aibo, 'Interaction' development with Neural Network Console and Libraries
summary
Aibo is a shipping consumer robot that uses camera, time‑of‑flight, position sensors, capacitive touch and multiple mics to find faces and interact, with on‑device behavior driven by collected sensor datasets and trained models. Sony uses Neural Network Console (GUI) to drag‑and‑drop networks and showed a classifier jump from ~19.5% to ~97% accuracy by adding a layer and to ~92% with convolution, and publishes Neural Network Libraries under Apache 2.0 with float16/tensor‑core support for fast GPU training.
AI-written summary · about
chapters (11)
- 0:00Introduction to aibo
- 1:35What is aibo robot
- 3:09aibo interaction structure
- 6:26Machine learning for aibo
- 8:54Neural Network Console intro
- 11:49Why use Neural Network Console
- 12:12Neural Network Console demo
- 15:45Applications of deep learning
- 17:12Neural Network Libraries overview
- 18:41Open source tools availability
- 19:45Conclusion and resources
