Weekly Review 1 April 2016

Some interesting links that I Tweeted about in the last week:

1. Valuing the AI market for 2016 http://techemergence.com/valuing-the-artificial-intelligence-market-2016-and-beyond/?utm_source=facebook&utm_medium=paid-promoted-post&utm_term=ai-market-size&utm_content=180last&utm_campaign=blog
2. Using machine learning to improve automatic speech recognition: http://spectrum.ieee.org/tech-talk/computing/software/machines-just-got-better-at-lip-reading
3. Resistive Processing Units to accelerate training in deep learning neural networks: http://www.tomshardware.com/news/ibm-chip-30000x-ai-speedup,31484.html
4. Paper on Resistive Processing Units for deep learning: http://arxiv.org/abs/1603.07341
5. Computers don't cause a net decrease in job numbers, but do increase inequality, with the lowest-paid hit hardest: https://hbr.org/2016/03/computers-dont-kill-jobs-but-do-increase-inequality
6. What I like to call "avoiding work by doing work" - doing small tasks to avoid doing larger tasks: https://www.insidehighered.com/blogs/gradhacker/two-one-deal-killing-boredom-procrastination 
7. UK's Wellcome Trust wants research they fund published in open access journals: http://www.theregister.co.uk/2016/03/26/sick_of_costly_research_journals/ 
8. Robots learning to pick things up using deep learning neural networks: http://spectrum.ieee.org/automaton/robotics/artificial-intelligence/google-large-scale-robotic-grasping-project 
9. This article seems to be arguing that it's better to get "off the shelf" machine learning than to develop your own: http://www.kdnuggets.com/2016/03/dont-buy-machine-learning.html 
10. AlphaGo and the declining advantage of big companies: https://hbr.org/2016/03/alphago-and-the-declining-advantage-of-big-companies?utm_source=twitter&utm_medium=social&utm_campaign=harvardbiz 
11. Lots of companies getting into AI now: http://www.informationweek.com/big-data/big-data-analytics/google-loves-machine-learning-cloudera-acquires-startup-big-data-roundup/d/d-id/1324845
12. AI hits the mainstream: https://www.technologyreview.com/s/600986/ai-hits-the-mainstream/
13. AI is getting big in Silicon Valley: http://www.nytimes.com/2016/03/28/technology/silicon-valley-looks-to-artificial-intelligence-for-the-next-big-thing.html?mwrsm=Twitter 
14. Note to post-grads: don't EVER use graphs like these in your dissertation, I will fail you! http://www.buzzfeed.com/katienotopoulos/graphs-that-lied-to-us#.scqWJelqk 
15. Neural network chip could bring convolutional neural networks to mobile devices: http://spectrum.ieee.org/computing/embedded-systems/bringing-big-neural-networks-to-selfdriving-cars-smartphones-and-drones 
16. One step to become a machine learning expert: http://www.kdnuggets.com/2016/03/become-machine-learning-expert-one-simple-step.html 
17. Building models is a skill, and like every other skill it must be practiced to be mastered: http://www.kdnuggets.com/2016/03/become-machine-learning-expert-one-simple-step.html 
18. How to tell if the performance of two classifiers is statistically significantly different: http://www.kdnuggets.com/2016/03/statistical-significance-two-classifiers-performance-difference.html
19. The fortunate failure of Microsoft's Tay: http://www.businessinsider.de/why-microsofts-chatbot-tay-should-make-us-look-at-ourselves?r=US&IR=T&utm_content=buffer919c9&utm_medium=social&utm_source=twitter.com&utm_campaign=buffer 
20. Some people would rather have a computer for a boss than a human: http://motherboard.vice.com/en_au/read/a-third-of-young-canadians-would-prefer-a-robot-boss
21. Is the next step for Google DeepMind playing poker? http://www.theguardian.com/technology/2016/mar/30/deepmind-poker-alphago-computer-casino 
22. Machine learning in signature detection for cybersecurity: http://www.darkreading.com/attacks-breaches/machine-learning-in-security-good-and-bad-news-about-signatures/a/d-id/1324888 
23. Google hypes machine learning to sell its cloud computing platform: http://www.informationweek.com/cloud/infrastructure-as-a-service/google-pumps-up-cloud-platform-with-machine-learning/d/d-id/1324822 
24. I'm sure I read / reviewed a paper about this - density-based unsupervised clustering: http://www.datasciencecentral.com/profiles/blogs/variance-clustering-test-of-hypotheses-and-density-estimation-rev 
25. Fighting China's - and especially Beijing's - smog with machine learning: https://www.technologyreview.com/s/600993/can-machine-learning-help-lift-chinas-smog/
26. Low-power, neuromorphic chips being applied in the US nuclear industry: http://www.computerworld.com/article/3049380/big-data/this-brain-inspired-supercomputer-will-explore-deep-learning-for-the-us-nuclear-program.html 
27. Machine learning in signature detection for cybersecurity part 2: http://www.darkreading.com/attacks-breaches/machine-learning-in-security-seeing-the-nth-dimension-in-signatures-/a/d-id/1324889 
28. How Google plans to solve Artificial General Intelligence: https://www.technologyreview.com/s/601139/how-google-plans-to-solve-artificial-intelligence/ 
29. Avoiding complexity in machine learning: http://www.kdnuggets.com/2016/03/avoiding-complexity-machine-learning-problems.html 
30. Artificial Intelligence still works best when AI is paired with humans: https://www.technologyreview.com/s/600989/man-and-machine/
31. Would the health care app space be a good place to apply machine learning? http://spectrum.ieee.org/the-human-os/biomedical/devices/ahead-of-apple-carekits-debut-physicians-still-skeptical-of-health-apps 
32. How Baidu is using AI, especially deep learning: https://www.technologyreview.com/s/600988/how-ai-is-feeding-chinas-internet-dragon/ 
33. I wonder if this approach could be used to generate real estate listings? They're not that different from clickbait: http://larseidnes.com/2015/10/13/auto-generating-clickbait-with-recurrent-neural-networks/