Artificial Intelligence Research Flashcards
What are 2 classifications of AI?
rule-based and neural-network based
What does rule-based AI focus on?
symbols and logic
From which question did rule-based AI start?
How to symbolize concepts or intelligence?
What are the main areas of interest with rule-based AI?
mathematical theorem proving and natural language processing
What did rule-based AI shift towards in the early 1980s?
expert systems applications
In what areas does rule-based AI perform well?
deductive reasoning, logical reasoning, expert systems
True or false: No groundbreaking methodology has arisen for rule-based AI since the mid 1980s?
true
What are some limitations of rule-based AI?
training difficult to implement, limitations in pattern recognition for text, voice, video, etc.
Where do neural networks take their ideas from?
the connections between neurons in the human brain
When and with who did the neural network research begin?
1957, Rosenblatt
After what did neural network research decline rapidly?
the publication of ‘Perceptrons’ in 1969
What is the latest learning method in the neural network family?
deep learning
What is the main difference between rule-based AI and ANN-based AI?
rule-based asks “What does the brain do?” and ANN-based asks “How does the brain work?”
What are the research fields for rule-based AI?
mathematical theorem proofs/problem solving, natural language processing (NLP), and expert systems
What are applications of NLP AI?
chatbots, translators, spam filters, search engines
How does expert system work?
input from the user interface goes through a rules engine to get information from a knowledge base created by knowledge from an expert, then returns output/advice/information to the user
What does an ANN consist of?
a network of numerous neural network processors, like a brain
What are research fields for ANN-based AI?
character, pattern, and image recognition, generative models, generative adversal network, and reinforcement learning
List some research topics for ANN-based AI.
biological neural networks, learning ability and characteristics of NN, printed and hand-written character recognition, real-time voice recognition, image recognition technology, object recognition within videos
What were the prominent characteristics of AI research in the Early era (60~70s)?
problem solving, processing and translation of natural language, games, character and number recognition (NN), proof of new mathematical theorem
When was the first chess playing program created and by whom?
1951 Dietrich Prinz
What were the prominent characteristics of AI research in the 1980~2000s?
expert systems, DeepBlue in 1997, intelligent agent research (ex roomba, self driving cars), ML such as Bayesian networks
What were the prominent characteristics of AI research from 2000s until now?
deep learning algorithms, Watson, AlphaGo, big data
What are 5 AI systems that impressed the world?
Mark I Perceptron, Mycin, DeepBlue, Watson, AlphaGo