Questions about AI (theoretical)

Just some questions about AI, which I don’t know where to ask. Will probably use it in the future.

  • Is machine learning just a “framework” which you can put into a random situation and it would learn?
  • what actual types of AI exist? Is ML just a type of AI?
  • what is the point of the turing test? How is it supposed to show intelligence?
  • why would AI lack imagination? Is it because it is always rational? But how would rationality block imagination?
  • what are the difficulties with AI
  • what is the difference between different types of AI?
  • Where I was looking for information, it showed that neural networks, machine learning and visual recognition were all sub-types of AI. So how does visual recognition work, if not for a neural network + some tricks to get the actual info
  • what actually is an AI? I got so many different descriptions; is it just an attempt to simulate intelligence?

Thank you in advance : D

4 Likes

oh and what actual models of AI exist?

1 Like

I share some of these questions in common too. It would be really helpful if someone provides an insight.

2 Likes

You won’t get a real answer because the field of AI is huge and over 60 years old.

Books over books have been written about it.

Tic Tac Toe

Here in the forum for example we discussed Tic Tac Toe.

Of course two humans can play each other. But you can also play against a computer. We used to call the computer player in the Sketch “AI” in the discussion. But of course we all knew it was a metaphor, and not intelligent / conscious / sentient like a human. (It would not be able to play chess or to learn it like a child only because it knows how to play Tic tac toe)
But the AI was able to make a move. Fast. And maybe better than a human in some cases.
How so? There are different ways the machine plays:

  • An algorithm like check all lines and make a move where you can win XXX, or hinder a win of your opponent OOX or make a line with 2 X’s where you can win later X_X etc…
  • Minimax Algorithm (see coding train) which can be used here and also in Chess. Minimax - Wikipedia
  • Alpha–beta pruning is similar. Remember these are technologies that beat a World Chess Champion.
  • train a neural network. That would be Machine learning
  • Googles Alpha Zero combines different techniques

But AI refers to a bunch of different techniques here. You can see AI is more a container word and so is machine learning.

image recognition

For image recognition check out OpenCV.

AI nowadays is just crunching of numbers and recognizing patterns in data. AI doesn’t have any awareness of what’s going on, or intention, goals or planning [not in a human sense when we pursue a project]. Therefore there is no imagination.
There are AIs who search for new concepts in the field of mathematics and can write certain proofs.

An interesting field is also

  • IBM Deep Blue, first machine to beat a World Chess Champion in a tournament and
  • IBM Watson who beat two Master Players in the game Jeopardy. see Watson (computer) - Wikipedia

AlphaGo

In 2016, Google DeepMind’s program AlphaGo beat Lee Sedol. It applies Deep Learning and combines several highly complex algorithms.

In October 2017, DeepMind announced a significantly stronger version called AlphaGo Zero which beat the previous version by 100 games to 0.

see Go (game) - Wikipedia

AlphaGo Zero

AlphaGo Zero is a version of DeepMind’s Go software AlphaGo. AlphaGo’s team published an article in the journal Nature on 19 October 2017, introducing AlphaGo Zero, a version created without using data from human games, and stronger than any previous version.[1] By playing games against itself, AlphaGo Zero surpassed the strength of AlphaGo Lee in three days by winning 100 games to 0, reached the level of AlphaGo Master in 21 days, and exceeded all the old versions in 40 days.[2]

Training artificial intelligence (AI) without datasets derived from human experts has significant implications for the development of AI with superhuman skills because expert data is “often expensive, unreliable or simply unavailable.”[3] Demis Hassabis, the co-founder and CEO of DeepMind, said that AlphaGo Zero was so powerful because it was “no longer constrained by the limits of human knowledge”.[4] David Silver, one of the first authors of DeepMind’s papers published in Nature on AlphaGo, said that it is possible to have generalised AI algorithms by removing the need to learn from humans.[5]

Google later developed AlphaZero, a generalized version of AlphaGo Zero that could play chess and Shōgi in addition to Go. In December 2017, AlphaZero beat the 3-day version of AlphaGo Zero by winning 60 games to 40, and with 8 hours of training it outperformed AlphaGo Lee on an Elo scale. AlphaZero also defeated a top chess program (Stockfish) and a top Shōgi program (Elmo).[6][7]

from AlphaGo Zero - Wikipedia

Machine learning

Machine learning ( ML ) is the study of computer algorithms that improve automatically through experience.[1] It is seen as a part of artificial intelligence. Machine learning algorithms build a model based on sample data, known as “training data”, in order to make predictions or decisions without being explicitly programmed to do so.[2] Machine learning algorithms are used in a wide variety of applications, such as email filtering and computer vision, where it is difficult or unfeasible to develop conventional algorithms to perform the needed tasks.

A subset of machine learning is closely related to computational statistics, which focuses on making predictions using computers; but not all machine learning is statistical learning. The study of mathematical optimization delivers methods, theory and application domains to the field of machine learning. Data mining is a related field of study, focusing on exploratory data analysis through unsupervised learning.[4][5] In its application across business problems, machine learning is also referred to as predictive analytics.

from Machine learning - Wikipedia

Artificial intelligence

Artificial intelligence ( AI ) is intelligence demonstrated by machines, unlike the natural intelligence displayed by humans and animals, which involves consciousness and emotionality. The distinction between the former and the latter categories is often revealed by the acronym chosen. ‘Strong’ AI is usually labelled as AGI (Artificial General Intelligence) while attempts to emulate ‘natural’ intelligence have been called ABI (Artificial Biological Intelligence). Leading AI textbooks define the field as the study of “intelligent agents”: any device that perceives its environment and takes actions that maximize its chance of successfully achieving its goals.[3] Colloquially, the term “artificial intelligence” is often used to describe machines (or computers) that mimic “cognitive” functions that humans associate with the human mind, such as “learning” and “problem solving”.[4]

As machines become increasingly capable, tasks considered to require “intelligence” are often removed from the definition of AI, a phenomenon known as the AI effect.[5] A quip in Tesler’s Theorem says “AI is whatever hasn’t been done yet.”[6] For instance, optical character recognition is frequently excluded from things considered to be AI,[7] having become a routine technology.[8] Modern machine capabilities generally classified as AI include successfully understanding human speech,[9] competing at the highest level in strategic game systems (such as chess and Go),[10] self-driving cars, intelligent routing in content delivery networks, and military simulations.[11]

Artificial intelligence was founded as an academic discipline in 1955, and in the years since has experienced several waves of optimism,[12][13] followed by disappointment and the loss of funding (known as an “AI winter”),[14][15] followed by new approaches, success and renewed funding.[13][16] After AlphaGo successfully defeated a professional Go player in 2015, artificial intelligence once again attracted widespread global attention.[17] For most of its history, AI research has been divided into sub-fields that often fail to communicate with each other.[18] These sub-fields are based on technical considerations, such as particular goals (e.g. “robotics” or “machine learning”),[19] the use of particular tools (“logic” or artificial neural networks), or deep philosophical differences.[22][23][24] Sub-fields have also been based on social factors (particular institutions or the work of particular researchers).[18]

The traditional problems (or goals) of AI research include reasoning, knowledge representation, planning, learning, natural language processing, perception and the ability to move and manipulate objects.[19] General intelligence is among the field’s long-term goals.[25] Approaches include statistical methods, computational intelligence, and traditional symbolic AI. Many tools are used in AI, including versions of search and mathematical optimization, artificial neural networks, and methods based on statistics, probability and economics. The AI field draws upon computer science, information engineering, mathematics, psychology, linguistics, philosophy, and many other fields.

The field was founded on the assumption that human intelligence “can be so precisely described that a machine can be made to simulate it”.[26] This raises philosophical arguments about the mind and the ethics of creating artificial beings endowed with human-like intelligence. These issues have been explored by myth, fiction and philosophy since antiquity.[31] Some people also consider AI to be a danger to humanity if it progresses unabated.[32][33] Others believe that AI, unlike previous technological revolutions, will create a risk of mass unemployment.[34]

In the twenty-first century, AI techniques have experienced a resurgence following concurrent advances in computer power, large amounts of data, and theoretical understanding; and AI techniques have become an essential part of the technology industry, helping to solve many challenging problems in computer science, software engineering and operations research.[35][16]

from Artificial intelligence - Wikipedia

4 Likes

An AI that solves Chess or Tic Tac Toe is a special purpose AI (weak AI).

The Turing Test is aimed at a strong AI that simulates a real Person (simulates /has consciousness).
To find out if the AI is truly intelligent the Turing Test was proposed. In a conversation situation where you can’t see the partner (conversation via chat/e-mail) the testing person has to decide whether it talks to a person or to a machine. That’s more about simulating and imitating an dialogue/personality.

see Turing test - Wikipedia

2 Likes

Thank you! It is really useful and helped me a lot in piecing together all the scattered information across the internet. Thank you for taking the time.

I wish you a nice day!

2 Likes

Would this be considered unsupervised learning?

1 Like

Yes. But there are different kinds of unsupervised learning

AI for Games

Also I looked here: Quarks Place

A library by quark

Tensor Flow

Also, Tensor Flow is interesting, not sure if it’s connected to processing

See TensorFlow - Wikipedia

See also

see Neural Network genetic algorithm game

Keywords

AI, KI, artificial intelligence, Künstliche Intelligenz, Machine Learning

2 Likes

The Coding Train has a dedicated playlist for it ig

@Chrisir Thanks for explaining!

2 Likes

By the way @Chrisir, given that the turing test is aimed at AGI (strong AI), could the judge just ask the computer to play chess with him? Or are there any limits to possible questions? On one hand you could play chess through text (like A6 to A7) and on the other you would need the whole graphical interface.

(Sorry for bothering you, but I found that topics releated to fast evolving technologies (especially AIs) are most accurate when asking a professional, rather than the probably outdated internet articles)

1 Like

Have you checked out rob miles on computerphile youtube channel. Or any of the other vids on the subject. Very interesting maybe not aimed at people with an already broad knowledge but excellent explanations

3 Likes

I am really not an expert.

There are different setups for the Turing Test, even with a jury and votes.

I guess it would be against the rules to play a game (and use a software for this).

4 Likes

This seems to be the reference

Not cheap though…

Also check out

  • Nick Bostrom, Superintelligence. Paths, Dangers, Strategies.
2 Likes

well, ML is a branch of AI. The confusion may arise because both ML and AI are very broad terms, of very broad disciplines! Before ML became pragmatically usable (because of some hardware and software breakthroughs ie GPU compute and backpropagation algorithm) most AI systems were rule-based. Rule-based system - Wikipedia

You would be surprised about the amount of applications that sell themselves as machine learning but are actually rule based systems under the surface.

1 Like