The TinyML Contest at IEEE COINS is a challenging, multi-month, research and development competition, focusing on advanced, real-world problems in the fields of AI, Embedded Systems, and IoT. It is open to multi-person teams worldwide. The focus of this year’s contest is on Human activity recognition. Human activity recognition is a common machine learning task that is widely used in multiple domains. This technology uses data from various sensors to help monitor health conditions, recognize abnormal human activities for surveillance, track fitness workouts, and much more.
There are, of course, various challenges associated with implementing complex human activity recognition. To achieve the required accuracy on low-power devices like wearables, embedded engineers have to find ways to run machine learning models on platforms with limited memory and processing capacity. Experts, therefore, need to design neural networks with a size-accuracy ratio that is appropriate for a proposed machine learning task on a memory-constrained device.

Please see more infromation in the PDF document:

paper-ieee-coins-neuton-human-activity-recognition-fin

Tentative Schedule:

Submission date: 31st May 2023
Conference date: 23-25th July
Notification date: during the conference (will be announced later)

Registration:
Interested parties must email the ST, Neuton, and and COINS contacts listed below.

Contact persons:
ST: marco.bianco@st.com
Neuton.AI: roman.rusak@neuton.ai
IEEE COINS: farshad.firouzi@duke.edu

Awards donated by ST Microelectronics to the top three entries:

1st place wins $1,500
2nd place wins $1,000
3rd place wins $500