Better Living through Deep Learning and Edge Computing Peggy James
Smart cities and big data
What is a smart city ? • Connected • Participatory • Organic
Networked Vertically and Connected Horizontally Multidirectional
Democratic Data Collection Mobocracy tamed and ordered Flattening of Participatory information influence Equity
Group functionality Individual identity • Organic Creation of community • I am you and you are me and we are all together • transindividuality and the multitude
1. Creation of Technology Relationships 2. Contextualize lots of data 3. Dynamic decision making at point collection and the tech infrastructure 4. Long term adaptations create resiliency 5. Short term adaptations create efficiency 6. The city environment is continuously responsive Concepts Schmoncepts– how does it work?
An environment that is built upon a P4 idea
Deep learning (3+ layers) Neural networks Let the algorithm decide what’s important to solve a problem, not a human..
DNN- Dynamic Neural Networks Extract features, layers provide a feature hierarchy Work with unstructured unlabeled data – raw data Provide the opportunity for unsupervised learning
Edge computing Time sensitive Process optimization Decentralized processing power
Your Gov allows citizens to Reported issues are easily report non- submitted directly emergency issues. into Waukesha YourGOV can be County's work order used to report issues system. Residents such as potholes, with an account can vandalism and view issues service requests for reported, and Waukesha County receive confirmation highways, parks and when the issue has facilities. been resolved
What about a smart community? Decentralized decision Dynamic making learning management Human Equitable machine access interfacing
RESILIENT SUSTAINABLE SENSIBLE EQUITABLE CREATIVE INCLUSIVE The more I dig into it, the more I realise the Smart City sector might not be a sector, it’s more of a way of thinking.” Batchelor, The Smart City podcast 9/3/2018
Songdo
Amsterdam
Masdar City
AoT in Chicago
@ParksideSSPS
Recommend
More recommend