Looking for physics based simulation for data generation

Click For Summary
SUMMARY

This discussion focuses on the search for physics-based simulation tools for data generation relevant to physics-based neural networks. The user seeks online tools that can produce data closely resembling real experimental results, specifically for proof of concept (POC) purposes. Suggestions include using Python to generate test data by applying theoretical models and adding noise with the math.rand function. Participants also recommend exploring datasets such as weather data, stock prices, and lottery data for potential analysis.

PREREQUISITES
  • Understanding of physics-based neural networks
  • Familiarity with loss functions in machine learning
  • Basic knowledge of Python programming
  • Experience with data generation techniques
NEXT STEPS
  • Research online physics simulation tools for data generation
  • Learn how to implement physics-based loss functions in neural networks
  • Explore Python libraries for data generation, such as NumPy
  • Investigate the use of real-world datasets like weather or stock price data for analysis
USEFUL FOR

This discussion is beneficial for machine learning practitioners, data scientists, and researchers focused on integrating physics into neural network training and those seeking efficient data generation methods for experimental simulations.

up2front
Messages
2
Reaction score
1
Good day,

I am working on physics based neural networks and I am investigating the role of integrating physics based loss function on the network's training.
So I am looking for physics based simulation for data generation (online tools) to generate data from experiments.

I know I may generate data using the explicit laws on excel but I am looking for relatively close to real experiments simulation. Please note that I am not looking for complicated simulation I just need data for POC.

So if anyone aware of such simulation software or tool please let me know.
 
Technology news on Phys.org
Welcome to PhysicsForums. :smile:

What kind of datasets? How many channels? How many bits of precision for each channel?
 
I would write a program in python to generate your test data.

use the theory to generate expected data and then use the math.rand function to add noise to the data.

we used this approach to create FFT test data that the FFT function could analyze and generate the frequency spectrum.
 
  • Like
Likes   Reactions: up2front
up2front said:
Good day,

I am working on physics based neural networks and I am investigating the role of integrating physics based loss function on the network's training.
So I am looking for physics based simulation for data generation (online tools) to generate data from experiments.

I know I may generate data using the explicit laws on excel but I am looking for relatively close to real experiments simulation. Please note that I am not looking for complicated simulation I just need data for POC.

So if anyone aware of such simulation software or tool please let me know.
Not exactly what you are looking for (means I don't know how complicate an implementation is), but it fits to your research: https://www.physicsforums.com/threa...ving-electronic-schroedinger-equation.997917/
 
  • Like
Likes   Reactions: up2front
berkeman said:
Welcome to PhysicsForums. :smile:

What kind of datasets? How many channels? How many bits of precision for each channel?
Well, any data set with enough variables that can be modeled into equation or loss function
 
up2front said:
Well, any data set with enough variables that can be modeled into equation or loss function
Sorry, please answer my questions explicitly and technically so that we can try to help you. Thank you.
 
You could always look for lottery data or weather data. Weather in particular could be interesting to analyze.

Or even stock/bond price trading data.
 

Similar threads

  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 3 ·
Replies
3
Views
1K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
Replies
48
Views
5K
Replies
11
Views
2K
Replies
1
Views
3K
  • · Replies 4 ·
Replies
4
Views
7K
  • · Replies 6 ·
Replies
6
Views
2K
Replies
4
Views
2K