दूरसंचार प्रणाली एवं प्रबंधन

Workflow of Artificial Intelligence: The Relationship between Speed and Accuracy

Abstract

Patrick Bangert

Abstract: The workflow of artificial intelligence projects begins with cleaning and labeling data. It proceeds over feature engineering, model selection, and hyper-parameter tuning to finally get to the training of the model. Once the model is ready, it can be used at inference to solve the original task. Each one of these major steps in the workflow can be assisted by tools that make that step either easier, faster or better than the, largely manual, process of today. Data labeling is augmented by Semi-automatic labeling methods. Feature engineering can be automated by creating synthetic features on the fly. Model selection and hyper-parameter tuning are essentially search problems that can also be automated. These possibilities will be illustrated and the common denominator discussed: All of these tools assume that the training of the model is fast. That can be accomplished by training the model on many computers simultaneously – distributed training. We illustrate the saving of time and the improvement of model accuracy in various domains such as natural-language-processing and image classification and segmentation Biography: Patrick is the Vice-President of Artificial Intelligence at Samsung SDSA where he heads the AI Engineering and AI Sciences teams. He is responsible for Brightics AI Accelerator, a distributed ML training and automated ML product that is also included in the Brightics AI platform. He is responsible for X.insights, a data center intelligence platform. Among his other responsibilities is to act as a visionary for the future of AI at Samsung. Before joining Samsung, Patrick spent 15 years as CEO at Algorithmica Technologies, a machine learning Software Company serving the chemicals and oil and gas industries. Prior to that, he was assistant professor of applied mathematics at Jacobs University in Germany, as well as a researcher at Los Alamos National Laboratory and NASA’s Jet Propulsion Laboratory. Patrick obtained his machine learning PhD in mathematics and his Masters in theoretical physics from University College London. A German native, Patrick grew up in Malaysia and the Philippines, and later lived in the UK, Austria, Nepal and USA. He has done business in many countries and believes that AI must serve humanity beyond mere automation of routine tasks. An avid reader of books, Patrick lives in the San Francisco Bay Area with his wife and two children. realization of Sigma Delta Radio over Fiber System. Dr Hadi is a Member IEEE, IEEE Photonics Society and IEEE Communications Society. He serves as a reviewer for IET Optoelectronics, IEEE Communications Letters and ASTES. Publication of speakers: 1. Bangert, Patrick. (2012). Optimization for Industrial Problems. 10.1007/978-3-642-24974-7. 2. Bangert, Patrick. (2012). Statistical Analysis in Solution Space. 10.1007/978-3-642-24974-7_2. 3. Bangert, Patrick. (2012). Overview of Heuristic Optimization. 10.1007/978-3-642-24974-7_1. 4. Bangert, Patrick & Jörg-A, C.. (2011). Increase of overall combined-heat-and-power efficiency through mathematical modelling. 91. 55-57.