There’s only so many times you can watch Elf or Love Actually over the Holidays before it drives you mad. Staying ahead of the trends for 2019 is on everyone’s minds, so we reached out to a few AI experts who have shared their predictions for the coming year which you can read here. If you’re keen to keep reading over the Christmas season we’ve also spoken to our AI and Deep Learning community who have recommended books, journals, podcasts and videos that they’re currently enjoying.

Learning to Reinforcement Learn

Recommended by Jeff Clune, Uber AI Labs

Jeff first joined us at the Deep Learning for Robotics Summit in San Francisco earlier this year and will be returning to the World’s Biggest Deep Learning Summit in San Francisco this January 24 - 25. Jeff’ presentation ‘Go-Explore: A New Type of Algorithm for Hard-exploration Problems’ will present Go-Explore, a new algorithm for such ‘hard exploration problems.’

Learning to Reinforcement Learn
In recent years deep reinforcement learning (RL) systems have attained superhuman performance in a number of challenging task domains. However, a major limitation of such applications is their demand for massive amounts of training data. A critical present objective is thus to develop deep RL methods that can adapt rapidly to new tasks. In the present work we introduce a novel approach to this challenge, which we refer to as deep meta-reinforcement learning. Previous work has shown that recurrent networks can support meta-learning in a fully supervised context. We extend this approach to the RL setting. What emerges is a system that is trained using one RL algorithm, but whose recurrent dynamics implement a second, quite separate RL procedure. This second, learned RL algorithm can differ from the original one in arbitrary ways. Importantly, because it is learned, it is configured to exploit structure in the training domain. We unpack these points in a series of seven proof-of-concept experiments, each of which examines a key aspect of deep meta-RL. We consider prospects for extending and scaling up the approach, and also point out some potentially important implications for neuroscience.

Other suggestions from Jeff:

Blog posts from the team at Uber:

Additional papers with some of the most exciting developments in ML:

Architects of Intelligence: The truth about AI from the people building it, Martin Ford

Recommended by Fiona McEvoy,

As well as joining the ethics panel discussion at the Deep Learning for Robotics Summit in June earlier this year, Fiona also appeared as a guest on RE•WORK’s Women in AI Podcast discussing the importance of collaboration in AI for ethically sound design.

Architects of Intelligence: The truth about AI from the people building it
How will AI evolve and what major innovations are on the horizon? What will its impact be on the job market, economy, and society? What is the path toward human-level machine intelligence? What should we be concerned about as artificial intelligence advances? Architects of Intelligence contains a series of in-depth, one-to-one interviews where New York Times bestselling author, Martin Ford, uncovers the truth behind these questions from some of the brightest minds in the Artificial Intelligence community.

Other suggestions from Fiona:

AI Super-Powers: China, Silicon Valley and the New World Order - Kai-Fu Lee

Deep Learning with Python, Francois Chollet

Recommended bu Hao Yi Ong, Lyft

“You've probably heard of it but I have been recommending Francois Chollet's Deep Learning with Python to colleagues who want to get started on the topic. You'll be able to go through it over a weekend.” Deep Learning with R introduces the world of deep learning using the powerful Keras library and its R language interface. Initially written for Python as Deep Learning with Python by Keras creator and Google AI researcher François Chollet and adapted for R by RStudio founder J. J. Allaire, this book builds your understanding of deep learning through intuitive explanations and practical examples. You'll practice your new skills with R-based applications in computer vision, natural-language processing, and generative models.

Other Suggestions from Hao Yi:
“Also try Christoph Molnar's Interpretable Machine Learning. He has very kindly made it available for free online on his website. Most folks coming out of school will only have an inkling of why interpretability would be important at work. But most folks in industry would already be working hard at tackling this practical challenge with their product, ops, and business colleagues.”

Foresight Report

Recommended by Will Millership, Good AI

Will is currently working on a case study for RE•WORK’s upcoming white paper, AI for Social Good. Their mission is to develop general artificial intelligence - as fast as possible - to help humanity and understand the universe.

Foresight Report
Foresight Institute is a non-profit organization to steer the beneficial development of technologies of fundamental importance for the future of life, focused on AI, cybersecurity, and molecular machine nanotechnology. Foresight selectively advances beneficial technologies via technical workshops, the Feynman Prizes and the Foresight Fellowships, and avoids the risks of technologies via strategy meetings and policy recommendations. The annual Foresight Institute AGI strategy meeting gathers representatives of AI safety organizations and academic institutions with experts in fields relevant to AGI strategy, including security, government policy, and international political economy. The 2017 Foresight Institute AGI strategy meeting on AGI Timeframes & Policy focused on AI timelines, with special consideration given to policy, cybersecurity, and coordination. A 72% majority of workshop survey respondents voted for the 2018 AGI strategy meeting to focus on avenues for coordination on the path toward AGI, especially in relation to the world’s greatest geopolitical powers.

Content from RE•WORK:

Women in AI Podcast

Women in AI is a weekly podcast from RE•WORK, meeting with leading female minds in AI, Deep Learning and Machine Learning. We speak to CEOs, CTOs, Data Scientists, Engineers, Researchers and Industry Professionals to learn about their cutting-edge work and advances, as well as their impact on AI and their place in the industry. Most recently we were joined by Laurence Perreault Levasseur from Flatiron Institute who spoke about deep learning in cosmology. Additional guests include Sara Hooker, Google Brain; Sarah Culkin, NHS; Raia Hadsell, DeepMind and many more.

Interviews and Presentations from RE•WORK Summits

We have just updated the RE•WORK YouTube channel with recent behind the scenes interviews and presentations from our most recent summits. Watch interviews with leading experts from Google Brain, DeepMind, Uber AI Labs, OpenAI, MIT-IBM Watson AI Lab, Royal Mail and many more.

Additional Recommendations:

The Master Algorithm: How the Quest for the Ultimate Learning Machine Will Remake Our World, Pedro Domingos

Recommended by Craig Metcalfe, Brisbane City Council

Artificial Intelligence, the Modern Approach, Stuart Russell

Recommended by Tarlan Mammadov, Shell

Your Best Year Ever, Michael Hyatt

Recommended by Eunice Chendjou, DataGig