Unit 6.3 - artificial intelligence

A branch of computer science dealing the the simulation of intelligent human behaviors by computers; it refers to a system’s ability to correctly interpret external data, to learn from it and to use those relevant learning to achieve specific tasks through flexible adaptation. Its data model of past experiences help it plan & predict.

• Collection of data • Rules for using/analyzing the data • Ability to reason (make decisions, detect patterns) • The ability to learn and adapt (update understanding)

Training AI machine to recognize and remember patterns by feeding it with lots of examples, then feeding the machine new data and letting it decide what to do with it/reason independent of human interference.

Huge amounts of data that allow deep analysis of patterns and predictions of behaviors by AI.

Systems that rely on humans to constantly update their data model and rules required to process the data in order to provide information to human experts by data analysis (which they perform by applying rules to the massive amounts of stored data in order to infer results) They imitate the decision-making ability of human experts and are designed to solve complex problems based on set of information and rules.

A type of programming where the intended outcome, along with the data model and rules are created but the precise steps to achieve the outcome are generated by the computer.

• Knowledge base - contains facts and data • Rules base - contains rules & logic (IF THEN conditions) which expert systems use to arrive at a conclusion • Inference engine - machine that processes rules & logic in order to arrive to a conclusion/result • Interface to allow the user to enter querries and view results

Machine learning is when a program has the ability to automatically adapt its own processes and/or data. A program/computer algorithm that improves automatically through experience. It’s trained to discover patterns in data and make predictions & decisions based on those past patterns; capable of updating their own data models when they encounter new data. It uses statistical techniques to learn how to get progressively better at a task, without having been specifically coded for the purpose.

• Self-driving cars • Smart assistants • Interpreting medical images • Fraud detection

• ANI - artificial narrow intelligence which specializes one area and solves one particular problem; machine learning • AGI - artificial general intelligence which refers to a computer that is as smart as a human across the board; machine intelligence • ASI - artificial super intelligence which is an intellect that is much smarter than the best human brains in almost every field; machine consciousness

Historical data

All types of data - images, sounds, text, numbers, temperature readings and more!

+ high level of accuracy & expertise, consistent results, faster response than human experts, can store large amounts of data & facts - only as good as the data entered into the system, optimal use requires training, responses can be cold & lack humanity, may still make mistakes

Diagnosing medical problems; financial calculations; diagnosing faults in equipment;

• Online customer support chat-bots • Self-driving cars • Tailored ads

• Expert systems have no learning capabilities and must be manually programmed while machine learning learns automatically from data inputs • ES mimics expert decision-making by using a set of rules while ML discovers patterns in data in order to make predictions • ES require humans to update their data models while ML can update them themselves when learning from new data

• Can help robot be more flexible/adapt to new situations by updating its rule base • Make it more efficient by having greater knowledge of (for example) surroundings