Hunting Dinosaurs with Intel AI Technologies
In this workshop, we will explore an end-to-end AI application with the goal of creating a dinosaur bone bed likelihood map to help guide your next dinosaur hunt! Join us as we use machine learning algorithms, powered by oneAPI, such as Intel Extension for Scikit-learn, NumPy, and OpenVINO as part of this AI dinosaur adventure.
We will first use Intel Extension for Scikit-learn to preprocess metadata and experiment with semi-supervised learning and clustering techniques on known dinosaur bone beds. We will then apply metadata labeling to aerial photos to be fed to a convolutional neural network using the Intel oneDNN library to classify the aerial images. Finally, we will use Intel OpenVINO to streamline the inference capability on Intel CPUs to process thousands of images and create a bone likelihood map that can be applied to other regions of the globe than the one they were initially trained on!
Note: Because dinosaur bones on US Federal lands are protected – we will have to obfuscate the actual map in deference to the dinosaur privacy.
Speaker(s)
Bob Chesebrough
Robert Chesebrough is currently a Technical Evangelist in the Intel Developer Academy. His educational background is physics. His industry experience has been software development and application/performance engineering for fortune 100 companies and national laboratories for over three decades. He is a data scientist, using machine learning/ deep learning for eight years while working for Intel and other high-tech companies.
Raymond Lo
Raymond Lo is the software evangelist at Intel focusing on the OpenVINO toolkit. Previously, Raymond was the founder and CTO of Y Combinator-backed augmented reality company Meta, the technology evangelist for Samsung NEXT, and Technical Solutions Consultant for Google Cloud AI. During his PhD, Raymond worked with Prof Steve Mann, who is widely recognized as the father of wearable computing from MIT Media Lab. Together they published numerous of research papers on the topic of GPGPUs, open-source HDR videos, and open-source wearable computing projects in the University of Toronto. In 2013, Raymond took his augmented reality (AR) eyeglasses invention and functional prototype from Toronto and co-founded Meta that raised over $80 million for their pioneering development of augmented reality eyeglasses. As the CTO of Meta, he successfully developed and shipped Meta 1 and Meta 2 with a team of over 150+ people. As the executive and Evangelist, he had the opportunity to speak at conferences such as TED talk, SIGGRAPH, Tencent WE Summit, and Hello Tomorrow. His unique combination of startup entrepreneurship and enterprise experiences helped Raymond to create connections to hundreds of startups and thousands of developers around the world.
Rahul Unnikrishnan Nair
Rahul is a Machine Learning Architect at Intel, with ten years of experience in data science, Machine Learning, and four years in applied Deep Learning. He joined Intel seven years ago in the OpenStack Cloud team after finishing a Master’s degree in computer engineering at the University of Texas. As a core member of the OpenStack security group, Rahul built tools for automated security analysis on the cloud. Following that, he joined the AI group in the Opensource Technology Center to develop Deep Learning and Deep Reinforcement Learning algorithms optimized for Intel CPUs. In this role, he developed Deep Learning pipelines that scale on the cloud, worked with startups, OEMs, and internal intel teams in optimizing AI workloads on Intel hardware, and open-sourced the learnings and code externally as blogs and code on GitHub. Rahul was also part of an Intel incubation startup as the technical lead in developing an AI platform for drug discovery, focusing on system design for AI and scaling machine learning solutions on the cloud. He was the architect and technical lead of the first group at Intel to release deep learning containers and held the role of container architect for oneAPI products.
