Code the Light Fantastic- Level 2
Share:
twitter social icon linkedin social icon facebook social icon
Finished
competition-bg

Code the Light Fantastic- Level 2

Simulate an adaptive matrix headlight

KTM AG
Machine Learning/AI
Enterprise
Total Prize 5,000
Join Datathon
Scroll To Top

Datathon Type

challenge_cup

Challenge with monetary prizes

No. of Users

20

No. of Submission

83

No. of Teams

5

Community chat

Join our Discord

Description

Description of the competition


Congratulations, Trailblazers! You've successfully navigated Level 1 of the KTM AG Inaugural Code Challenge, showcasing your progress in developing an algorithm for an adaptive high-beam lighting system. Now, let's dive into Level 2, where the challenges become more demanding, and the bar is set higher.

Level 2 Overview:
In Level 2, participants are confronted with a new video featuring more challenging conditions. Additionally, another set of Key Frames is provided, adding complexity to the evaluation process. The core task remains consistent: refine and elevate your algorithm to detect objects, map regions onto the pixel matrix, and dynamically control the high-beam lighting system.

Task Refinement:

  1. Challenging Conditions: The new video introduces more complex scenarios, pushing participants to enhance the adaptability of their algorithms.
  2. Tighter Tolerances: The tolerance on Key Frames becomes tighter, demanding even greater precision in replicating the desired illumination states.

Assessment Process:
Similar to Level 1, the algorithm undergoes both automatic and manual assessments. The automatic evaluation ensures accuracy in reproducing Key Frame outputs, while the manual on-board assessment verifies the compatibility with KTM AG's target hardware.


Level 1 Reference:
In Level 1, participants tackled a video filmed in "easy" conditions with accompanying Key Frames showcasing the ideal algorithmic outputs. The algorithm's accuracy in mimicking these Key Frames determined participants' success, and those meeting the minimum score progressed to Level 2.

Progression to Level 3:
To advance to Level 3, participants must surpass a higher minimum score threshold, demonstrating a refined algorithm capable of handling the intensified challenges posed in Level 2.

Your journey in shaping the future of two-wheeler technology with KTM AG continues in Level 2. Best of luck, and may your algorithms shine in the face of increased complexity! 🏍️💡

Description of the competition


Congratulations, Trailblazers! You've successfully navigated Level 1 of the KTM AG Inaugural Code Challenge, showcasing your progress in developing an algorithm for an adaptive high-beam lighting system. Now, let's dive into Level 2, where the challenges become more demanding, and the bar is set higher.

Level 2 Overview:
In Level 2, participants are confronted with a new video featuring more challenging conditions. Additionally, another set of Key Frames is provided, adding complexity to the evaluation process. The core task remains consistent: refine and elevate your algorithm to detect objects, map regions onto the pixel matrix, and dynamically control the high-beam lighting system.

Task Refinement:

  1. Challenging Conditions: The new video introduces more complex scenarios, pushing participants to enhance the adaptability of their algorithms.
  2. Tighter Tolerances: The tolerance on Key Frames becomes tighter, demanding even greater precision in replicating the desired illumination states.

Assessment Process:
Similar to Level 1, the algorithm undergoes both automatic and manual assessments. The automatic evaluation ensures accuracy in reproducing Key Frame outputs, while the manual on-board assessment verifies the compatibility with KTM AG's target hardware.


Level 1 Reference:
In Level 1, participants tackled a video filmed in "easy" conditions with accompanying Key Frames showcasing the ideal algorithmic outputs. The algorithm's accuracy in mimicking these Key Frames determined participants' success, and those meeting the minimum score progressed to Level 2.

Progression to Level 3:
To advance to Level 3, participants must surpass a higher minimum score threshold, demonstrating a refined algorithm capable of handling the intensified challenges posed in Level 2.

Your journey in shaping the future of two-wheeler technology with KTM AG continues in Level 2. Best of luck, and may your algorithms shine in the face of increased complexity! 🏍️💡

Organisers & Sponsors

Timeline

22

Challenge starts

November at 19:00 UTC

Add to Calendar

31

Challenge ends (Public leaderboard)

December at 21:30 UTC

Add to Calendar

31

Challenge ends (Private leaderboard)

December at 21:30 UTC

Add to Calendar

Description of Timeline

🚀 Welcome to the Level 2 of Code the Light Fantastic challenge ! 🚀

📅 Event Duration: November 21, 2023 - December 31, 2023

This round is exclusively for approved participants and invited applicants. Please consult the Level 1 material if you didn't previously participated: https://www.datasource.ai:443/en/home/data-science-competitions-for-startups/code-the-light-fantastic-cdd4094f-59de-46a4-b0d5-eacd8c752219/description

Please note that the timeline for the Datathon may be adjusted without prior notice. To stay updated, we recommend regularly reviewing the platform and enabling notifications. This ensures that you don't miss any important updates or changes to the schedule.

Evaluation

Performance & Evaluation


Task description for the 2nd Phase:
Unlike Level 1, entry to level 2 is not open to all participants. Only participants which meet the following criteria may be admitted to level 2:
o Successfully won a prize in Level 1 OR
o Provided a submission above the 90% quality threshold, but no more
prizes were available.
o Provided a submission just below the 90% quality threshold, however
the submission was deemed interesting enough to grant admission.
o Have been privately invited to take part due to their track record in
similar competitions
Prior to receiving the data resources for Level 2, participants will be asked to sign an NDA. The evaluation procedure for Level 2 shall be as follows:
o Participants must ensure their submission respects all the requirements listed in Section 6.1 of the Terms and Services.
o An automated assessment will be made of the CSV file to understand how closely this matches the key frames that have been selected from the video. This result will be made public on the leaderboard.
o Entries will automatically be ordered in terms of output accuracy. Entries which score more than 95% will be assessed without waiting time. This involves downloading of the software onto the target hardware and verifying the output.
o The computing speed of the system will be measured on the target hardware. At Level 2 only software which has a maximum per frame calculation of 100ms or less is eligible to win a prize.
o The Weights and Biases used for Neural networks shall be limited to INT16 at Level 2.
o The quality of the submission will also be verified. Where the quality and computing speed are acceptable, a Level 2 prize shall be awarded to the participant. These prizes shall be allocated on a first come basis.
o In the case that not all Level 2 prizes have been awarded by the Level 2 deadline, submissions between 90 and 94.9% will be examined in order of score to assess if they are prizeworthy.

To be eligible to be consider for a prize, as solely determined by The Organizer, 
the entry MUST: 
  • Be aware of the hardware limitations upon which the code must be able to run, namely: 
  • System on Module: AGX Xavier [32GB]: 512 Core GPU, 64 Tensor Cores, 32 TOPs
  • Operating system: Jetson Linux 35.3.1 (Ubuntu 20.04) 
  • Jetpack version: 5.1.1
  • Graphics: AI Accelerator NVDIA Jetson SOM, Single GPU support with 
    NVIDIA Volta Architecture 
  • Be aware of the software limitations: 
  • Programming Language : Python3.8.10 [GCC 9.4.0 – Linux ]. 
  • Neural networks (if applicable) must be utilized via the ONNX runtime 
Supply all the submissions requested: 
  • The CSV File with the Key frame output 
  • Code submission in a *.py file which adheres to the provided code template 
  •  A text file which explains the function of your code, and provides a list of all libraries/packages used for it’s execution 
     
    Respect the performance limitations with regards to per-frame processing time 
    and Weights and Biases, which differ for Level 2 and 3
  • be in English; 
  • be the Individual’s (/Team’s) own original work; 
  • not have been submitted previously in any promotion of any kind; 
  • not contain material or content that: is inappropriate, indecent, obscene, offensive, sexually explicit, pornographic, hateful, tortious, defamatory, or slanderous or libellous; or promotes bigotry, racism, hatred or harm against any group or individual or promotes discrimination based on race, gender, ethnicity, religion, nationality, disability, sexual orientation, or age; or promotes alcohol, illegal drugs, or tobacco; or violates or infringes another’s rights, including but not limited to rights of privacy, publicity, or their intellectual property rights; or is inconsistent with the message, brand, or image of The Organizer, is unlawful; or is in violation of or contrary to the laws or regulations of any jurisdiction in which the Entry is created; 
  • not generate the Entry by any means which violate these Rules, The Organizer’s Terms of Service or The Organizer’s Privacy Policy; 
  • not engage in false, fraudulent, or deceptive acts at any phase during participation in The Challenge; and 
  • not tamper or abuse any aspect of The Challenge.
Join our private community in Discord

Keep up to date by participating in our global community of data scientists and AI enthusiasts. We discuss the latest developments in data science competitions, new techniques for solving complex challenges, AI and machine learning models, and much more!

Join our Discord