The Deceptive Coronavirus Has Captivated Us in Our Homes or Cottages. The Silver Lining ...

The deceptive coronavirus has captivated us in our homes or cottages. The silver lining is that many people have more time to read. What books should I recommend to those interested in artificial intelligence? Today, I will introduce you to three books that I am currently reading during this period. If you have read them too or are planning to, I would appreciate your comments.

1. Life 3.0 – I listened to it as an audiobook a few years ago, as a birthday gift from the Alfas. The book was so brilliant that I also bought the Czech translation, which was published in 2020 by Argo / Dokořán. The book by Swedish-American cosmologist Max Tegmark playfully explores the impact of AI on human civilisation. According to the author, this new form of artificial life will gain the ability to improve itself (perhaps it all started with my digital Descartes:)). The author contemplates various scenarios, as well as ethical and philosophical implications. Like me, he questions how consciousness can even be defined.

2. Superpowers in Artificial Intelligence – China, Silicon Valley, and the New World Order. I also read this years ago in English, and this year it was released in Czech. The book reflects on the idea that in AI, significant innovations and brilliant AI experts are no longer as decisive as access to vast amounts of data. Lee Kai-Fu sees essentially only two major players in the game: the USA and China. At the end, the author predicts that the biggest challenge of the future will not be the takeover of human jobs by artificial intelligence, but the question of what it actually means to be human. Personally, I was captivated by the author's experiences from both American and Chinese startups.

3. Deep Learning Approaches to Text Production – released at the end of March this year, currently only in English. This book may not appeal to as wide an audience as the first two, but it will be appreciated all the more by experts in NLP. It was only thanks to this book that I understood why the over-trained xSum architecture is so much better than CNN, yet does not perform well on longer summaries. I look forward to it providing me with a quick overview of modern approaches to paraphrasing, sentence merging, and also text simplification (which two customers have already asked me about). The book also addresses other issues I am currently dealing with, such as how to take context into account when generating text. I am particularly excited about the chapter "REINFORCEMENT LEARNING APPLICATIONS," as I have not seen much on this topic in the context of text generation. Unfortunately, the complexity here is rather high, making it suitable primarily for postgraduate students or established researchers.

Sources: • https://argo.cz/?post_type=book&p=131976
• https://www.knizniklub.cz/knihy/447846-supervelmoci-v-oblasti-umele-inteligence-cina-silicon-valley-a-novy-svetovy-poradek.html
• https://books.google.cz/books/about/Deep_Learning_Approaches_to_Text_Product.html?id=JefYDwAAQBAJ&source=kp_cover&redir_esc=y

Originally published on Facebook — link to post