Below static programme is indicated in UTC (Coordinated Universal Time).
If you check the agenda details through the buttom Programme on the left of this page, instead, time appears adjusted to your local time zone (i.e. adjusted to your computer system time). For info see Time.
DAY 1 - Monday, 6 June 2022
07:00 Welcome to the Workshop 30' Speaker: Marco Zennaro (ICTP)
07:30 Introduction to TinyML 1h0' Speaker: Vijay Janapa Reddi (Harvard University)
08:30 Introduction to ML 1h30' Speaker: Mehran Behjati (National University of Malaysia)
DAY 2 - Tuesday, 7 June 2022
07:00 Introduction to ML part 2 1h0' Speaker: Mehran Behjati (National University of Malaysia)
08:00 Edge Impulse Intro and Workflow 1h0' Speaker: Louis Moreau (Edge Impulse)
09:00 Seeed TinyML kit intro and interface to EI 1h0' Speaker: Momodou B Jallow (SeeedStudio)
DAY 3 - Wednesday, 8 June 2022
07:00 Hands-on Motion classification/Anomaly detection 1h30' Speaker: Momodou B Jallow (SeeedStudio)
08:30 University Case Studies 30'
09:00 Ethical issues in TinyML 1h0' Speaker: Susan Kennedy (Santa Clara University)
DAY 4 - Thursday, 9 June 2022
07:00 Hands-on Sound/KWS 1h30' Speaker: Momodou B Jallow (SeeedStudio)
08:30 Hands-on Computer Vision 1h0' Speaker: (Video) Brian Plancher (Harvard University)
09:30 University Case Studies 30'
DAY 5 - Friday, 10 June 2022
07:00 FOMO 1h0' Speaker: (Video) Shawn Hymel (Edge Impulse)
08:00 Case Study: Our Experience in Building an Industrial Internet of Things (IIoT) Solution Business 30' Speaker: Apinun Tunpan (SMART Sense Industrial Design Company Limited, Thailand)
08:30 Case Study: Design and Deployment of IoT Devices, from Edge to Data Centers 30' Speaker: Reginald Juan Magpantay Mercado, (GTek Research, Philippines)
09:00 Artificial Intelligence as a driver for the sustainable development Application to disaster risk reduction 1h0' Speaker: Soichiro Yasukawa (UNESCO)
10:00 Discussion: Specific TinyML Research Asian challenges 1h0'
* * *
TinyML is a subfield of Machine Learning focused on developing models that can be executed on small, real-time, low-power, and low-cost embedded devices. This allows for new scientific applications to be developed at an extremely low cost and at large scale.
The TinyML process starts with collecting data from IoT devices, then training the collected dataset to extract knowledge patterns; these patterns are then packaged into a TinyML model that considers the target microprocessor’s limited resources such as memory and processing power.
The resulting model is then deployed on embedded devices where it is used to evaluate new sensor data in real-time. Typically, power requirements are in the mW range and below which enables a variety of use-cases targeting battery operated devices. TinyML represents a collaborative eﬀort between the embedded power systems and Machine Learning communities, which traditionally have operated independently.
• ML general concepts
• Introduction to TinyML
• Getting started with the TinyML training kit
• Examples of TinyML applications
• Scientific Applications of ML
* * *
Priority given to Asian candidates that are part of the ICTP TinyML Academic Network.
Interested candidates from other regions:
- Potential candidates from Latin America are invited to consider applying for a dedicated edition of this online activity: Latin American Regional Workshop on SciTinyML: Scientific Use of Machine Learning on Low-Power Devices (see: smr.3721)
- The African Regional Workshop on SciTinyML: Scientific Use of Machine Learning on Low-Power Devices has been held from 25 to 29 April 2022 (see: smr.3708)