I love your enthusiasm @pavankumarinfo.
Perhaps after Course 3 launches, I can work with @marcozennaro to coordinate a kick off meeting to see how like minded people such as ourselves can get together to get this project afloat.
If we as a community believe in it and doing it, we can do it.
Hi everybody!
I contacted a professor of Nutrition at the University of Sao Paulo. He is very interested in going ahead with everything, from sharing knowledge to implementing some TinyML projects (I have some in mind). However, due to the pandemic crisis, he is a bit busy with rearranging classes & course schedules. He will be available in a monthâs time.
This guy created and is the head of the Population Analysis Laboratory, he has a very good database with measurements of the Brazilian population, which could be used to model data.
Projects suggested by the professor:
Anemy studies
Children and teenager population obesity (a real problem in our world)
Diabetis checking and control
Blood pressure and oxygen level
Other biochemical data
I understand that these projects could really help developing countries. By the way, the professor also told me that the Nutrition Faculty has very close relationships with the corresponding course in Harvard.
Thatâs awesome @hochberg. Thanks for following up on this.
I know that feeling about being tied up with courses; we are having to do the same at Harvard.
I tried looking up a Website URL for the lab. Is there one you can share with us as I couldnâtâ find it. @JoydeepG mentioned that we could create a Kaggle competition around some ideas (not necessarily these but in general). Just connecting the dots here to help get some energy going.
Hi, @vjreddi .
The link for the lab page is http://www.fsp.usp.br/lanpop/
Kaggle competition: excellent idea! I will try to check with Professor Conde about the data he has available, structure, etc.
Hello @hochberg,
Allow me piggyback off your suggestions. I would like to append âsleep monitoringâ to your list. There are many different opportunities here from diagnosis to treatment. The key characteristics here are that data is collected continuously over a few hours.
Kind regards.
Thanks a lot, @baqwas !
Yes, it could be useful, too.
My main interest is in public health & inexpensive devices that could be bought by development countries and that could help in early disease diagnosis, reducing the costs for the public health system and, of course, providing the population with a better life quality.
early disease diagnosis
Precisely! The benefits especially when the locations are remote can provide the most significant benefits under budgetary constraints which are inevitable.
Kind regards.
We are organizing a Special Track on the use of ML on Embedded Devices: applications for Social Good at the ACM International Conference on Information Technology for Social Good (GoodIT 2021).
This track will focus on applications for social good that make use of the fast-growing branch of ultra-low power machine learning technologies. These devices have a huge potential, especially in developing countries, as they enable the creation of a healthier and more sustainable environment for all. They are low power by design (typically in the single milliwatt and below power range) and do not require a stable connection to the Internet.
The topic is timely as the recent explosion of ML-enabled embedded devices will have an impact on applications for social good: increasingly common examples include applications of image recognition with ML on inexpensive devices in wide varieties of fields like agriculture, public health, anti-poaching surveillance etc. Moreover, this is a very new topic and there are few conferences that focus on the applications of ML in embedded devices that have an impact on society.
Topics may include, but are not limited to:
Important Dates
Submission deadline: May 1, 2021
Notification of acceptance: June 22, 2021
Camera Ready: July 10, 2021
Publication
Papers must be in English and should be up to 4 pages in length. Previously published work may not be submitted, nor may the work be concurrently submitted to any other conference or journal. Such papers will be rejected without review.
Submission
Papers should be submitted via the HotCRP submission website.
More info here.
Cheers,
Marco
Hi @hochberg!
I also live is Sao Paulo. Iâm a doctor (Obs-Gyn) and I also come from Escola Politecnica (Mechatronics Engineering).
I believe in developing solutions to improve health and help people live better lives.
If you think that creating a TinyML cell in Sao Paulo worth, lets go.
Spreading knowledge as @vjreddi propose and /or developing projects that help improve health here would be of great value.
Regards,
Hi, @mateuslfreitas !
Thanks for your reply, I am a bit late with the course, I am retaking it today. I have been sick with covid, I am in the 3rd week of symptoms, improving now. I will spare you all the details.
My wife is both a lawyer and a nutritionist (I am feeling a little bit of an underdog here⊠
: and I have some good contacts at nutrition/USP.
I have some experience in KM (Knowledge Management) and, yes, I really believe in spreading knowldge. Count me in for the Sao Paulo cell! And letâs keep in touch!
Regards,
Carlos
This is possible, I am in the university so I can help out here,
Marco,
A pleasure to meet you. I met you I think in 2019 at Republic of Benin in a one week workshop on IoT, where you trained us on IoT, at the Institute of Mathematics and Applications, University dâAbomey, Calavi. A nice session we had with you!
Roselyn
Welcome @Roselyn! Great to see you are interested in this thread. @brian_plancher is working on making our materials publicly accessible so that other teachers can build on what we have to share. Stay tuned for that!
The slides/readings/colabs from the edX course can all be found here: GitHub - tinyMLx/courseware
My goal is to also get more materials posted from anyone who teaches other TinyML courses so that we can have a repository for all teachers to mine from as they develop new courses targeting new audiences! So if you create a course please consider submitting a PR with some of your materials!
@marcozennaro it might be time to kick off a meeting with like minded folks here to start a tinyML4D gathering of minds 
What do you think?
Hi @vjreddi and @brian_plancher. I am a Brazilian (living in Chile) and a half away from Course #3. I am very focused on your idea of TinyML4D. This week, I gave a speech at an IoT conference in Brazil talking about TinyML (here a 3â glimpse of the presentation: https://youtu.be/kAjIDlswjpI).
I am an Electronics Engineer with a Masterâs in Data Science. TinyML is a perfect match in both worlds (Electronics and DS). I used last yearâs pandemic time to write and publish some projects, mainly in EdgeML (with RPi and TFLite). Here, some of them:
End of last year, I teach a special class about EdgeIA for engineerâ students (https://youtu.be/aLiu-xt4cDs) at my university (UNIFEI - Federal University of ItajubĂĄ), and this year I started supervising an Engineering End of Course project, where the idea is to use TinyML. The university invited me to present a TinyML course proposal that I did, based on what we have seen on the EDX courses, the Harvard course, and de Coursera one. (https://www.dropbox.com/s/bd6c3psa8buc0f6/UNIFEI%20-%20Proposta%20curso%20TinyML%20VersĂŁo%201Âș%20Semestre_2021.pdf?dl=0).
If everything works fine, I will start the course next month. I really appreciate all your efforts to help developing countries. I believe that TinyML can be of great help!
Thanks a lot!
Hi Brian, certainly I will include all material that I will generate for the TinyML course of UNIFEI in Brazil (please see my last post regarding it). My idea is to teach the classes in Portuguese, but keeping the materials in english, so others can use it, same I am doing with HarvardX. Thanks a lot.
Amazing! This all sounds fantastic and your talk looked great (and from the google auto-translate also read great)! I think it will really help everyone to get materials out there in many different languages! Let us know what we can do to help â I got all of the slides, readings, and colabs up on the github now so hopefully that is a helpful start!
@mjrovai I love the cascade architecture!!! Thatâs awesome. Thanks for sharing the links.
Sounds wonderful.
One of our interests in supporting tinyML4D is providing resources to bootstrap. As @brian_plancher mentioned we opened up the slides. Videos we canât 
But hopefully as Brian said the slides and collabs are helpful to jumpstart you.
Then of course is the issue of the kits. Let me know if you folks need to have some sponsored kits. I can pull some strings to make that happen (no promises though â 
) but it would be nice to have industry sponsor kits as right now things are still costly ⊠we did get feedback that $50.00 for a kit is still way off the charts.
Ideally, we should shoot for $5.00 tinyML devices. I pulled the number out of thin air. Do you have any ideas on what would be a reasonable cost? â this question is to the broader audience as well.