Michael Nielsen recognizes that Open Access is often argued about in the abstract. To help the discussion move from the conceptual to the concrete, he recently decided to openly share his experience of writing an open-access book, “Neural Networks and Deep Learning” http://neuralnetworksanddeeplearning.com/chap1.html to illustrate the positive impact and far reach of online publishing.
Nielsen, an Australian quantum physicist living in the San Francisco Bay Area, chose to tell his story in a novel way – posting it as series of tweets in mid August. Although he wrote that being transparent about the details of who read his book and where violated “conventional modesty norms,” he thought it was important to do so.
“A friend suggested it’d be helpful for open science if more facts of this type were widely known,” Nielsen explained. “Putting myself in the shoes of a potential author, I could see that hearing about the impact of making materials freely available could make a difference to my publishing decisions.”
Nielsen says he wrote the book for himself and a small number of targeted readers who he knew well. His hope was it that would reach both people with just a little technical background who might enjoy an accessible introduction to the topic, and readers with a deep background such as professional mathematicians and scientists.
By making the information available freely online, he found he reached a much larger audience than that.
“Neural Networks and Deep Learning” was released chapter-by-chapter (as he wrote it) from November 2013 until completion in July 2015, and has received 14 million page views from 3.5 million users in 231 countries. The fact that each chapter was a distinct marketing event was a real benefit, says Nielsen.
On the increased reach of the work, Nielsen tweeted:
A more informative number: total time spent by readers is about 250,000 hours, or roughly 125 full time working years.
An editor at a major academic press tells [Nielsen] great sales figures for a similar technical textbook in a “hot” field are typically about 5,000-10,000 a year. So open access has a factor 200x or more here.
The geographic reach was astounding. The top cities where it was downloaded: Bengaluru, Beijing, London, New York, Shanghai, Chennai, Hyderabad, San Francisco and Hong Kong. The top countries: United States, India, China, Germany, United Kingdom, Canada, Japan, France, South Korea and Russia.
Nielsen says he was also struck that his book was read in St. Barthelemy, Palau, Kiribati, Anguilla, Montserrat, Comoros, and Mayotte. He anticipates few of those countries would see the book with conventional closed-access publishing. Dozens of translations of the book been begun and a few include Japanese, Chinese, Russian, German, Portuguese, and Lithuanian.
“This started as a hobby project,” says Nielsen. “It makes me very happy – and gives me goose bumps – to know that people in countries I’d never previously heard of have used the book.”
On revenue and credit, Nielsen tweeted:
Would any of this have been possible closed access? Of course some of it would have. I might have made more money. But on nearly every other metric, I suspect being open access was a 100x or more multiplier on the impact.
An unexpected personal benefit has been a steady flow of PayPal donations through my inbox, often accompanied by thank you notes. You better believe I read and appreciate them.
When came to people who wanted his book cited, he suggested a simple format: Michael A. Nielsen, “Neural Networks and Deep Learning,” Determination Press, 2015. He didn’t give it much thought, and didn’t even bother looking at citations for over a year, but was surprised to see that people followed his suggestion and that the book received 786 citations between 2014 and 2018.
Nielsen came up with the name “Determination Press” as a joke, but found that it has now taken on a life of its own. “People occasionally contact me and tell me that some editor needs to know where Determination Press is located. So far I’ve resisted the temptation to say ‘Everywhere…”, and demurely reply “San Francisco’” says Nielsen.
In his Twitter thread, Nielsen sums up his experience this way:
Open access makes material freely available to people who would otherwise never even hear about it. This amplifying effect is not small, it is enormous. And it applies in parts of the world woefully underserved by the existing publication system.
For a general audience interested in knowing more about the book, Nielsen recommends short, stand alone essays on whether there is a simple algorithm for intelligence within the books found at: http://neuralnetworksanddeeplearning.com/sai.html
Next project for Nielsen? He’s writing an interactive book about augmenting human cognition. The second chapter is out at http://augmentingcognition.com/ltm.html
