Featuring our little Hue-powered-by-mood hack, Siri goes neural-net, a tour of machine learning algorithms, emotion-displaying blankets by British Airways, tracking your every cough, and Jaguar’s smarter cars.
This Week in Context
Your Weekly Update on All Things Context, July 11 2014
As I’m writing this, the general mood in our office has evolved from tired to cheerful. We’ve not been feeling intoxicated, aggressive or nervous today. How do I know? The guys put together a neat little hack that ‘ubiquitously listens’ for our mood, and changes the colour of these Philips Hue lights accordingly. Pretty awesome, no?
Now, on to this week’s must-knows for sensors, mood detection & AI:
A neural-net boosted and more talented Siri is on the way, writes Wired. The company is hiring in the artificial intelligence branch known as ‘deep learning’ for their speech recognition team.
Deep learning? What is the difference between supervised learning and unsupervised learning? As it comes to popular machine leaning algorithms, is k-Nearest Neighbour even remotely similar in function and form to Self-Organizing Map or to Decision Stump? Time to start your tour of machine learning algorithms.
An algorithm schedules and manages the nightly engineering work for the Hong Kong subway, and does it more efficiently than any human could. This supports Bringsjord vision (quoted below) and demonstrates the power that niche artificial intelligence can have, opposed to the idea of a human-level intelligence. (via Roel)
“I have no small amount of optimism about what machines that fall short of human intelligence, but leverage brilliant human engineering, can accomplish. In this regard, I think there are very few limits within those constraints.”
– Selmer Bringsjord
quoted in Forget Turing, the Lovelace Test Has a Better Shot at Spotting AI
Mood, Sensing & Wearables
British Airways is tracking their passengers emotions using bluetooth & fiber-optic equipped blankets that respond to electrical fluctuations in the brain measured by a headpiece. The “happiness blankets” turn red when you are relaxed or happy and blue when you are angry, irritable, or anxious.
A prototype wearable, a ‘piezoelectric detector’, picks up sound waves transmitted through the skull, it can detect subtle clues about the activity – eating, coughing, breathing, .. – or emotional state of the person wearing it. The system consists of a microphone that attaches behind the user’s ear, and someday could be built into the frame of a device like Google Glass.
Talking about Glass, wouldn’t it be much better if you just had to think ‘OK Glass’? This Place adjusted NeuroSky’s MindWave Mobile headset to create MindRDR, an open-source project that allows you to command Glass using mere thought. (via Roel)
Smarter Car, Connected Home
Jaguar too is working on self-learning technology that will figure out your habits, schedule and preferences so it can automatically and independently act. The cars could pre-heat the cabin, or send you smarter alerts. More examples on businessinsider.com and in this video. (Hat tip to Niels van Weereld)
(Related in a previous #WIC: Mercedes-Benz Research & Development and their aim to turn your car into the family’s friend & concierge.)
Sentri, a smart home solution that aims to guard everything from your home’s security to its physical state such as humidity levels and air quality, has more than reached its $200,000 Kickstarter home. (via David)
Privacy & iOS 8
With iOS 8 and OS Y Yosemite, Apple introduced many changes to the way it deals with user security and privacy. Luis Abreu has posted a great overview.
‘Detecting’ Your Health
Will your health insurance soon ask you to submit a selfie? ‘Face My Age‘ focusses on facial-recognition and age-progression technology, research that will help with estimating people’s life spans and future health based on a photograph.
With ‘And what if your health plan knows how you shop?‘ the New York Times has a look at which data sets hospitals and insurers are using to make predictions about your health and doctor visits. These predictive analytics get used to timely steer patients to primary care, rather than wait for their emergency room visit, but also to target “scientifically identify well-insured prospects.”
Papers, Talks & Research
- Volvo intelligent news: A context aware multi modal proactive recommender system for in-vehicle use (context-aware, paper)
- Ad Click Prediction: a View from the Trenches (machine learning, paper)
- Online Matching and Ad Allocation (algorithms, paper)
- The 380-millisecond-tracking, decision-making software that reveals what you really want (predictive analytics, article)
- Future progress in Artificial Intelligence: A Poll Among Experts (future, poll)
- Group Affiliation Detection Using Model Divergence for Wearable Devices (sensor data, paper)
- Networks with emotions: an investigation into deep belief nets and emotion recognition (deep learning, paper)
- BodyBeat: A Mobile System for Sensing Non-Speech Body Sounds (mobile sensing, paper)
If you encounter – or have written – an interesting article on emotion & mood detection, using sensor data to make technology respond smarter, or the future of artificial intelligence, let us know!
Enjoy the weekend, see you next week.
If you’d like to get the highlights of what is moving and shaking in our industry delivered to your mailbox once a week, sign up here.