Artificial Intelligence A very short introduction by Boden, Margaret A. Book Review
Hello everyone, in this article i will talk about Artificial Intelligence A very short introduction book by Boden, Margaret A.
Very Short Introductions is a series of books published by Oxford University Press. The books are short introductions to specific topics, aimed at a general audience, but written by experts. Most of them are less than 200 pages long. While the authors can offer personal perspectives, the books are intended to be “balanced and complete” as well as thought-provoking.
The book does not give overly detailed information about artificial intelligence that it does not have such a purpose, but if you have a slight interest in artificial intelligence, if you want to learn a few new things, or if you are looking for an ideal alternative to enter the artificial intelligence, it will be an ideal book. The book provides a relatively satisfactory knowledge of artificial intelligence and contributes to a good background.
In the first part of the book, a general introduction to artificial intelligence is made, the main objectives of artificial intelligence are mentioned, and chronological developments of artificial intelligence are introduced by referring to the chapters explained in more detail in the book. Many models and concepts are explained very well in the book with concrete examples. For this reason, the book conveys the basic information of artificial intelligence in a refreshing way, intriguing, and more concrete examples which away from boring.
In the second part of the book, the concepts are explained in detail and deeper information about artificial intelligence is conveyed to the reader. Supercomputers, computer vision, heuristics, frame problems, agents and machine learning are discussed in this section in a very descriptive manner and with examples. In the machine learning part, supervised, unsupervised and reinforcement learning are briefly mentioned. Concepts are dealt with in more detail without boring the reader again after the introduction.
In the third part of the book, the effects of language, creativity, and emotion modeling on artificial intelligence are discussed in a very descriptive way. The challenges and issues that the NLP needs to address are reinforced by giving concrete examples. The fact that machine translation was thought to be an impossible and wasted field at first, made this field turn to semantic context, words & phrases rather than syntax. At the same time, this topic has been made quite understandable with the example of Google Translate. The contribution of NLP to big data and data mining is also explained with explanatory examples. In the “creativity” section about artificial intelligence, which has confused my mind many times before, the fact that even exploratory AI depends on human judgment is explained with examples and I was surprised.
AI creativity has many applications. Sometimes it can meet or even exceed human standards in a small corner of science or art. But in the general situation pairing human creativity is a completely different matter. At the same time, I was thinking that emotions are far from artificial intelligence but I have learned that If we are ever to achieve AGI, emotions such as anxiety will have to be included and used. This is explained in a very self-explanatory way; it changed my mind.
As mentioned in the fourth part of the book, I thought neural networks were virtual machines of symbolic artificial intelligence, but I learned that the network is usually simulated by a Von Neumann machine, that is, ANNs are parallel computed virtual machines implemented in conventional computers. The book generally corrects known mistakes with a simple expression.
In the expressions related to deep learning networks and the brain, the names of regions of the brain are used as terms (for example, cerebral cortex), but for those who do not know these concepts, this part can become a bit difficult to understand and leads to research.
In the part of the book where artificial neural networks are described, the ideas that rendered artificial neural networks dysfunctional starting in the 1960s and how this field entered a dark period are explained very well. How a scientific scandal begins and ends is covered without boring the reader. However, it is narratively explained that the most appropriate method for the human brain approach is sequential and parallel. At the same time, it is not only dealt with from a psychological and philosophical perspective but also its commercial interactions. Although this part is very interesting to me, I can say that this is the hardest part of the book to understand.
In the fifth part of the book, which deals with robots and insects, the evolution of robots from the past to the present is explained by strengthening examples.
The sixth chapter was one of the chapters that I enjoyed the most. Because it enabled me to think in different ways about the philosophical questions that I & most people think about AI. I knew about the Turing Test that it was passed by an artificial intelligence thought to be Eugene Goostman, but in this episode, Ian McDonald’s quote also came to mind: “Any artificial intelligence smart enough to pass a Turing Test is smart enough to know to fail it.”
This chapter also guided me to be able to think in a field (artificial intelligence and morality) that I had never thought about artificial intelligence.
The last part of the book explains the singularity, which means that if artificial intelligence reaches the human level, it can replicate and develop itself, so it can become superior to us. There are actually two different views in this part, the first of which is that when artificial intelligence reaches this level, war, hunger, disease, etc. It argues that our problems will end and this will be for our benefit. However, the second argues that ignoring the threat of artificial intelligence would be our biggest mistake, as scientists like Stephen Hawking have said. I think, in addition to thinking like S-sceptics AI is less promising than many people assume, it needs more and more researches but it is also important to think that it could be a threat.
At the same time, I think the thing to be aware of is the people who can manage artificial intelligence and use it for malicious purposes, rather than the threat of artificial intelligence to us. So, although we may be a little far from being hostile to us by artificial intelligence, it is clear that humanity has no enemy other than humans at the moment.