Course

Syllabus

Review of Machine Learning Basics: learning types, classification/regression, model types, dimensionality issues, linear/logistic regression. Evaluation metrics for classification, regression, and anomaly detection. TinyML overview and key use cases. 

Support vector machines for regression and classification. Neural Networks, Ensemble methods, Introduction to Convolutional Neural Network (CNN), Reinforcement Learning. End-to-end development of ML applications using Edge Impulse: data collection from sensors, Data pre-processing, feature extraction, feature selection, model training. 

Model optimization techniques for embedded ML: quantization, pruning, model size vs performance trade-offs. Hardware constraints and selection criteria for ML deployment. Multi-sensor fusion. Deployment frameworks: TensorFlow Lite, TensorFlow Lite for Microcontrollers (TFLM). 

Pre-requisite: Nil

Course Objectives 

• To introduce foundational concepts in machine learning and their relevance to embedded systems. 
• To develop practical skills for designing, training, and deploying ML models using platforms like Edge Impulse. 
• To explore model optimization and deployment strategies tailored for resource-constrained embedded environments. 

Course Outcomes 

CO1: Explain key machine learning concepts and evaluate models using appropriate metrics for classification, regression, and anomaly detection.

CO2: Apply algorithms such as SVMs, neural networks, and CNNs to solve classification and regression problems using embedded datasets.

CO3: Build end-to-end ML applications on Edge Impulse, including data acquisition, feature engineering, model training, and validation. 

CO4: Optimize and deploy trained models to embedded platforms using frameworks such as TensorFlow Lite and TFLM, considering hardware constraints and sensor fusion. 

CO-PO Mapping 

1. Pete Warden, Daniel Situnayake, TinyML: Machine Learning with TensorFlow Lite on Arduino and Ultra-Low-Power Microcontrollers, O’Reilly Media, 2020.
2. Simon Haykin, Neural Networks and Learning Machines, Pearson, 2020.
3. Xiaofei Wang, Yi Pan, Edge AI: Machine Learning for Embedded Systems, Springer, 2022.
4. Daniel Situnayake, Ian Buckley, Practical TinyML: Deploying Machine Learning on Microcontrollers with TensorFlow Lite, O’Reilly Media, 2022.
5. Tom M. Mitchell, “Machine Learning”, McGraw-Hill, 1 st edition 1997.