Media & Publications
This page lists various publications, patents and presentations I have done over the years.
Publications
- I. Preet, O. Boydell and D. John, “Class-Separation Preserving Pruning for Deep Neural Networks,” in IEEE Transactions on Artificial Intelligence, doi: 10.1109/TAI.2022.3228511, link.
- Aswin K Ramasubramanian, Robins Mathew, Inder Preet, Nikolaos Papakostas, “Review and application of Edge AI solutions for mobile collaborative robotic platforms”, Procedia CIRP, Volume 107, 2022, link.
- Alastair McKinstry, Oisin Boydell, Quan Le, Inder Preet, Jennifer Hanafin, Manuel Fernandez, Adam Warde, Venkatesh Kannan, and Patrick Griffiths, “AI-Ready Training Datasets for Earth Observation: Enabling FAIR data principles for EO training data”. EGU General Assembly, 2021, link.
Patents
- Class Separation Aware Artificial Neural Network Pruning Method, International (PCT) Application No. PCT/EP2022/085998
- System and method for optimized PV placement and battery energy storage capacity.System and method for optimized PV placement and battery energy storage capacity. P23-1731WO01
Videos
Pruning & Quantization for Deep Neural Networks> Invited for a talk on my publication on optimizing DNNs for deployment at the edge.
Hear experts from IBM, Inpher and Oblivious.AI for a discussion on three technologies that can be used to protect data and models - differential privacy, secret computing and secure enclaves.
Main challenges for deploying and training ML algorithms is discussed here and some solutions both on the hardware side and design of neural networks are presented.
A proof of concept for automating the optimization, pruning and quantization of DNNs for edge deployment is presented for the problem of anomaly detection in heartbeats using wearable sensors.