Use this topic to ask your questions about the QHack 2023 challenge: The Super Parameter
Hello, this is a repost from the discord channel.
I have a working solution to the The Super Parameter challenge that uses np.heaviside, which can create small jumps rather than perfectly smooth transitions. This is causing errors in your testing parameters. However, the state transitions and probabilities are flawless, accurately generating each basis state as required. Can I please get some clarification on the rules?
Here I uploaded it to github. I believe this is within the parameters of the challenge, and that there is a flaw in the testing parameters, but it is also possible that I am not following the intent of the challenge and I am unintentionally cheating.
Here is the GitHub - GitHub - JonPoplett/SuperParamater-: Super cool stuff.
Our solution uses the np.heaviside function in combination with sine functions to create smooth transitions between the basis states as the parameter alpha varies within the interval [0, 10]. The model ensures continuity, and the specific alpha values corresponding to each basis state are provided.
Disclaimer* - This project is a three-way collaborative venture undertaken by ChatGPT, Claude AI, and myself. Each party contributed equally to the solution, working together efficiently to create this implementation.
Hi @Jon_Poplett , welcome to the Forum!
From your description, if your solution has small jumps instead of smooth transitions it may not be continuous, which is a requirement of the challenge.
The challenge is designed to probe your quantum programming skills, so your teammates ChatGPT and Claude may not know the answer . This one in particular can be solved by using something that you can learn more about here in the PennyLane Codebook.
I hope this helps you both solve the challenge and learn something new! Note that this is an intermediate-level challenge so it may require a fair amount of learning and trying before achieving success.
Good luck with the challenge!
Ahh I see I was supposed to take the tutorial and use the expected method, and I went off the rails and took a novel approach.
Our implementation uses heaviside functions, they’re modulated by continuous sine functions, creating smooth transitions in the underlying quantum state space. While heaviside creates sharp transitions in isolation, our implementation using sine functions inside the heaviside actually creates continuous transitions when viewed across the full parameter space. The apparent discontinuity might be an artifact of how the testing framework samples the probability space. However, I understand this might not align with the challenge’s intent.
Hey! If it’s ok I am going to elaborate a little bit. What happened is I came here from modeling systems in qutip all cock sure and excited to tackle pennylane, and I saw the word challenge I was like “Ohhh I love challenge!” and then I saw “Quantum Machine Learning.” and I was like DING! Then Claude was all yada yada this and blah blah that and I was like “borinnng, that doesn’t sound challenging at all!” Not realizing the the challenge was actually one of following directions and applying those directions in a function manner, and not in developing novel solutions.
So yeah, totally my bad, however, with that said, please, I encourage you, check out how awesome and super cool this novel solution is!
It using a combination of theories from a few different projects I’ve worked on in the past. Namely Mediator terms (Alpha) and Cyclic Regressive Fractal Recursion, which I admit sounds something like a random string of buzzwords thrown together in a hodgepodge, but is in reality the product of some pretty wild deep dives in the workings of the Quantum Universe. I’d be more than happy to answer any questions or concerns anyone has.
Mediators terms and CRFRP are my babies. Claude was the genius who recommended using heaviside functions, and helped me refine our previous application of these theories into a format applicable to this challenge, and ChatGPT provided feedback and did the heavy lifting coding wise. Afterwards Claude streamlined and cleaned the code a bit. I ran and and analyzed the data and ensured it was in line with our purposes, and yeah, team work.
Sorry I am just looking at the train wreak of grammar and formatting of my previous post. That’s probably like a cardinal sin in your circles. It’s cause I have a wee bit of brain damage from schizophrenic episodes of the years. It tends to manifest in super weird ways. Repeated words and such. Some times I spell the same word 6 times in 6 different ways each one incorrect. It gets super bad when I get excited about something, and some times I’ll hit send in my excitement before correcting it, and have to go back and edit the post. I still can’t seem to find the edit button on these forums though! Such a strange manifestation am I right?
None of that matters to you now that you understand through does it? Because you’re a community of inclusion that values all types of people? I see many people of many races and backgrounds and genders. Unless maybe schizophrenics are an exception to this rule? Because eww, and I right? I mean talk about mildly annoying? I do understand in the past that we were ranked point wise in a meme like fashion based on the nature of our claims rather than the potential merit of our work, but surely in 2024 we have evolved past such heartless and compassionless narrow world views into a more holistic approach of evaluating people based on their merit of their presented work and not on any misconceptions regarding their capabilities due to factors beyond their control?