A Holiday Reading List to Satisfy Your Inner Nerd
It’s the end of the year and time for a well-deserved break. In addition to spending time with family and friends, if you’re like us, you’re probably looking forward to some down time and the chance to cuddle up with something interesting to read. It’s with that in mind that we wanted to share some our favorite picks. From science fiction to cutting-edge academic research, we’ve compiled a reading list to challenge and delight you. In the process, you’ll hopefully also expand your perspective on some of the more fascinating aspects of machine learning and artificial intelligence.
Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy
In “Weapons of Math Destruction,” The New York Times best-selling author Cathy O’Neill introduces readers to the dark side of big data. She explains that while the pervasiveness of algorithms in our daily lives should bring greater fairness to society, the opposite is actually true. All too often, algorithms reinforce discrimination with potentially disastrous consequences. Well-written and thoughtfully argued, “Weapons of Math Destruction” should be required reading for anyone interested in bias, transparency, and trust.
Automating Inequality: How High-Tech Tools Profile, Police, and Punish the Poor
For even more on the ethical risks of algorithms, check out “Automating Inequality” by Virginia Eubanks. In it, she looks at the role of data mining, policy algorithms, and predictive risk models on underprivileged and working-class people in the United States. What we like about this book is that she does a wonderful job showing that the issue isn’t just with the algorithms themselves. Rather, it starts with the data we collect and the fundamental constructs of the problems we’re trying to solve.
The Three-Body Problem
If you’re a science fiction fan, you’ll love “The Three-Body Problem” by Chinese author Liu Cixin. This trilogy tells the story of a secret military project designed to establish contact with extraterrestrial life that inadvertently sets plans in motion for an alien invasion. We particularly love the human motherboard scene in the first book because it’s a wonderful example of using creative metaphors to make abstract technology concepts visceral and easy to understand. At the same time, it shows the power of having a massive number of people focus on a problem.
Life 3.0: Being Human in the Age of Artificial Intelligence
The impacts of artificial intelligence are as wide-reaching as they are transformational. In “Life 3.0,” MIT physicist Max Tegmark delivers a thoughtful look at what each of us can do to position ourselves to benefit from AI while avoiding the potential risks associated with it. This is a great book because it not only incites deep reflection on some of the most critical questions of our time, but is also written in a way that’s accessible for all. Buy, read it, and then share it with your friends and family.
Variational Autoencoders for Collaborative Filtering
In terms of research papers, we like “Variational Autoencoders for Collaborative Filtering” from Dawen Liang and Tony Jebara (Netflix), Rahul Krishnan (MIT) and Matthew Hoffman (Google AI). Recommendation systems are par for the course in consumer enterprises looking to solve next best action problems with customers. Here at integrate, we’re big users of generative models and we found the approach outlined in this paper gave us greater flexibility in our recommenders.
Neural Ordinary Differential Equations
Last, but certainly not least, we love “Neural Ordinary Differential Equations” from the University of Toronto and the Vector Institute’s Ricky T. Q. Chen, Yulia Rubanova, Jesse Bettencourt, and David Duvenaud. We work on time series data at integrate, and it’s exciting to consider how this new architecture could allow us to change how we sample discrete windows of past behavior to make even more accurate predictions.
Pressed for Time?
Don’t worry, we’ve got something for you too. In this fireside chat from ICLR 2018, Daphne Koller talks about her work in machine learning and probabilistic graphical models. Our team loved this and keeps returning back to it for inspiration as we think about the impact that we want to have in our personal vision exercises. Plus, it’s only 36 minutes long making it easy to squeeze in to even a jam-packed holiday season.
