tinyML Courses

Oral Tongue Lesion Detection Using TinyML on Embedded Devices

A case study of using tinyML models on embedded devices to detect benign and premalignant tongue lesions for low-cost oral cancer screening.

• YouTube
• 34 minutes
• Self-Paced
• Free Video

CMSIS-NN & Optimizations for Edge AI

Optimize Edge AI inference on Arm Cortex-M processors using CMSIS-NN, cycle analysis, operator tuning, and TensorFlow Lite for Microcontrollers.

• YouTube
• 48 minutes
• Self-Paced
• Free Video

MicroTVM: A Tensor Compiler for Bare Metal

Explore microTVM, a tensor compiler that converts machine-learning models into deployable code for bare-metal microcontrollers, with demonstrations of compilation, quantization, and execution.

• YouTube
• 58 minutes
• Self-Paced
• Free Video

Binarized Neural Networks on Microcontrollers

Learn how binarized neural networks reduce memory and computation for real-time deep learning on microcontrollers, with training, benchmarking, and person-detection deployment.

• YouTube
• 59 minutes
• Self-Paced
• Free Video

Security of Edge AI Against Hardware Attacks

Examines how side-channel and fault attacks can reverse-engineer neural networks running on edge devices, with masking and other hardware countermeasures.

• YouTube
• 57 minutes
• Self-Paced
• Free Video

Towards Software-Defined Imaging - Adaptive Video Subsampling for Energy-Efficient Object Tracking

Explores software-defined imaging that uses adaptive video subsampling and programmable CMOS sensors for energy-efficient object tracking.

• YouTube
• 37 minutes
• Self-Paced
• Free Video

Power Efficient Neural Network Based Hardware Architecture for Biomedical Applications

A research presentation on SABiNN, a power-efficient binarized neural network hardware architecture for real-time sleep apnea detection on wearable devices.

• YouTube
• 27 minutes
• Self-Paced
• Free Video

Software-Driven TinyML Hardware Co-Design for AI Power Smart Devices

Explores end-to-end TinyML software–hardware co-design using processing-in-memory accelerators and PIM-optimized neural networks to reduce inference power.

• YouTube
• 28 minutes
• Self-Paced
• Free Video

In-memory Computing and Dynamic Vision Sensors for Internet of Video Things

Explore dynamic vision sensors and in-memory computing for energy-efficient tinyML systems in Internet of Video Things applications.

• YouTube
• 32 minutes
• Self-Paced
• Free Video

TinyML Talks - Low-Cost Energy-Aware Sensor Data Acquisition at Scale

Low-cost, low-power sensor networks collect data at scale for tinyML through LoRaWAN, battery-aware design, and device-to-cloud time-series replication.

• YouTube
• 49 minutes
• Self-Paced
• Free Video

On-Device Continuous Event-Driven Deep Learning to Avoid Model Drift

Explores continuous, event-driven deep learning on embedded devices to reduce model drift, data bias, and user intervention in changing environments.

• YouTube
• 21 minutes
• Self-Paced
• Free Video

Continual On-Device Learning on Multi-Core RISC-V Microcontrollers

Explores continual learning on multi-core RISC-V microcontrollers, using latent replays, quantization, and parallel training kernels to adapt TinyML models on-device.

• YouTube
• 25 minutes
• Self-Paced
• Free Video

Optimizing Data-Flow in Binary Neural Networks

Optimizing data flow in binary neural networks through quantization and normalization choices, with attention to accuracy results.

• YouTube
• 20 minutes
• Self-Paced
• Free Video

Autonomous Nano-UAVs: An Extreme Edge Computing Case

Explore how tiny CNNs and ultra-low-power microcontrollers enable real-time autonomous navigation and human-robot interaction on nano-UAVs.

• YouTube
• 31 minutes
• Self-Paced
• Free Video

TinyML AutoML Tutorial: Building Production-ready Models with Imagimob AI

Learn to use Imagimob AI’s low-code workflow to build production-ready time-series neural network models for resource-constrained devices.

• YouTube
• 57 minutes
• Self-Paced
• Free Video