Inspiration

To predict/prevent errors and make gadgets more reliable for all humans

What it does

Oraclum is an optimized device tester. It saves the customer time and money.

How we built it

We've used simulator provided and improve upon on it. Then created oraclum as a component tester with basic AI techniques and statistical approach. Oraclum dynamically learns optimal action it should perform.

Challenges we ran into

  • Lack of real world data
  • Simulating a test behaviour close to real world applications
  • Trying to apply novel AI techniques to a statistical problem

Accomplishments that we're proud of

  • Improvement of the throughput on %94 of the cases.
  • Throughput is doubled on average on said cases. (Total time dropped by %51 percent on average)

What we learned

We've learned the importance of creating a good simulator with realistic approach.

What's next for Oraclum

  • With real world data and prior knowledge on component models, we wish to implement novel AI techniques
  • Increase the complexity of the simulator for customer benefit.

Built With

Share this project:

Updates