This document discusses the history and future of context-aware technologies. It begins with a timeline of early context-aware tools and research projects. It then discusses the value of context awareness in mobile devices, with goals of adaptively delivering personalized information without increasing cognitive load. Key aspects are high-speed hybrid inputs, adaptive analytics, and distributed display methods. The document presents biological models like the brain's default mode network as an example. It envisions a future where phones act as hybrid sensors to underpin augmented reality, ubiquitous computing, and hyper-personalized services through constant context gathering and adaptive delivery of information.
4. Alfred North Whitehead-(1861-1947)First context-aware tools theorist “Civilization advances by extending the number of important operations which we can perform without thinking of them.” ------ …Context-aware services shouldn’t spike cognitive load.
5. Context Awareness Timeline Efficient power usage Context-aware tools begin todisappear. CALO project Commercial aviation HUD 1st car HUD:Cutlass Supreme 1st Context Awareness paper Rear-view mirrors Sailboat tell-tales US DoD air-combat testing “Augmentedenvironments,” sensor nets Startups & Acquisitions Secretary invented Siri founded Now 2013 ~1900 1970s 1974-8 1988 1992+ 2003 2007 1994 ? Prehistory, soon after invention of the sail. Horse carriages Schilit, Theimer, Want, & Adams Tom Caudell &David Mizell “UbiComp”:Mark Weiser Sutherland/Sproull, ‘68: 1st HMDs Hands-up displays
6. The Value of Now Contextual data Constantscenariosupport = + A.I.
7. Reasons for Context Aware Mobile 1. Diversity of use cases / user. 2. User attention = scarce resource. Tool should learn the user. Context awareness manages datadelivery:What,when, how much,what modality, what priority. ------------------------------------------------------- Especially when not-to.
8. Goal of Context Aware Mobile Adaptively deliver info, hyper-personalized via n dimensions of user + environment states, with high space/time precision, without spiking your cognitive load. Context + Personal Salience = Value
11. The most important rule we can import from biology:Nature is Always On.The 2nd most important rule:Hybrid sensing & hybrid response.
12. The hardware emulation: Your phone is the hybrid Master Sensor for your life; everyone else’s phones help, all the time. Together, they’re the platform for: ) Augmented Reality Pervasive computing Context awareness Semantic Web Ubicomp&sensornets Ambient Intelligence Things that Think Internet of Things & Hyper-personalized services… )
13. Human Context-Aware Model: The Brain’s Default Mode Network • Active only when the brain is notfocused on a particular task • 2 linked brain areas: Empathy re: intent of others,-+ your own state awareness Personal memories + visualization of future scenarios Scenario modeling State & intent awareness Brain photo: Omikron/photoresearchers Network overlay: Olaf Sporns/Indiana Univ. (modified by J. Korenblat)
14. “It’s all about long-term, sustaining relationships.” -Alan Kay (2009), on the trend toward a service-based model. Context Aware Services: long-term, sustaining,contextuallydynamic, hyperpersonalized-relationships between users & info services.
15. Context Gathering • Constant autodiscovery of data feeds & sensor feeds. • Constant 2D & 3D feature I.D. and object recognition. • Constant motion analysis + evaluation linked to object recognition. • Constant indoor mapping, 2D & 3D: Ultrasonic& IR-sensing for position, spatialanalysis & modeling. • Integrated gesture & speech recognition, plus voice ID. • Constant audio + visual logging with visual & semantic tags.
16. Context-Driven Use Case Entertainment Any licensed character: a useful buddy, a delivery method on top of context-aware, locational, hyper-personalized services. The character behavior, persona, & animation: Character Skin.
18. The Ad Hoc Value of Now Ad hoc P2P mobile sensor networks, constantly re-forming and informing each other. In each locale: • Pooling realtime data • Stored only in the swarm • Hand-off & erase upon exiting • Short-distance wireless • No use of cloud or carriers • Data tied to location • Every phone • Every moment Duration of data retention on any device: A function of # of devices transiting the defined area.
26. Context-driven Use Case: Entertainment First rule of theme parks: You bring the fun in with you - your friends-& family. • The attractions you pay for aren’t the experience. • Now, a new type of friend. For as long as you want. A licensed character becomes a useful buddy, a delivery method for context-aware, locational, hyper-personalized services. c. Lynne LaCascia, all rights reserved
27. Context-Aware Biological Model: The Brain’s Default Mode Network Functions of consciousness: 1) Constantly evaluate surroundings (physical + human) 2) Constantly judge salience of each aspect of current situation in context of your needs 3) Constantly construct scenarios for next step Parallel function of context-aware devices: Assess, select, shape, & deliver salient data (Device awareness augments user’s awareness) History+current state= basis of decision for next state
41. Schilit, Bill.; Adams, Norman; Want, Roy. Context-Aware Computing Applications. 1st International Workshop on Mobile Computing Systems and Applications. (1994) 85-90
For good context aware services, using our phone to both gather and deliver data, …We need to be sensing, sorting, and prioritizing many dimensions of personal context simultaneously -- physical proximities, time proximities, and all the daily priorities of your life. Then we need predictive ability to estimate what key elements of that data may be so important to you that you’ll want to know, NOW.Finally, we need the awareness of what you are doing, in what sort of physical environment and with whom, to be able to decide how best to deilver you the information. Could be audio, could be tactile, could be an alert followed by looking at a screen. But it’s VERY important that your device be able to deliver that data in the optimum medium at the right momemt to you, in a useful amount of detail, and NOT at the wrong time.To do this, it needs to learn you. A lot about your habits plans, and preferences, in a lot of layers. And it’s going to be another big, profitable facet of the disruptive impact of our mobile digital tools. Context aware tools need broadband highly parallel multimodal data sensing Fast powerful analyticsAnd hybrid means of delivering their conclusions to you, methods that constantly adapt to what’s best for you in the moment..
Fortunately we have models…for broadband data gathering, alertness, for anomaly identification and threat analysis, and for processing that information.Nature has been developing ways to simultaneously attend to many dimensions of one’s immediate environment, discard the irrelevant data, prioritize what’s left, and decide what to do NEXT, very quickly. We already do it ourselves. It’s called the Default Mode Network of the brain. Whenever we’re not on a task, it’s constantly evaluating our current situation, compariing it to all the history we’ve ever had, and deciding what possible things might be best to do next. We just need to find ways to do that with hardware. That’ll help a lot.
Many apps that haven’t even been developed yet will have to move into the OS, always functioning.Much more sensing must be going on, all the time. Your phone needs to know the shape of the room you’re in, how many light sources and people there are in it, who the people are, what trace elements are in the atmosphere, and how that changes over time, and Compared to the biological models, our tools are at a disadvantage; our emotions, for example, are adept at automatically finding important patterns based on our experience, and driving our reactions before we even know what’s happening, but tools need to calculate and approximate that --- very difficult. Then they need to be able to respond and notify us in ways that aren’t irritating, and that definition changes frequently every day. Big task, huge payoff.