when iot meets artificial intelligence
TRANSCRIPT
When IoT Meets Artificial Intelligence
Veselin Pizurica
Internet of Things Event, 5th edition, 8/06/2016
Why so much interest in IoT?
Why now? - Perfect storm● Cost of adding new connected sensors/actuators has come down dramatically● Connectivity ● Cloud● API economy ● Big Data/Analytics● AI● Robotics
Devices are becoming widely available
Off-the-shelf gadgets Programmable devices
API economy● APIs have become new patents● Who holds the data, holds the knowledge● Companies don’t share their know-how, but they are willing to share their
know-what (via Application Programming Interface API)● API economy will be the major driver of the profit for many companies
Weather API - monetization
Big data analytics
What connects these two pictures?
Intelligence - where?
“Swarm” intelligence Logic in the gateway
“Fog” computingLogic in the cloud
Logic in the device
Swarm Intelligence - in sensor networks?
● Limited storage, power and processing power● Sensors and actuators local
Fog computing● Anomaly detection● Compress sensing (not for computing, but bandwidth optimization, as data
leaves the edge)● Fast reaction time● No privacy issues if data doesn’t leave the edge● Doesn’t work for LoRA and Sigfox, as data deduplication happens in the cloud● Mostly in factory settings - transition from SCADA (legacy) systems to more
internet oriented solutions
Why NOT intelligence in the cloud?● Latency requirements● Failure (in)tolerance (lack of redundancy) – adding more blocks system even
less stable● Cost of pushing data in the cloud (storage, bandwidth)● SW cost of integration● Lack of standardization● Security concerns: Authentication/Authorization● Privacy concerns
Why intelligence in the cloud?● Device-agnostic and decouples logic from the presentation layer● Combination of the sensor data with API “economy” ● Integrating multiple IoT vertical solutions● Cloud-capacity scales horizontally, while distributed HW often needs to be
swapped when HW resources are no longer sufficient● Cloud intelligence also allows easy generation of analytics regarding the
usage of the logic itself. Which rules fired and why? How often?● An architectural model arises where logic is built together with a REST API
Our vision
IoT reference model is suboptimal
Critical in IoT is the ability to process data in real-time as they come in, i.e.the ability to act on data in motion.
We need a technology that effortlessly can blend event-based and query based data in real-time, not one or the other!
So let’s talk about AI!Y = f (X) Y = f (X)
How do we evolve to a programmable world?
Rule engine is a
knowledge modeling problem
Y = f (X)
IoT/API Rule Engine Challenges● Changes of the (customers’) environment and requirements.● Lack of compact representation, leading to difficult simulation, debugging and
maintenance.● Rule engines don’t provide us with easy ways to gain additional insights: why
a rule has fired and under which conditions?● Combining data from the physical world (PUSH mode) with data from the “API
world” (PULL mode). ● How long do we wait for the next information to come before deciding to move
on in decisions? ● How long is the measurement is valid?
Bayes and inference engine to the rescue!
Waylay Rule Engine is a Cloud Smart Agent
Waylay platform