Links of the week


Using Machine Learning to Explore Neural Network Architecture – Google
Designing Neural Network Architectures using Reinforcement Learning – MIT
How neural networks can generate successful offsprings and alleviate the burden from human designers using reinforcement learning.

Data as Agriculture’s New Currency: The Farmer’s Perspective – AgFunder News
A classification of three types of agricultural data and how they related to the farmer’s needs.

The AI Cargo Cult: The Myth of a Superhuman AI – Kevin Kelly
The founding executive editor of Wired explains why he believes superhuman AI is very unlikely. Instead, we already see many form of extra-human new species of intelligence.

Everything that Works Works Because it’s Bayesian: Why Deep Nets Generalize? – inFERENCe
Finally, Bayesian can also say that they can explain why Deep Learning works! Jokes apart, this article overviews several recent useful interpretations of Deep Learning from a Bayesian perspective.

AI: Its nature and future

AI: Its nature and future is a little book by Margaret Boden, research professor of Cognitive Science at the University of Sussex. It is a quick (too quick?) overview about the history of artificial intelligence (AI) from the first symbolic reasoning systems to the more recent recursive deep neural networks. Boden discusses philosophical and social implications of AI advances and also delves into the hotly debated singularity idea. Boden is a self-declared Singularity-skeptic, but that doesn’t prevent her from acknowledging the threats that AI could pose to society in the near future. I wish she had gone deeper into her arguments, to better motivate her position and offer a clearer understanding of the topics covered.

Books of the month

  • Science: A History in 100 Experiments – John and Mary Gribbin
    A compelling brief history of the key scientific experiments that helped science progress: from Archimedes’s Eureka to the detection of gravitational waves, passing through x-rays, DNA, nuclear energy and much more.
  • Predictably Irrational – Dan Ariely
    Amusing anecdotes and deep observations about why human behavior, especially for economic decisions, is far from being rational. The good side is that our irrational choices are predictable and systematic, allowing us to develop better decision strategies and avoid recurring mistakes.

Artificial Intelligence links of the week

2016 Books List

These are the books I read in 2016. Some of them are truly filled with insights and wisdom.

What is your list?

  1. Murakami – 1Q84
  2. Sam Harris – Waking Up
  3. Sam Harris – Free Will
  4. Winifred Gallagher – Rapt
  5. Neal Stephenson – Seveneves
  6. Roger Peng and Elizabeth Matsui – The Art of Data Science
  7. Yuval Harari – Homo Deus
  8. Cal Newport – Deep Work
  9. D. and D. Meadows, Randers – Limits to Growth: the 30 years update
  10. Italo Calvino – Le Cosmicomiche
  11. Nick Bostrom – Superintelligence
  12. Tanpinar – L’istituto della regolazione dell’orologio
  13. Italo Calvino – Mr. Palomar
  14. Yuval Harari – Sapiens
  15. Daniel Kahneman – Thinking, fast and slow
  16. Nicholas Nassim Taleb – Antifragile

The Best Technology Writing 2009

What’s happening right now in the technology world? What are the social implications of the digital tools now available to everyone with an internet connection? How is our thinking mind changing due to the constant use of the web? Why do people blog? How can people find the time to write Wikipedia? Who are those annoying groups that are disrupting your second life?

These are only some of the questions that are tackled by the inspiring collection of essays edited by Steven Johnson in The Best Technology Writing, 2009. As he emphasizes in the introduction, there has been a shift in technology writing in the last years. The focus is more and more the analysis of the current trends and phenomena and less and less speculations of how the future might surprise us. The digital world has become so complex and varied that it requires the most effort to be explicated. And the future? The future is now and, at the same time, we understand that it is increasingly difficult to anticipate what it will unveil. Especially if we consider that today’s most advanced and popular trends were utterly unpredictable just some years ago. Think about the Facebook phenomenon. Who could have foreseen that it would have grown to contain more people that the United States? Clive Thompson studies the new social media in his “Brave New World of Digital Intimacy”, introducing us to the concept of ambient awareness.

Like the status updates of Twitter and Facebook, the essays in the book provide us a glimpse of the present (what is happening right now?) technological realm. But how is this constant presentness and exposure to an interminable flow of information affecting our thinking? Nicholas Carr tries to answer in his “Is Google Making Us Stupid?”.

The closing essay is by Clay Shirky and well explains where people find the time to spend online socializing, uploading videos to YouTube or writing new articles in Wikipedia. There is a cognitive surplus that is slowly sobering up from decades of TV drunkenness. The Wikipedia effort consists of approximately 100 million hours of human thought. Every years in the US alone, people spend 200 billion hours in front of the TV screen. That’s the equivalent of 2,000 Wikipedias every year! If only 1 percent of that time is carved out for producing and sharing, we could have other 100 Wikipedia-like projects every year!

Stop watching TV. Read The Best Technology Writing, 2009 and contribute to the expanding knowledge world we are creating.

P.S. If you want to know why I blog, read Andrew Sullivan‘s excellent essay “Why I Blog?”, also contained in the book.


pretesco /pre’tesko/ (preƒteƒsco) agg. $ [av. 1556; der. di prete con –esco] stereotipo, spregiativo, considerato tipico o proprio dei preti in quanto ipocrita, untuoso e insincero: ipocrisia pretesca, atteggiamento p.

trovato nel romanzo “Le Braci” di Sandor Marai