Practical AI Generating "hunches" using smart home data đźŹ
May 4, 2021
Evan Welbourne, who leads Amazon's smart home machine learning team, dives into the complexities of synthesizing smart home data to enhance user experiences. He reveals how the pandemic has influenced consumer behavior, creating anomalies in data. Evan elaborates on the innovative 'hunches' feature, which personalizes interactions by predicting user needs. He discusses the importance of trust and the challenges of data integration and security, while also touching on future integrations of smart home tech with automotive advancements.
AI Snips
Chapters
Transcript
Episode notes
Defining Smart Homes
- Smart homes aren't new; home automation tech has existed since the 80s.
- Amazon's vision focuses on an "actually smart home," using AI (Alexa) to simplify home management and achieve high-level goals like sustainability and safety.
Smart Home Interaction
- Smart home interaction modes include directed control (voice commands), programmed control (routines), and intelligent control (hunches).
- Intelligent control allows Alexa to act autonomously, based on learned behaviors and user goals, for tasks like energy saving and security.
Smart Home Adoption Challenges
- Early smart home adopters were tech-savvy and willing to overcome initial hurdles.
- Amazon aims to simplify adoption for less tech-savvy users by making setup and error recovery easier, enabling broader access to the technology.