This week, we give you a first glimpse of our mood-aware music API ‘Melody’, think about the ethic settings for connected cars and purchasing a safety premium, look forward to the sci-fi movie Automata and cover predictive algorithms for health and location detection based on car speed.
This Week in Context
Your Weekly Update on All Things Context, August 22 2014
We’re ready to show the results of an algorithm we’ve developed that knows how a song will be emotionally perceived. Find out all about this research and our upcoming Melody API in Vincent’s post ‘Mood for Music: Emotion Recognition on Acoustic Features‘.
The emotional response to a music fragment depends on a large set of external factors, such as gender, age, culture, and context. However, these external variables set aside, humans beings consistently categorise songs as being happy, sad, enthusiastic or relaxed.
Featured: Ethic Settings for Connected Cars
We’ve previously covered the subject of ethical considerations for intelligent things, and the possible ways cars could be programmed to make that though decision, as any accident that requires choosing one victim over another is fraught with ethical concerns and liability. Inspired by ‘If death by autonomous car is unavoidable, who should die?’, Wired’s Patrick Lin published a great piece on adjustable ethic settings, and how this still would not clear car manufacturers – or programmers – from liability.
A steady increase in SUVs on the Belgian roads, means one car in five is now SUV-sized. This trend is mostly driven – and promoted – by the idea of purchasing ‘safety’ at additional monetary cost. This, in combination with the results of the Tunnel Problem, means we may safely assume people will be gladly paying more for car that prioritises their own survival?
“The way this would work is one customer may set the car (which he paid for) to jealously value his life over all others; another user may prefer that the car values all lives the same and minimizes harm overall; yet another may want to minimize legal liability and costs for herself; and other settings are possible.”
Automata: How Hollywood deals with the possible rise of unfriendly singletons?
Whilst HER was a beautiful story about an advanced (and loveable) humanlike UI for a self-aware and empathic AI agent, Automata will have robots becoming sentient and kidnapping a human. The sci-fi film, set to be released October 10th, paints us a dark future where robots are quickly gaining consciousness and things go bad when a small number of them actually become sentient. (Hollywood rewriting Superintelligence, the orthogonality thesis, and the catastrophic outcome due to lack of pre-planning?)
Machine predicts heart attacks 4 hours before doctors
Using data from 133,000 patients, and parameters such as vital signs, age, blood glucose and platelet counts, researched trained an algorithm that is able predict Code Blue events, such as cardiac or respiratory arrest. The algorithm guessed correctly about two-thirds of the time, while doctors using a scorecard flagged just 30 percent.
Further on data, technology and health, you want to check out these two free publications by O’Reilly: ‘The Information Technology Fix for Health‘ (Andy Oram) and ‘How Data Science is Transforming Healthcare’.
Darkly: Protecting User Privacy from Perceptual Applications
Darkly is an interest concept towards a new field of systems research: privacy-preserving perceptual computing. Darkly is a multi-layered, domain-specific privacy protection system that leverages the structure of perceptual software to insert protection at the platform level APIs, where Darkly intercepts the app’s sensor accesses and inserts multiple privacy protection layers.
Speed tracking and privacy
Drivers who avoid jackrabbit starts and sudden stops are typically lower-risk drivers, and insurance companies benefit by rewarding such behavior. So some companies are offering lower premiums to customers who install a device that constantly measures, records, and reports their speed. However, using ‘elastic pathing‘, it is possible to figure out the route and destination based on starting point and speed, Futurity writes.
An international research team transmitted words from one person’s brain to another by mapping electrical currents in the brain and the spine. Newsweek explains: Researchers used EEG to translate greetings, such as hola and ciao, into code. This was sent from a participant in Thiruvananthapuram, India to Strasbourg, France. There, a computer interface translated the message from code to words and implanted them into the receiver’s brain through light electrical stimulation. Participants didn’t report feeling anything in the process, and only saw flickers of light in their peripheral vision – but they did hear the message.
Papers, Talks & Research
- A Robust Daily Human Activity Recognition and Prediction System (activity recognition, time-sequence, prediction, abstract)
- Gyrophone: Recognising Speech from Gyroscope Signals (microphone, smartphone, sensors, paper and project page)
- A Scanner Darkly: Protecting Privacy From Perceptual Applications (privacy framework, context-awareness, ubiquitous computing, paper)
- Conscious Brain-to-Brain Communication in Humans Using Non-Invasive Technologies (telecommunications, robots, paper)
- Large Scale Visual Recognition Challenge 2014: The Results , and a summary on the NY Times’ Bits Blog.
- Elastic Pathing: Your Speed is Enough to Track You (privacy, destination prediction, automotive, paper)
Enjoy the weekend, see you next week!
(Well, if you subscribe to the Week in Context here.)