Integrated machine-learning hardware for near-sensor computing applications
IEMN, Villeneuve d’Ascq, France

Organizers : Kévin Hérissé, Benoit Larras, Antoine Frappé (Univ. Lille, Yncréa Hauts-de-France, IEMN) Contacts : ; ;

With the growing amount of Smart Sensors, decreasing the energy consumption of the devices must be a priority to increase the batteries lifetime and enable wearable and continuous monitoring. Since communication interfaces are the most energy-hungry parts of the sensor nodes, the “Near-Sensor Computing” concept aims at pre-processing the input data in order to keep only relevant information and thus limit the amount of data to transmit. Machine learning techniques are used to determine the relevance depending on the targeted application. The objective of this workshop is to detail how the embedded processing circuits can be integrated into the hardware and interfaced as close as possible to the sensor.

Three excellent speakers are scheduled to cover many aspects of integrated processing and machine learning hardware. The application fields range from biomedical signals (EEG, ECG) to audio signals (silicon cochlea, voice activity detection). The contributions will address circuit-level, system-level and integration issues.

Jerald Yoo, National University of Singapore, IEEE CASS Distinguished Lecturer On-Chip Epilepsy Detection: Where Machine Learning Meets Wearable, Patient-Specific Wearable Healthcare

Minhao Yang, EPFL
Towards Near-Zero-Power Audio Inference Sensing

Deepu John, UC Dublin
Low Power Sensor Design for Wearable Health Monitoring

Additional talks will be scheduled following the call for contributions. Registration is mandatory:

Call for contributions :
We invite you to present your work during this technical workshop. The format will be 30-min slots (20’ presentation – 10’ questions).
Please send us a title and a short abstract (1/2 page) before September 30th at and The final program will be issued on October 15th!