機器學習與模擬加速器.pdf

Major Programme: Master of Science in Microelectronics & Master of Philosophy in Microelectronics □ CM – Compulsory Major Course Type: □ L&S – Languages and Skills □ CPE – Community and Peer Education  RE – Required Elective Course Title: Machine Learning and Analog Accelerators (in Chinese and English) 機器學習與模擬加速器 Duration:  Semester Course Grading System:  Letter Grade Medium of Instruction: Course Description: Intended Learning Outcomes (ILO): □ * GE – General Education Suggested Year of Study: Year 1 □ Yearly Course Credit Units: 3 □ P/NP Pre-requisite: None (if any) □ MI – Minor □ FE – Free Elective English This is an introductory course in machine learning tailored for IME students. It covers the topics from classification, regression and statistical signal processing, to more recent techniques such as neural network and deep learning. It also covers the analog approximate computing integrated circuit design considerations for acceleration purpose. The course aims to offer students the fundamental concepts in the advanced artificial intelligence theory with emphasis on hands-on experience through practical examples such as intelligent hardware system implementation and case studies with MATLAB/Python. The verified algorithm can be further implemented on an FPGA for applications such as image/audio recognition. This course enables students to have:  To introduce the essential knowledge in machine learning and deep learning.  To introduce analog accelerators with practical circuit considerations.  To teach students with hands on experience on designing, training and verifying neural networks for image/audio classification problems using MATLAB/Python.  To teach students with hands on experience on implementing the neural networks on an FPGA board for real-time classification. Others (please specify) Listening & Oral Assessments / tests Reading & Writing Assessments / tests Company visits Field study √ Internship Exercises & problems √ Service learning Writing Assignment Group discussions Group project / paper Individual project / paper Student Presentation Role Playing Case Study Major Assessment Methods: Class Participation / Discussion _______% Assignment(s) 30 % Test(s) % Examination % Others: Project 70 Course Content: (topic outline) Template revised on 20 Oct 2017 % √ √ ‐ Introduction: basic concepts and the evolution of the artificial intelligence with examples and applications ‐ Preliminaries: matrix algebra, probability, random process ‐ Machine learning techniques: classification, regression and statistical signal processing ‐ Neural networks and deep learning: perceptron, feed-forward multilayer neural networks, backpropagation algorithm, deep networks, deep belief networks ‐ Approximate computing cases study ‐ Practical labs: classification, regression, prediction for practical application such as image/audio signal processing using MATLAB/Python and implementing on an FPGA 1