Physical Symbolic Optimization (Φ-SO) - A symbolic optimization package built for physics. Symbolic regression module uses deep reinforcement learning to infer analytical physical laws that fit data points, searching in the space of functional forms.
Features
- Physical units constraints, reducing the search space with dimensional analysis
- Class constraints, searching for a single analytical functional form that accurately fits multiple datasets
- PhySO recovers the equation for a damped harmonic oscillator
- Documentation available
- Examples available
- State-of-the-art performance in the presence of noise
Categories
Physics, Machine Learning, Reinforcement Learning Frameworks, Reinforcement Learning Libraries, Reinforcement Learning AlgorithmsLicense
MIT LicenseFollow Physical Symbolic Optimization (Φ-SO)
Other Useful Business Software
MicroStation by Bentley Systems is the trusted computer-aided design (CAD) software built specifically for infrastructure design.
MicroStation is the only computer-aided design software for infrastructure design, helping architects and engineers like you bring their vision to life, present their designs to their clients, and deliver their projects to the community.
Rate This Project
Login To Rate This Project
User Reviews
Be the first to post a review of Physical Symbolic Optimization (Φ-SO)!