Learning universal approximations for partial differential equations with Physics-Informed Broad Learning System
arXiv:2606.19754v1 Announce Type: new Abstract: Partial differential equations (PDEs) play a central role in modeling complex physical, biological, and engineering systems. While traditional numerical solvers are robust, they often incur prohibitive computational costs due to mesh dependencies, whereas recent Physics-Informed Neural Networks (PINNs) offer a mesh-free alternative but frequently suffer from slow convergence and optimization instability. To bridge this gap, this article proposes the Physics-Informed Broad Learning System (PIBLS), a novel backpropagation-free framework that reform
The continuous push for more efficient and robust AI-driven scientific computation, particularly for complex systems like PDEs, drives the development of new architectures like PIBLS.
This development addresses key limitations of current Physics-Informed Neural Networks (PINNs), potentially accelerating scientific discovery and engineering R&D with more stable and faster simulations.
The proposed Physics-Informed Broad Learning System (PIBLS) offers a backpropagation-free approach to solving PDEs, overcoming convergence and stability issues common in PINNs, thereby making AI-driven simulations more practical.
- · Scientific research
- · Engineering sectors
- · AI compute providers
- · Drug discovery
- · Traditional numerical solver developers (if not adapting)
- · Companies reliant on slow simulation cycles
More accurate and faster simulation capabilities for complex physical systems.
Reduced R&D cycles and costs across industries dependent on simulation, leading to faster innovation in materials science, aerospace, and climate modeling.
Democratization of advanced simulation, enabling smaller research groups and companies to tackle problems previously reserved for institutions with massive computational resources or specialized expertise.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG