Welcome 10 Selected Teams for the LETU Contest

Hello, Everyone.
Frist of all, I welcome all of you for your registration to enter the contest.
This is a blog to encourage each other as we share the ideas and tips for successful projects. Please, read the following judging criteria as well. Then, I pray and cheer for you and your fun and successful project!

Judging Criteria

Judges

  1. Qeexo (2 members)
  2. Arduino (1 member)
  3. LETU (2 members)

Judges will score entries for entered category based on the following judging criteria:

Originality & Creativity (10 points)

  • Problem statement and motivation: which problem do you want to solve for, why?
  • What are the benefits of using embedded ML to solve this problem?
  • Novelty, the contest idea should not be introduced in Qeexo website or its existing show cases.

Functionality / Technical Qualities (10 points)

  • How well does it work?
  • How would you go about improving the solution?
  • How does using Qeexo AutoML make implementing the idea easier?

Demonstration / Presentation (30 points)

  • Video Demo + Presentation (PPT/slide) – explain #1 and 2 in presentation
  • Two to five forum entries on Qeexo website, documenting your progress and thought process, ideally on a regular basis
  • Real Demo with final round teams (tentative)
3 Likes

Good luck to everyone!

Hi Teams & Professor @hookim,

This is Tina, Product Marketing Manager at Qeexo – you may have seen my brief presentation about Qeexo and Qeexo AutoML a few weeks ago in Professor Kim’s class.

Today I have a couple of announcements and updates to make regarding the competition.

  1. Our most recent update to Qeexo AutoML now allows ML model exports for the Arduino Nano 33 BLE Sense hardware. This means that you can download and embed the ML model(s) in your application(s) as a statically-linked classification library.

@GilbertTsang, our Product Manager, will share additional details about how to use this functionality shortly. Please note that everyone will need to update Qeexo AutoML (by reinstalling) to continue working with the tool, whether or not you want to take advantage of this new feature.

  1. Additional bonus points (up to 20 points total) will be awarded to teams who successfully implement functionalities based on classification results, using the library export feature described above.

  2. If you encounter any technical issues or have questions for Qeexo about using Qeexo AutoML, please post to this forum, under the “LETU Qeexo AutoML Competition” category, and @tina_shyuan so I can follow up.

  3. A reminder to everyone to introduce your teams and project ideas here - these entries will count for part of the final grade, as outlined in the judging criteria above.

See you around!
Tina

1 Like

Hello Teams & Professor @hookim:

Hope this post finds you safe and healthy.

I am excited to share that the step-by-step guide on using Qeexo AutoML to build custom applications for Arduino Nano 33 BLE Sense is live at:

https://github.com/qeexo/Qeexo_AutoML_Arduino_Nano33BLE_pub/blob/automl-arduino-nano-33ble-sense/README.md

Please let me (@GilbertTsang) or @tina_shyuan know if you have any questions on the guide, or building applications with Qeexo AutoML. We can’t wait to see what your ideas are about!

Best regards,

Gilbert

Hello!

Our team is coming up a potential issue in the Qeexo AutoML data collection tab.
When we attempt to build a new environment to begin training, we enter in our desired sensors and sensitivities into the “configure sensors” tab. From there, the website tells us that it is “flashing the application” and that it will take a few minutes. Every time we have tried this it sits and flashes the program for the next six hours or so before telling us that it has timed out.
Do you have any suggestions for where to go from here?

Thank you!
-SurfaceSense Team (Micah, Levi, Corbin)

I am having a similar issue to the Surface sense team. the application will not flash to my Arduino. I uninstalled and tried to reinstall, but I keep receiving an error. I will add an image to show the error.

Welcome, SurfaceSense!

Sorry, I’ve just seen this post (please @tina_shyuan and/or @GilbertTsang, otherwise it does not notify us), and passed it to Gilbert, our Product Manager, for help.

It really should just take 1-2 minutes. If it goes longer than that, it probably has failed; please do not wait for 6 hours… My suggestion is to press-and-hold the (only) grey button to reset the device and try again. Sometimes the factory-settings do not work out-of-the-box with Qeexo AutoML, but can be fixed easily. (The Installation Guide should have included this step.)

If you are still having trouble, Gilbert can look into more details with you.

Welcome Jacob!
Please see my response to SurfaceSense above. :slight_smile:

However, looking at your screenshot, it looks like your issue is with the installer of Qeexo AutoML, not the flashing of the Arduino. I have not seen this particular issue before and would have to refer to our engineers and Gilbert.

Hello @jacoblandreth32, thank you for your post and screenshot. Do you mind sending me the installation log file? This can be found at C:\Program Files (x86)\Qeexo AutoML\install.log. My email address is gilbert.tsang@qeexo.com.

Looking forward to receiving your log file!

Gilbert Tsang

Hello @SurfaceSense Team, thank you for your post. Assuming you are using Windows 10, do you mind sending me these 2 files for troubeshooting?

  • C:\Program Files (x86)\Qeexo AutoML\install.log.
  • C:\Users\“username”\.local\share\qeexo\devicemanager.log

My email address is gilbert.tsang@qeexo.com.

Looking forward to receiving your log files!

Gilbert Tsang

If you are using macOS, feel free to check out this forum post on generating a diagnostic report of Qeexo AutoML.