Blog - Jürgen Kanz

Logo Jürgen Kanz
Go to content

Artificial Intelligence: A Guide for Thinking Humans

Jürgen Kanz
Published by in Book Review · 15 November 2019
Tags: AIArtificialIntelligenceDeepLearningCommonSenseThinking
Almost every day we can read about Artificial Intelligence (AI) in media. Many articles describe the potential threat of AI, combined with the question “When will robots take over the world?

Are we already so close to the “singularity” – the moment when AI has the same intelligence level as a human?

When we hear about AI, we have to distinguish to stages.


1. Artificial narrow intelligence (ANI)
ANI is the technology we use with Alexa, Siri, Web search, movie recommendations at Netflix, and so on. We are talking here about help or assistant functions for our daily life.

2. Artificial general intelligence (AGI)
This technology does not exist yet and we do not know when it will come or whether it will come. It is the area of “superintelligence”, an area of “singularity” and beyond.

Therefore, we can discuss the topics of ANI and we can question the status and quality of this technology today.

With a critical eye, we can make assumptions about AGI, we can already now define guidelines what we want to allow by application of this technology level.

Recently, Melanie Mitchell launched her new book “Artificial Intelligence: A Guide for Thinking Humans”. Melanie writes with incomparable clarity and wisdom. She makes clear what are the AI opportunities and limitations today. She makes clear that we should not believe in all media topics about “superintelligence”. We are far away from machines / computers with such a level of intelligence. Why?
Nobody knows how to explain and train "common sense"[1] to a neural network. Deep Learning algorithms do not know what logic is.



Melanie's book is devided into five parts:

Part I. Background
1.  The Roots of Artificial Intelligence
2.  Neural Networks and the Ascent of Machine Learning
3.  AI Spring

Part II. Looking and Seeing
4.  Who, What, When, Where, Why
5.  ConvNets and ImageNet
6.  A Closer Look at Machines That Learn
7.  On Trustworthy and Ethical AI

Part III. Learning to Play
8.  Rewards for Robots
9.  Game On
10.  Beyond Games

Part IV. Artificial Intelligence Meets Natural Language
11.  Words, and the Company They Keep
12.  Translation as Encoding and Decoding
13.  Ask Me Anything

Part V. The Barrier of Meaning
14.  On Understanding
15.  Knowledge, Abstraction, and Analogy in Artificial Intelligence
16.  Questions, Answers, and Speculations

I highly recommend Melanie Mitchell’s book, because it can close knowledge gaps and destroy inappropriate worries.

In case you are not sure, whether this book is something for you. A couple of days ago, Melanie presented insights of her book at the Santa Fe Institute  [2]. Please look, listen and decide if you want to dive deeper into this subject.



Resources:
[1] Jürgen Kanz, "How to become a 'common sense practitioner'?", https://www.juergen-kanz.de/blog/index.php?how-to-become-a--common-sense-practitioner--,  last accessed on November 15, 2019
[2] Melanie Mitchell & Santa Fe Institute, Video "Artificial Intelligence: A Guide for Thinking Humans", https://vimeo.com/372705678, last accessed on November 15, 2019



Back to content