I: Artificial Intelligence Flashcards
What can AI be viewed as?
AI can be viewed as mechanical intelligence that requires computational systems to perform somewhat intelligent behavior.
What is artificial intellligence?
The exact definition of AI is generally vague but general consensus is that AI are systems that exhibit behavior similar to human thinking, such as handling decision-making automatically, finding deviations in data sets and carrying out conventional conversations.
When did the first AI-technique become available?
In 1943, the first AI-technique started. Since, other things that have been carried out are very important for AI and for autonomous systems carrying out tasks in an environment.
What is the Turing Machine?
The Turing machine comes from 1936 and is the first theoretical automatic machine that was used for solving problems surrounding computation using rigorous mathematical models. Turing applied the idea of problem-solving and intelligence as a basis for computers.
Who is John McCarhty? What is LISP?
John McCarthy coined the term “Artificial intelligence” in 1995 and is generally considered one of the founding fathers of AI. He invented the early programming language LISP. LISP was historically one of the more used program languages for AI, alongside Prolog, and over the years several programming languages have come from it.
What is the Turing test? Has it been passed?
It is a test modeled by Alan Turing in 1950 and its main principle is to test whether or not a human can tell apart a human from a computer. If a human can guess that the computer system is not a human, then the computer system is not intelligent enough. However if people cannot guess this in over 30% of the trials, then the computer system is considered intelligent. When using AI this system is obviously a chat-bot.
In 2014, a computer AI called Eugene Goostman supposedly passed the Turing test, however this supposed “passing” has been questioned over the years.
What is the Chinese Room?
John Searle argued against the Turing test by saying that no matter the intelligent behavior exhibited by a computer, a computer system cannot be shown to have mind, understanding or consciousness. The Chinese Room is a closed room with one person in the room, the person gets a number of arbitrary questions written in Chinese and has to answer in Chinese. The person does not understand Chinese but has detailed instructions and a set of rules on how the answers would be formed. Is this person intelligent?
What is IBM Deep Blue?
IBM Deep Blue is a computer system that in 1997 played against chess master Garry Kasparov and won by calculating 200 million calculations per second. Is it intelligent?
What is an issue with AI?
They do not seem to employ human kinds of intelligence.
They do not have intelligence above what they have been specifically taught.
“As soon as it works, no one calls it an AI” many systems that use aspects of AI are not called AI.
What are AI winters?
AI winters refer to the periods of times when there have been a long pause in between breakthroughs in the technology. There have been several hype cycles with meh results and critique after it. A major factor that played into these winters have to do with the fact that after successes this did not scale into other problems and could not be used in a generalized setting.
What areas are there in AI?
- Natural language processing
- Machine learning
- Neural networks
- Deep learning
- Bio-inspired AI
- Cognitive Science, Cognitive computing and computation
- Expert systems, knowledge-based systems, decision support systems
- Multi-agent systems, agent-based systems
- Robotics Systems
- Perception and computer vision, virtual reality and augmented reality
What are the three most common and different categories of intelligence?
- Artificial narrow intelligence
- Artificial general intelligence
- Artificial superintelligence
Shortly explain artificial narrow intelligence?
Artificial Narrow Intelligence (ANI) specializes in one area at a time. In ANI, the task provided can be performed well however it cannot be generalized to handle several different tasks. Weak AI and ANI are not the same thing: Weak AI explains which AI tasks can be performed while ANI has the ability of computer systems to perform a single task extremely well.
Shortly explain artificial general intelligence:
Artificial General Intelligence (AGI) refers to a system intelligence that can understand and learn the same intellectual task as a human and would be capable of computing the same tasks as a human. AGI represents a generalized version of human cognition and is general. AGI is often compared to strong AI: strong AI refers to a system with intellectual abilities indistinguishable from those of a human and which are able to work with different tasks in several complex domains.
Shortly explain artificial superintelligence?
Artificial superintelligence (ASI) refers to systems that possess intelligence surpassing human minds and are thus smarter. ASI can handle scientific creativity, has general wisdom and social skills.
