Feature Engineering Challenges
Daniel discusses the challenges of feature engineering on embedded devices, particularly with high frequency sensor data like audio. He explains the need to distill out pertinent information before it reaches the model, and the trade-off between signal processing algorithms and model size. The goal is to build tools that allow practitioners to optimize the entire system, rather than just focusing on the model.In this clip
From this podcast
The Gradient
Daniel Situnayake: AI on the Edge
Related Questions
What problems do developers face when building AI applications as discussed in the episode Daniel Situnayake: AI on the Edge and the clip Streamlining Embedded Machine Learning?
How do you leverage different models in machine learning as discussed in the episode 549-william-falcon-optimizing-deep-learning-models and the clip Data Transformation Challenges?