AI for Good blog

The days of “bigger is better” are over: Say hello to the power of tinyML

Artificial Intelligence | Machine Learning

Big problems sometimes require small solutions. That is one of the key takeaways from the first ever tinyML challenge focused building solutions for smart weather stations.

This year ITU has partnered with the tinyML Foundation to launch a challenge to the machine learning (ML) and Artificial Intelligence (AI) communities. tinyML refers to the use of machine learning algorithms on devices with small computational resources, such as microcontrollers or single-board computers. This technology helps edge devices to process and respond to data in real time, without using as much computational resources.

2022 was an exciting year, as participants were challenged to create a low-cost, low-power, reliable, accurate, easy to install and maintain weather station based on tinyML. This competition was co-organized by the tinyML Foundation and supported by other partners.

Overall, the challenge received overwhelming response from the tinyML community and 60 teams signed up which comprised 102 students and professionals. Over the course of eight months, the participants worked on various stages or check-points of the competition.

During the Challenge finale on 15 December 2022, six teams were invited to present their smart weather station solutions. Presentations by teams were interesting, of high quality and were highly rated by judges.

The first place prize was awarded to team CSEM which comprised of prize money 3’000 USD, a certificate and SenseCAP S1000 10-in-1 Compact Weather Sensor with CO2 Measurement sponsored by SeeedStudio. The Team developed a smart weather station called Aurora with several capabilities to track the temperature, humidity and pressure in the environment as well as detecting wind and rain.

The second place prize was awarded to team BrainCircuit which also received the Best use of Edge Impulse prize.

The third place prize was tied and the judges decided to split the award amongst three teams. These are;

· Icfoss

· AI4AFRICA

· Tinycow

Based on the presentations from teams during the Challenge finale, Judges decided to award an additional prize of Best Sustainable Solution to Kasun.

We encourage teams that have been awarded prizes and those that did not make it to the tinyML Challenge finale to continue developing the projects/prototypes further. The solutions are planned to be showcased in upcoming tinyML for Good Forum in February 2023 as well as AI for Good events.

You can watch a replay of the tinyML Challenge Finale here or access the github if you are interested in seeing the solutions in detail.

We would like to thank the tinyML Foundation for the support as well as Marco Zennaro (a Research Scientist at the Abdus Salam International Centre for Theoretical Physics (ICTP)) for the collaboration and putting together the idea for the 2022 tinyML Challenge.