Dimensions overall score 3.0

GitHub Code Pulse

No public code linked for this paper yet.

• The Physics-Informed Neural Networks and Sequence Encoder (PINN-SE) model can be applied to the realistic case of heating and early cooling of the thermo-stamping process.method 95%
• The PINN-SE model can extend its inputs to multimodal data, including sequences of temporal 2D images.method 95%
• Combining multiple encoders with the previously proposed PINN-SE method is feasible.result 90%
• Training the PINN-SE model on synthetic data generated from experimental data can improve generalization to unseen real experimental data.result 95%
• A Convolutional Neural Network (CNN) is used to encode sequences of temporal 2D images into feature vectors.method 90%
• A Deep Set is employed to encode the temporal dimension of latent vectors generated from image sequences.method 90%
• Fourier feature layers are utilized to encode spatial dimensions (x and y) for faster model convergence.method 90%

Competitive landscape

Competitor map is still being generated for this paper. Enable generation or check back soon.

Startup potential card