers Potential for Significant Cost Savings	Consistent workflow improvements: Offers repeatable and consistent enterprise-level improvements to the maintenance workflow. Proactive readiness: Improves maintenance scheduling, allowing for improved logistics, scheduling, virtual kitting, and part/component availability. Customizable: Customers can build and tailor their PMx solutions according to their needs, providing increased data control and an alternative to costly and limited solutions.	8793
Image-Based Disease Severity Algorithm	The technology utilizes an image-based algorithm and automatic methodology to detect biomarkers in disease. It leverages a well-trained neural network on a dataset of healthy instances to learn the feature distribution of the healthy data. Anomaly detection algorithms generate anomaly scores based on properties of the neural network, providing a measure of severity by assessing how anomalous an instance appears compared to the healthy distribution.	9097
Generative Machine-Learning Model for Preference-Based Attribute Editing and Photorealistic Image Generation	User-controlled image attribute editing and generation through paired queries Effective extraction of difficult-to-articulate preferences for personalized content Applications in social media, e-commerce recommendations, and entertainment industries	9200
Custom-Fit Reusable Respiratory Protective Device (RPD) with Continuous Fit Monitoring Improves Comfort and Protection	Vastly improves comfort and protection while bringing peace of mind to wearers through a custom fit and continuous monitoring Reduces pressure injuries caused by extended RPD use through both a customized fit and alerts to the wearer when the RPD should be adjusted or removed Minimizes exposures to workplace hazards by immediately detecting leakage during RPD use and alerting the wearer to adjust the device	8876, 9089, 9090, 9091
Deployable Critical Active Learning (DECAL) System Offers Better Characterization for Improved Diagnostics and Imaging	The deployable critical active learning (DECAL) system’s bi-modal interface enables better characterization with large training sets that generate more accurate output for higher precision. The algorithm used during the initialization period creates a framework that can be expanded and generalized to any data method and application. The DECAL algorithm samples uniquely important, unlabeled data to aid existing active learning systems in generalizing more quickly, providing faster diagnosis statistics.	9099
High-Precision, Hands-Free Remote Control of Complex Robotic Systems Via Eye Movements	This two-camera eye-tracking system (TCES) integrates a commercial eye tracker with machine-learning technology for continuous real-time classification of gaze and eye directions for robotic arm control. The system provides highly accurate classification for four directions of eye movement and has achieved 99.99% accuracy in studies. Using simple eye movement, the TCES offers low-cost, high-precision control of external devices/hardware for people with disabilities, surgery robots, warehouse systems, construction tools, and more.	9184
Enhanced Cybersecurity for Networked Motion Control Systems with Somewhat Homomorphic Encryption	Increase cybersecurity with expression rewrite rules for encrypted dynamic control schemes that reduce the multiplicative depth of somewhat homomorphic encryption and improve numerical stability. New approach generates motion commands to servo systems without creating a security hole. Information decryption and control signal calculation can be performed and executed inside the plant, while all sensitive system information outside of the plant is always encrypted.	9012
Airborne Acoustic Sensing System for “Ear in the Sky” Capabilities on Unmanned Aerial Vehicles (UAVs)	Optimizes speech clarity: Uses acquired empirical data and iteratively refines the processing capability until the optimal signal to noise ratio (SNR) is achieved Improves covert detection: Enables acquisition of a live voice signal from a person of interest without detection by using UAS platforms to remotely acquire acoustic signals and using digital signal processing algorithms to filter, enhance, and localize the speech signal of interest Enhances warfare tools: Provides increased situational awareness, enhanced ISR capabilities, and helps identify potential security threats to deployed assets	6985, 7223
RobocallGuardPlus	Advanced: Uses a unique smart virtual assistant with a voice interaction model that generates natural conversation to filter out robocalls Robust: Blocks robocalls that other methods may miss, such as those that evade call block lists through caller ID spoofing Unique: Provides the first known system to block mass, evasive, and targeted robocalls using natural, interactive conversation methods rather than relying solely on caller ID and block lists	8817
Generative Causal Explanations for Black-Box Classifiers	Increased trust: Has the potential to improve the transparency and fairness of machine learning systems and increase the level of trust users place in their decisions Flexible explanation vocabulary: Provides a rich and flexible vocabulary for explanation that is more expressive than feature selection or saliency map–based methods Deeper insights: Offers visualizations that can be much more descriptive than saliency maps, particularly in vision applications, leading to better user understanding	8507
IMUTube: Converting Videos of Human Activity into Virtual IMU Data Streams	Automated: Automates wearable sensor data collection, which can in turn be used to train HAR models by simply giving them access to large-scale video repositories Accurate: Demonstrated recognition accuracy in some cases comparable to models trained only with non-virtual (“real”) sensor data in benchmark studies Superior: Outperformed models using real sensor data alone by adding only small amounts of real sensor data to the virtual sensor dataset in benchmark studies	8538
Miniaturized Modular-Array Fluorescence Microscopy	Powerful: Offers high fluorescence sensitivity, efficiency, and spatiotemporal resolution (~3 µm and up to 60 Hz) Configurable: Offers compatibility with conventional cell culture assays and physiological imaging, providing accessibility to upright physiological imaging and integration with biochemical sensors under the cell platform Efficient: Provides effective parallelization of multi-site data acquisition	8580
Improve the Reliability of Quantum Computers by Mitigating Measurements Errors	Increased reliability by 2-3x: Some current quantum computers have average measurement error rates ranging from 6%–8%, with worst case rates ranging from 25%–30%. Georgia Tech’s methods improve measurement fidelity by 2–3x. Improved usability: Implementation of the EDM uses the top-four mappings produced by the underlying mapping policy. Proactive: Using AIM to predict the state that is being measured and the error rate proﬁle of the machine, the predicted state can be proactively mapped to the strongest state using a speciﬁcally designed inversion string.	8207, 8208
Enabling Privacy-Preserving Search Over Fuzzy Databases	Privacy-preserving: Improves the security of biometric-based surveillance, identification, or searches for individuals using their biometric data over private databases Timely: Addresses the requirements of recent and emerging privacy protection regulations and policies Practical: Fills a gap in privacy-preserving search technologies, which currently do not accommodate searching of fuzzy data such as biometrics	8506
Virtual Sensing Integrated for Predictive Reliability (VIPR)	Advanced: Acquires system data via virtual sensing, without the need for physical sensors in all desired locations Cost-reducing: Aims to reduce maintenance and planning costs for solid rocket motors and potentially other vehicle propulsion systems Predictive: Provides useful, repeatable, extensive data sets to predict rocket performance and enable proactive plans for propulsion system maintenance, upgrades, and replacements, potentially reducing downtime and failures	8555
Microfluidic CODES with Innovative Machine-Learning Analysis	Scalable: Set-up allows for the integration of multiple microfluidic devices to increase throughput volume. Versatile: Machine learning analysis can be applied to other microfluidic devices integrated with the same Coulter sensor network. Pragmatic: This innovation improves the capabilities of lab-on-a-chip systems to provide an affordable solution for low-resource settings.	7107, 8032
Utilizing Impedance Spectroscopy for Advanced Characterization of Particles	Comprehensive: Simultaneously measures both spatial and dielectric properties of particles Streamlined: Helps resolve signal interference from coincident cells—a challenge of other assessment techniques Precise: Offers complex impedance as a new level of particle feature assessment, integrating it with other measurements such as elasticity and size for more predictive analyses	8049
Privacy-Preserving Remote Biometric Authentication	Protective: Designed to eliminate privacy concerns associated with databases of people’s faces Flexible: Supports various biometrics, including face, voice, iris, fingerprint, and gait Remote authentication: Eliminates the need to always possess the enrollment device	7885
‘Intelligent’ Power Amplifier Architecture	AI-assisted: Provides the PA self-reconfigurability and robust adaptive operation over environmental changes High performance: Features low latency and fast response capabilities required for many fifth-generation (5G) applications Improved learning efficiency: Incorporates into control algorithms multiple RL frameworks, including CAB, multi-armed bandit, contextual bandit, and actor-critic with experience replay	8377
Contextually Aware, Machine Learning–Based Smart Notification System	Increases user efficiency Increases user productivity Reduces user stress	8034

Technologies