Four short links: 10 August 2016
Language Research, Probabilistic Cognition, Medical Data, and Teaching State Computer Security
- AI’s Language Problem (Wired) — very readable overview of the challenges and research in making more sense of the text we are processing, and why deep learning is a great start but not the end of the work to be done.
- Probabilistic Models of Cognition — web book from Stanford profs behind webppl, a probabilistic programming language built on Javascript.
- Your Medical Data Misappropriated (BoingBoing) — “Property” is a terrible framework for understanding personal information—it’s led to a situation where people aren’t allowed to know what’s going on in their own bodies, and where corporations can use anti-theft laws to attack scientists, security researchers, and the people whose bodies generated the data the corporations have turned into crown jewels.
- PwC Australia’s Game to Teach Cyber Security Lessons (Computer Weekly) — a friend has played this, says it is really good at conveying the complexity and the relentlessness (and is accessible for normal people).
