🧠 Introduction to AI
Neu-Ulm University of Applied Sciences
February 12, 2024
Which thoughts and feelings come to your mind when you think of AI?
Fiction
or future?
I believe it’s going to change the world more than anything in the history of mankind — even more than electricity. Kai-Fu Lee
The pace of progress in artificial intelligence is incredibly fast. Unless you have direct exposure to groups like Deepmind, you have no idea how fast—it is growing at a pace close to exponential. The risk of something seriously dangerous happening is in the five-year timeframe. 10 years at most. Elon Musk
Forget artificial intelligence—in the brave new world of big data, it’s artificial idiocy we should be looking out for. Tom Chatfield
AI is the science of making machines to
Cognitive science is the study of the human brain and its processes — it examines how the human brain may be functioning. Cognitive science requires analytical observation and experimentation.
We can learn about human thought in three ways (Russel and Norvig 2022):
Some of the most powerful AI models are a result from observing human thinking experimentally (e.g., deep neural networks).
What is rational thinking about?
The “laws of thought” refer to fundamental axiomatic rules upon which rational discorse itself is often considered to be based.
Socrates is a man and all men are mortal, thus, it can be concluded that Socrates is mortal Aristotle (384-322 BCE)
Computers have been able to solve any solvable problem, as long as
The Turing Test1 (Turing 1950) tests if a computer has the ability to mimic peoples’ behavior.
To pass the test, it would need following capabilities:
Want to do a Turing Test? Play “Bot or Not”
Would ChatGPT pass the turing test? Why (not)?
Large language models in general, have the ability to produce human-like responses that can fool even experienced evaluators ChatGPT has shown it can. However, depending on your prompting skills, those models may still produce a lot of nonsense.
An agent is something that acts, an rational agent is one that acts so as to achieve the best coutcome (i.e., does the right thing), or, when there is uncertainty, the best expected outcome (i.e., does the appropriate thing) based on the objective that is provided to the agent (Russel and Norvig 2022).
The approach goes beyond the “laws of thought” approach as it involves actions based on
Do you see any issues with the rational agent approach?
Machines that are provably beneficial to humans
According to Russel and Norvig (2022) two refinements to the standard model of AI are needed:
‘AI system’ means a machine-based systems designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment and that, for explicit or implicit objectives, infers, from the input it received, how to generate output such as content, predictions, recommendations, or decisions, that can influence physical or virtual environment (European Commission 2024).
Systems that perceive, learn, think and act human-like.
A series of interviews to some of the AI pioneers.
The full documentary is available here
1943—1956
The inception of AI
1966—73
A dose of reality (AI winter)
1970—90
Expert systems (knowledge-based approaches)
1990—present
AI spring (statistical approaches)
2012—present
New excitement
Improvement in performance obtained from increasing the size of the data set by two or three orders of magnitude outweighs any improvement that can be obtained from tweaking the algorithm Banko and Brill (2001)
Which AI-powered systems are you using?
The exercises are (inspired) by Russel and Norvig (2022)
Define in your own words:
If and to what extent are the following computer systems instances of artificial intelligence?
Various subfields of AI have held contests by defining a standard task and inviting researchers to do their best. Examples include the DARPA Grand Challenge for robotic cars, the International Planning Competition, the Robocup robotic soccer league, the TREC information retrieval event, and contests in machine translation and speech recognition.
Investigate one of these contests and describe the progress made over the years.
Read the statements (one after the other) and discuss if the second sentence of each statement is true and if it does imply the first.
Surely computers cannot be intelligent
—they can do only what their programmers tell them.
Surely animals cannot be intelligent
—they can do only what their genes tell them.
Surely animals, humans, and computers cannot be intelligent
—they can do only what their constituent atoms are told to do by the laws of physics.