# Benchmarking TinyML Continual Learning under Concept Drift
> Explore a research framework for evaluating streaming AI on ARM Cortex-M4 hardware, focusing on accuracy, RAM, and energy trade-offs for on-device learning.

Tags: tinyml, continual-learning, concept-drift, edge-ai, arm-cortex-m4, machine-learning-benchmark, embedded-systems
## Evaluating Streaming Continual Learning on TinyML Hardware
- **Objective:** Master thesis proposal to evaluate AI learning on true TinyML hardware under conditions of concept drift.
- **Presentation by:** Alba Huti, JKU - Streaming AI Initiative.

## Proposed Analysis Tools & Contributions
- **Timeline:** 6-month research plan covering Literature Review, Implementation, Experimentation, and Writing.
- **Open-source Contribution:** Development of a benchmark framework for evaluation on Microcontroller Units (MCUs).
- **Data focus:** Empirical trade-off data involving Accuracy vs. RAM, latency, and energy consumption.
- **Alignment:** Focus on on-device learning to reduce cloud dependency and environmental impact (CO₂).

## Research at a Glance
- **Algorithms:** Implementation of 2-3 Continual Learning algorithms including Online SGD, Replay Buffer, and EWC-lite.
- **Hardware:** ARM Cortex-M4 MCU with 256KB RAM and 64MHz clock speed.
- **Metrics:** Evaluation based on Accuracy, Latency, Memory, and Energy usage.
---
This presentation was created with [Bobr AI](https://bobr.ai) — an AI presentation generator.