![Surrogate Model - Arize Docs](https://1591756861-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F-MAlgpMyBRcl2qFZRQ67%2Fuploads%2F7ukGOrpB3s6xfQAIog51%2Fimage.png?alt=media&token=3abc6b10-9b20-4fd6-a9ed-bfa449803a4d)
Complex brain simulations with Spiking Neural Networks (SNNs) demand vast computational resources. We have previously leveraged GPUs and FPGA/Dataflow systems and are now focusing on Artificial Neural Networks (ANNs) as surrogate models. Key research areas in this umbrella topic include:
– Transfer & Meta-Learning: Rapidly fine-tuning ANNs for new SNN architectures.
– Dynamic Surrogates: Crafting adaptive ANNs for desired accuracy levels.
– Optimization: Exploring techniques to boost ANN performance.
– Hybrid Training: Developing tools for on-the-fly data generation and training on GPUs or FPGAs.