Assignment 3 – Tony

These two papers mainly focused on how to design a wearable system with a software agent, a heads-up display, a one-hand chording keyboard, and IR sensors sensing preseted beacons, named Remembrance Assistant (RA). RA is capable of proactively monitoring a user’s contextual activities and pushing just in time information.

RA automatically makes “notes” related to a user’s contexts by recognizing people’s faces whom the user is talking to and by sensing beacons which surround. The user can also type in “notes” with the one-hand chording keyboard. RA exports outputs which generated by the soft agent equipping designed algorithms calculating correlations between notes and contexts and shows them on the heads-up display.

We can say the intentional type-in information, the facial recognition and preseted location beacons are kind of explicit contexts. Rather than that, I am more interested in how to incorporate implicit contextual information, such as environmental atmosphere, personal emotion and physical condition to predict users’ needs. These will make the wearable assistant more embedded as being part of our mind instead of being part of our body.

Assignment03_Jifei

When a bike is broken in the middle of the road, and we happened to have a future gadget like Google Glass with us. We wear the glass and try to fix the bike. the question is, what kind of information the Glass would provide to us?

According to the JITIR system, the Glass would automatically search the potential solutions base on the condition of the environment and the bike. Or it would suggest us different tools to fix it. As a user, what we have to do is just perceiving the information and filter those which are not fit.

This is a pretty typical approach of artificial intelligence. In this approach we assume that human perceptions are passive. our eyes are like a camera, statically sitting there and waiting for something happen. It also implies that if something could be done in a way without any effort, we human are likely to chose that way. As in the bike example, if the Glass could provided me directly what’s wrong with my bike and an precise procedure of fixing it, that would be something that I want.

I am skeptical about this approach. For me human perception is not a mode of passive receiving, but an interweave between receiving and actively searching.  Alva Noë wrote in his book <action in perception> that human’s perception is not something just happens in brain, but a skillful activity of human as a whole^[i]. According to him, the sense of vision, for example, is acquired not just by passively watching but also bodily engaging in the world. If one cannot interact with the world with his/her sensorimotor knowledge, even if his/her eyes physiologically function well, he/she still cannot perceive. This radical theory stresses the importance of sensorimotor skills in perception.

I therefore think, in the frame of just-in-time learning, it is not just about information providing, but more about encouraging active searching. It is about trying things out. what a computer can do in this system is to provide cues that could accelerate this interaction loop, not end it as soon as possible.

(more discussion to come….)

Assignment 3 – Sophia

I’m interested in using similar systems to the remembrance agents, etc. to augment the creative process.  Computers currently expect users to make the leap to remember where things are, what they are named, and that they even exist.  Similarly, tools and systems that creators rely on also put the burden on the user.  Creators must search for connections between disparate sources and ideas, and they must develop their own systems to help them do this.  Most artists I know rely on keeping a sketchbook, having an inspiration wall, or free writing to help them generate new ideas.  Few I know are able to use a computer to help them with this process, and many have complex rituals they use to put them in the right frame of mind for creation.  Often, the most important “aha moments” seem to come from nowhere.  What if a system could be devised that would help the user stay in this creative, associative mode more of the time?  How could inspirations, interesting connections be given to the user proactively?  It is much more difficult to determine what might be relevant to a creative process than in the memory agent.  Maybe this is where generative systems might be come into play (which have long been used by many artists, musicians, and writers in aiding them in their creative process).  I wonder how such generative systems might be able to work with the user’s current context? People are very good at immediately knowing if an idea is a compelling, but the hard part is blending information and concepts in new ways to come up with those ideas.  Maybe here the computer could be a help.  Would such a system work as a secondary task, something running on the side and not requiring the user’s full attention? 

I also felt interested in the idea of computers worn as jewelry.  Using metaphors of traditional objects changes the user’s expectations of what the computer can do.  How might one of these systems be different if it were incorporated in a bracelet, a bracer, a necklace, a locket, a ring, an earring, a pocket watch, a wristwatch, a headband, a crown, etc. as opposed to a magnifying glass, a book, glasses, etc.?  So many of these traditionally are ornaments and sometimes status symbols and only a few serve practical functions.

Assignment 3 – Champika

When thinking about how JITIR systems can be used in the support of learning it seems relevant to consider what we define as ‘learning’. Personally – I think it can be many things. We learn through creating, through conversation with others, through experimentation and tinkering, etc.  I would contrast this with rote memorization of information – which I would consider a more superficial learning as opposed to understanding. Given that definition of learning I think JITIRs can serve as something that remove’s inefficiencies in those activities. I think the Remembrance Agent Bradley discusses already does this for conversation. By allowing you to more easily pull up relevant and meaningful references in context – you could arguably engage in conversations that lead to more connections and learning. In terms of creating I could see JITIR systems playing a useful role as a source of inspiration – providing you with examples (similar, contradicting, provoking) and connecting you to helpful people (mentors?).

Three – Dhairya

I found some interesting insights mostly relating to the psychology and the ergonomics of a wearable display and a just in time system. I hope to use these in my thinking process.

-Factors of a context aware system: physical, temporal, information
-Users of a wearable system (henceforth called ‘wusers’) use more of the same information than a user who has no convenient way of accessing that information
-Present information to wusers in ways that can be ignored but is accessible
-Proactively queried information can be more distracting than suggestively queried information
-Zipf’s Principle of Least Effort: People will try to minimize their total future work, given their best estimates at the time and chooses the strategy with the best cost/benefit trade-off.
-More than two seconds of response delay is unacceptable and will result in fewer uses of a particular tool, even at the cost of decreased accuracy
-Focused attention and divided attention
-Proximity compatibility principle: High display proximity (similarity) helps in tasks with similar mental proximity, and where information is related and needs to be treated together
-’Ramping interface’ where information is conveyed in stages.

Assignment #3: Perovich

I was interested in how remembrance tools are meant to map the computer’s framework into the realm of human memory to make interactions more fluid and positive. Though I did begin to wonder how necessary this would be in the future–are human brains instead fully mapping themselves into the digital framework? There is increasing research on cognitive differences for this generation growing up immersed in technology. Sophia’s comments in class about her own experience also made me wonder how much we’re in fact meeting in the middle with technology on this front.

I thought the context tags as a memory structure was a very human and “brain friendly” approach. But it also made me think about how individuals store memories differently from each other–maybe I remember how to get to school based on landmarks, where another person remembers it based on geometry. It would be interesting to personalize systems based on these difference–and possible use this as a way to experience and explore someone else’s memory framework.

Furthermore, many of my strongest memories are “tagged” not by normal, easily captured, context clues (e.g. visual, location). Instead I reference them through an emotion, or a smell, or a texture. I wonder how these triggers might be supported–and whether these ways of remembering fall to the side if they’re not reinforced by technology support while others are.

Just-In-Time Learning Technologies

In order for just-in-time learning to work, it requires accurate delivery of content in various situations and in real-time. The problem with most technologies today is that the content is often not “smart” in the sense that its relevance is dependent on the user rather than the machine. In addition, it is difficult for real-time communication to occur, since there are few natural triggers that our devices utilize in order to present us with timely information.

In order for true situated learning to take place, technology must begin to recognize optimal learning places and times so that it can effectively present content. If a user must initiate the delivery of some type of information, the learning experience is limited to what the user already knows or can easily identify a need for. In reality, information that is often most impactful to the learner is not recognizable by the individual and therefore is often left inaccessible.

An intelligent system will be able to overcome the individual’s deficiency in judging cognitive load capabilities and awareness of content relevance. A JITIR agent will constantly scan the environment, recognize opportunities, and match those opportunities with ones that the learner is cognitively able to ingest. Over time, I imagine this could greatly increase learner self-efficacy and stimulate new areas of interest. In addition, it could refine knowledge in certain domains and filter unnecessary information, freeing up mental capacities for more meaningful tasks.