Artificial Intelligence

“Let’s clear this out once and for all!” Luco Ny Aina Yannick RAMAROMANANA
R&D Developers Forum
July 2018

© 2018 Ceridian Corporation. All rights reserved.

Summary
• What is AI?
• Benefits and Risks of AI
• How to AI?
• Demo
What is AI?
What is AI?
AI : Artificial Intelligence
• Intelligence demonstrated by machines
• Similar to human/animal intelligence but relatively more/less efficient
• Computer Programs capable of
• Reasoning
• Planning
• Learning
• Processing Natural Language
• Deciding actions without explicit programming
• Performing tasks that require some kind of intelligence
• …
What is AI?
AI : Artificial Intelligence

Sophia: First non-human Ava: A sentient robot

citizen. (AGI) in the Ex-Machina
movie
What is AI?
The roots of AI
• Philosophy
• Logic/Mathematics
• Computation/Programming
• Psychology/Cognitive Science
• Biology/Neuroscience
• …
What is AI?
Types of AI
• Type I: Weak AI, Narrow AI, Applied AI, Reactive Machines
(Everyday life AI)
• Type II: Strong AI, Self-Aware AI, AGI (As Intelligent as humans)
• Capable of emotions, creativity and originality
• Not limited by its program boundaries
• Completely autonomous, spontaneous
• Seen in Sci-fi only (for the moment)
• Type III: The singularity, ASI, Superhuman Intelligence, (Way more
intelligent than humans)
• Can’t be imagined
• Unlikely
What is AI?
Types of AI
What is AI?
Fields of AI (Type I)
What is AI?
Applications of AI in Engineering (Type I)
• Spam filters (gmail)
• Non-player character / Computer opponent in games (fifa games)
• Recommendations (facebook ads, youtube, google search)
• Predictions (google maps traffic prediction, weather prediction)
• Chatbots (clevert bot, messenger bots)
• Speech recognition (smart assistants like siri, cortana, alexa)
• Facial recognition (facebook pictures auto-tagging, iphone X face ID)
• Object detection (google image search, people counting)
• Self-driving cars (Pathfinding, Computer vision, Speech recognition, Object detection)
• Flights auto-pilot (Physics simulation)
• Military autonomous drones
• Space/Planet exploration (Mars’ Curiosity rover)
• Surgery assistant
• …
Benefits and Risks of AI
Benefits and Risks of AI
Benefits of AI
• Automation: repetitive/laborious boring tasks, smart assistants
• Speed and precision: heavy computation, rapid decision making
• Safety/Security: auto-pilots, smart home, fraud detection
• Simulation: crowd simulation, chemical experiments, plague prevention
• Exploration: space, ruins, mines
• Entertainment: video games, cinema, robot pets
• Research: cognitive science and neuroscience, mathematics, biology, AI Safety
• …
Benefits and Risks of AI
Risks of AI
• Actual (Type I):
• Loss of jobs
• High cost
• Low number of specialists
• AI arms race
• Programming bugs/failures
• Future (Type II & III):
• Human cockroach
• War of the machines
• Decrease of human mental capacity
• High dependency on the machines
How to AI?
How to AI?
1. Familiarizing yourself
• Read introductory articles/books/blogs/videos
• Choose a subfield of interest (eg. Machine Learning)
• Use third party tools/libraries (eg. Azure ML, Tensorflow)
2. Getting proficient
• Write/Expand your own library
• Learn appropriate Algorithms and Maths
• Recreate existing AI programs or Create your own using existing techniques
3. Becoming Expert/Researcher
• Find new AI techniques/algorithms/subfields
• Improve existing ones
• Research on AI safety
Demo
Demo
Taxi Planner ML
• What?: Auto assignment of taxi/van users for Ceridian
• Why?: Manual assignment takes about 1 and half hour every day for hundreds
of employees VS. Auto assignment takes a few seconds
• How?: Using a combination of a Machine Learning (Clustering) and Pathfinding
techniques
• Tools used: .NET Web API (C#), VS 2015, SQL Server 2016, Google Map API
• Credits: The MAU R&D AI Team A
• Luco Ramaromanana (SSE)
• Nivedita Bissoon (QA)
• Brayen Summogum (ASE)
• Hok Fei (ASE)
Demo
Fandrika game
• What?: A board game based on Isolation game
• Why?: For the sole purpose of Learning AI
• How?: Using Minimax & Alpha/Beta pruning algorithm (based on Udacity
courses)
• Tools used: Netbeans IDE, Gradle, LibGDX (java)
• Credits: Luco Ramaromanana (SSE)
Useful Links
• https://en.wikipedia.org/wiki/Artificial_intelligence
• http://jmc.stanford.edu/artificial-intelligence/what-is-ai/index.html
• https://www.techopedia.com/definition/190/artificial-intelligence-ai
• http://theconversation.com/understanding-the-four-types-of-ai-from-reactive-robots-to-self-aware-beings-67616
• https://en.wikipedia.org/wiki/Outline_of_artificial_intelligence#Types_of_artificial_intelligence
• https://community.hpe.com/t5/Digital-Transformation/The-6-types-of-artificial-intelligence/ba-
p/6976199#.WvWCwaSFOM8
• https://www.quora.com/What-are-the-types-of-artificial-intelligence
• https://en.wikipedia.org/wiki/Slaughterbots
• https://www.youtube.com/watch?v=WSKi8HfcxEk
• https://www.youtube.com/watch?v=XAgXwUwQoPA
• https://www.youtube.com/watch?v=TlO2gcs1YvM
Questions?