Book Info:

Description
Title Selected Papers on Computer Science [Amazon]
Author Donald Knuth
Pages 276

This is the most accessible book from Don Knuth. Although I was published nearly 20 years ago, it still is a classic computer science book. In Amazon, there is an interesting comment about the book from Peter Norvig (Director of Google Research). The major topic in the book is the origination of computer science in the period of which the discipline is a new thing. Through chapters, we can see there were many debates and struggles among scientists about whether if computer science is truely a “science” and not a branch of mathematics [Chapter 1, 2, 3].

This makes me think about the current state of Deep Learning. Maybe 5 or 10 years later, Deep Learning will become a separate discipline as Computer Science segragated from mathematics several decades ago.

There are especially interesting chapters in the book which I can describe as below:

History of Computer Science and Deep Learning

While reading the book, I had the feeling that the period of 1950-1975 perhaps exploded into computer science research, which is pretty much similar to Deep Learning nowadays. That was the time we there are few universities opened Computer Science department and people still debated about the name of this science that whether its name is “Computer Science”, “Information Technology”, or “Information Processing”. That is also the time when people thought that the problem which can only solve on $\mathcal{0}(N^2)$ actually can be solved in $\matchcal{O}(N \lg N)$, the time when there is not any mathematical tools for analyzing algorithms.

How about Deep Learning?

Nowadays, people remain skeptical about Deep Learning in many aspects. Somebody compared it to alchemy because of the lack of rigorous analysis and mathematical fundamentals. However, according to the currect development of Deep Learning and avoiding the media hype, I am confidient about the future of Deep Learning. At least, in my opinion, it will become an interdiscipline and completely transform into “chemistry”. In retrospect, perhaps Computer Science was compared to “alchemy” in the period of 1940-50 due to its lack of rigor.

Study about History of Computer Science

I was fascinated by the analysises in detail about “ancient” algorithms as well as the elaborate way to discuss the draft of merge sort from John von Neumann. IMO, this really is the way people should study about the history of Computer Science. It is not about memorizing who and when algorithms were created, it is about the motivation and approach which the inventor tackled the problem as well as analyzing the methods in the circumstances which the technology is limited. Through these insights, we will respect the contribution from algorithms and their authors. In addition, we can gain more methods, approached for another problems.

Don Knuth is especially interested in this particular subject, there are many videos and documents in which he discussed exhaustively: