In the News
Facebook: new milestones in AI research
List of a few impressive milestones reached by Facebook's AI teams. This includes object segmentation in scenes, natural language & image understanding, predictive learning and planning.
Apple's secrecy is hurting its AI efforts
Apple is known for its opacity. But what works for the development of a mobile OS may not work for research in AI. Will this secrecy hurt Apple's effort to recruit top-tier scientists and build the next generation of intelligent devices?
Toyota starts $1 billion AI center
Toyota Motor Corp. is spending $1 billion to form a research institute focused on the artificial intelligence and robotics technology it needs to make cars that can overcome driver errors and reduce traffic fatalities.
Computer, respond to this email
Google announced the roll-out of a feature where Inbox will now prepare full responses to incoming emails.
Workplace
Data science interview questions
Long list of questions that may be asked during data science interviews. Mostly technical.
How to figure out the gaps in your Data Science skill set
With your unique mixture of academic and non-academic projects, you will feel like there are gaps in your current background. You've searched around the web to see if you can find some insight into your situation, but so far no recommendations on what to do and what to learn have been personal enough for you.
Learning
Baidu found China’s "ghost cities"
Data analysis helped Baidu find so-called "ghost cities" in China. At this point they seem to be keeping the locations secret though.
Inside view of language technologies at Google
Few organizations are more heavily invested in Natural Language Processing than Google. This interview with Enrique Alfonseca (NLP scientist) sheds some light on his work at Google.
Understanding the Bayesian approach to false discovery rates (using baseball statistics)
Learn to apply a Bayesian approach to a method usually associated with frequentist statistics, namely false discovery rate control.
Software tools & code
Apache Spark Machine Learning scenario with a large input dataset
What if you want to create a machine learning model but your input dataset doesn’t fit in memory? Normally you would use distributed computing tools like Hadoop and Apache Spark. However, Apache Spark is also able to process your data in local machine mode and even build models in such cases.
Texture Synthesis with deep CNNs
Impressive library of textures that were generated from a given source images. The results look great. See paper and code for more details.
Some thoughts
Philosophy will be the key that unlocks artificial intelligence
AI is achievable, but it will take more than computer science and neuroscience to develop machines that think like people
Human values should be programmed into robots, argues a computer scientist
As Artificial Intelligence becomes increasingly sophisticated, there are growing concerns that robots could become a threat. This danger can be avoided, according to Stuart Russell, computer science professor at University of California, Berkeley, if we figure out how to turn human values into a programmable code.
About
This newsletter is a weekly collection of AI news and resources. If you find it worthwhile, please forward to your friends and colleagues, or share on your favorite network!
Share on Twitter · Share on Linkedin · Share on Google+
Suggestions or comments are more than welcome, just reply to this email. Thanks!