What we're about

• What is the purpose of the group?
To spread the word and educate the industry on "tinyML" (broadly defined as machine learning architectures, devices, techniques, tools and approaches capable of performing on-device analytics for a variety of sensing modalities--vision, audio, motion, environmental, human health monitoring etc.) at “mW” or below power range targeting predominately battery operated devices. The tinyML meetup group is an informal monthly gathering of researchers and practitioners working on various aspects of machine learning technologies (hardware-algorithms/networks- software-application) at the extreme low-power regime to share latest developments in this fast growing field and promote collaborations throughout the ecosystem. The format will be presentations with Q&A followed by networking.

• Who should join?
Experts in machine learning technologies at the edge, especially in the low power battery operated regime. This includes hardware architects, software engineers, systems engineers, ASIC designers, algorithms and application developers, low power sensor providers and end users. “Newbees”, i.e. people interested in joining this field and getting up to speed by listening start-of-the-art presentations and interacting with established players are very welcome to join too, both from the industry and the academia.

• What will you do at your events?
Communicate to the attendees the “latest and greatest” in tinyML by watching a presentation from a tinyML expert from the industry or the academia and interfacing with the member of the tinyML Community.

Upcoming events (2)

TinyML Challenge 2022: Smart weather station

Network event

Link visible for attendees

Announcing Challenge 2022 on June 22nd, 2022

IMPORTANT: Please register here
https://us02web.zoom.us/webinar/register/2016554048715/WN_uyEeGRJzQ9ucMtfPsf28Rg

Once registered, you will receive a link and dial in information to teleconference by email, that you can also add to your calendar.
8:00 AM - 9:00 AM Pacific Daylight Time (PDT)
Thomas Basikolo, Programme Officer, ITU
Marco Zennaro, Research Scientist, Abdus Salam International Centre for Theoretical Physics in Trieste, Italy
Alessandro Grande, Director of Technology, Edge Impulse
"TinyML Challenge 2022: Smart weather station"

Developing Countries is the area of the globe where land-based, in situ monitoring of weather and climate is at its scarcest, but at the same time has arguably the most potential to benefit society.
Rainfall and temperature can have high spatial variability due to the strong feedback that can exist between the land and atmosphere. Temperature can change rapidly in space due to land-cover heterogeneity and changing altitude over complex mountainous terrain. This means that a weather station tens of kilometers away may measure conditions that have little relevance to your location, making it hard to make informed local decisions.
The goal of this challenge is to create a low-cost, low-power, reliable, accurate, easy to install and maintain weather station, with no mechanical moving parts for measuring all weather conditions with a focus on rain and wind, based on TinyML, that can be deployed locally.
This talk will introduce the 2022 tinyML Challenge and how you can participate in this Challenge.

Thomas Basikolo works with the ITU coordinating and managing the AI for Good’s ML5G activities and as an advisor of the ITU-T Focus Group on Autonomous Networks. He received a PhD in Electrical and Computer Engineering from Yokohama National University, Japan. During his studies, he was awarded the Japanese Government (Monbukagakushō) scholarship. He was also a recipient of grants for Non-Japanese Researchers from the NEC C&C Foundation, and a visiting researcher at the NEC Data Science Research Laboratories. Prior to joining ITU, he worked as a Research Engineer in the Engineering Department of Microwave Factory Co., Ltd, Tokyo, Japan. He is recipient of multiple Best Paper Awards, the IEEE AP–S Japan Student Award and the Young Engineer of the year award by IEEE AP–S Japan in 2018. He has co-authored peer-reviewed journal and conference papers, predominantly in the areas wireless communications and antenna engineering. He serves as a Reviewer of IEEE and IEICE Journals.

Marco Zennaro is a Research Scientist at the Abdus Salam International Centre for Theoretical Physics in Trieste, Italy, a Category I UNESCO Institute, where he coordinates the Science, Technology and Innovation Unit. He received his PhD from the KTH-Royal Institute of Technology, Stockholm, and his MSc degree in Electronic Engineering from the University of Trieste in Italy. His research interest is in ICT4D, the use of ICT for Development, and in particular he investigates the use of Internet of Things for Development (IoT4D). He acts as TinyML4D Chair and TinyML Academic Network Co-Chair, in the framework of the TinyMLEdu initiative. Over the years he has organized more than 30 training activities on IoT in Developing Countries. Marco is a Visiting Professor at Kobe Institute of Computing (KIC) in Kobe, Japan.

Alessandro is a physicist, an engineer, a community builder and a communicator with a visceral passion for connecting and empowering humans to build a more sustainable world through the aid of technology. Alessandro is the Director of Technology at Edge Impulse and co-organizes the tinyML Meetups in UK and Italy. Prior to Edge Impulse, Alessandro worked at Arm as a developer evangelist and ecosystem manager with a focus on IoT and TinyML. While at Arm Alessandro launched a weekly live stream – Innovation Coffee with his colleague Robert Wolff. Alessandro holds a master’s degree in nuclear physics from the University of Rome “La Sapienza”.

We encourage you to register earlier since on-line broadcast capacity may be limited.

Note: tinyML Challenge slides and videos will be available on the tinyML website and tinyML YouTube Channel afterwards, for those who missed the live session. Please take a moment and subscribe to the YouTube channel today: https://www.youtube.com/tinyML?sub_confirmation=1

#@#

tinyML Talks by Koteswararao Chilakala from TU Delft

Network event

Link visible for attendees

Announcing tinyML Talks on July 9th, 2022

IMPORTANT: Please register here
https://us02web.zoom.us/webinar/register/3616559312186/WN_jVPbo9yAQ8m65e4hLHRcvw

Once registered, you will receive a link and dial in information to teleconference by email, that you can also add to your calendar.

9:30 AM - 10:30 AM Pacific Daylight Time (PDT)
Koteswararao Chilakala, Masters graduate in Embedded Systems, TU Delft
"Neural network framework using emerging technologies for screening Diabetic Retinopathy"

Diabetic Retinopathy (DR) is one of the leading causes of permanent vision loss. Its current prevalence is around 45 millions across the globe and is projected to 70 million by 2045. Most of the people with this disease condition belong to remote and low income settings. We can reduce this incidence, if quality medical care is accessible in remote areas. With the current advancements in imaging technologies, fundus examination can be carried out on a handheld device. We need to improve such devices to deliver high quality services through auto detecting DR based on convolutional neural networks(CNNs) in an offline setting. Addressing these challenges, we aim to develop an integrated solution which delivers high compute at ultra-low power consumption. Firstly, we have created 3 datasets of different sizes merging multiple public datasets to create a vigilant model training process. This is to make the CNN model robust to real-world noise. CNNs trained on smaller datasets have shown a 15% accuracy drop on evaluation datasets where as CNNs trained on large datasets showed consistent performance. Secondly, we have proposed a new binary labelling scheme using multi-class output to maximize the utility of its softmax probabilities. We have achieved 90.26% accuracy on evaluation dataset with the new scheme. These high performing models along with compression techniques are implemented on resistive random access memory (RRAM) based computational in memory (CIM) architecture. These implementations resulted in atleast 200x improvement in energy consumption for inferring on one image when compared to CPU, GPU and mTPU (google coral dev board). Similarly, latency improvements 28x,130x and 2000x compared GPU, CPU and mTPU are registered. Model quantization with 4-bit precision for weights have preserved the original accuracy and showed 4x improvements in energy consumption on RRAM implementation.

I am a recent masters graduate from TU Delft in Embedded Systems. I develop software and embedded platforms for autonomos and robotic systems. I have experience in developing screening and diagnostic devices for eye care for 4 years. I aspire to develop systems enabling embedded Machine Learning. I am here to share the work of my thesis where we developed a neural network framework using emerging technologies for screening diabetic retinopathy.

We encourage you to register earlier since on-line broadcast capacity may be limited.

Note: tinyML Talks slides and videos will be available on the tinyML website and tinyML YouTube Channel afterwards, for those who missed the live session. Please take a moment and subscribe to the YouTube channel today: https://www.youtube.com/tinyML?sub_confirmation=1

#@#

Past events (149)

tinyML Talks by Alexander Timofeev from Polyn.ai

This event has passed

Photos (238)