Malicious Login Detection background cover

Malicious Login Detection

Detect malicious login attempts in the BETH dataset

novaceneai

Hosted by

novaceneai

Overview
Docs
Code
Rules
Discussion
Leaderboard
PeterYS
PeterYS

2

Posted

Runtime

More of a comment rather than a question:
I've tried with both the quantum and classical implementation. It seems that with the classical approach the runtime is much faster (not so bad accuracy), with existing classification model. So, is it still worth doing the quantum ML with this problem? Or shall I seek other quantum network with less time cost?
Order by:
Jannes Stubbemann

2

Posted by stubbi •

Thanks for the comment! What do you mean by
It seems that with the classical approach the runtime is much faster (not so bad accuracy), with existing classification model.
Is it because of simulation overhead or the algorithm as such? As from Aqora perspective, any solution are supposed to harness quantum computing.
Have a great day!
PeterYS

2

Posted by PeterYS •

What I mean is that the simulation overhead for the classical method, SGDOneClassSVM—which is the one I am using and has an accuracy of around 0.85—is much smaller than that of the quantum counterpart.
My quantum computational solution is simply to add angle embedding for the classical data, but the simulation takes hours to run...
I suppose I need to find a way to work around this.
Jannes Stubbemann

1

Posted by stubbi •

Gotcha! Is there anything we could provide you with to come to a solution for this kind of problem?

Want to join this discussion?

Join our community today and start discussing with our members by participating in exciting events, competitions, and challenges. Sign up now to engage with quantum experts!