Project ideas, ELLI

Walled Garden vs. Open Community
In thinking about studying online behavior within SL, there are choices of using a walled garden or keeping the community open. The walled garden would control for the services, applications, content, that will be availabel to the comunity (communities) of study within SL. In a real walled garden, however, the user is unable to escape the walled garden to use other learning paths within SL. I'm not sure whether or not there is a need or a use for the walled garden.

Slideshare on Open vs. Walled Gardens

Mentoring
Would mentoring via available personnel or intelligent agent be easier within a walled garden? If the user is identified as part of the study and gives consent, can we simply follow their id within the open community without the need for walls?
Could the SL platform with the ELLI replace the human interface of mentors?
Can we formalize the number of recommendations based on what we expect the experts to do?
Regarding the mentoring of participants and the ELLI score, could I interview the group of mentors that have been trained in ELLI mentoring by Ruth? In my interviews, I would work to discover the kinds of comments and interventions the mentors will make, and the time elapsed before they will make them. Could I learn enough from the ELLI mentors to design an intelligent agent in SL that would mentor?

I served as a mentor teacher in Los Angeles Unified School District many years ago. I functioned as an intelligent agent would, albeit in the real instead of virtual world. I would visit teacher's and talk about what they were doing, what they wanted to do, what help, support, and skills they needed to reach their goals. I also was in a women's accountability group. All of us were in different places, but learned to trust each other to share our goals in life. We made suggestions, recommendations, and followed-up on the things that each of us had agreed we would accomplish over that month. We always had the option to change our mind and decide that the goal was not for us, but the group always kept us accountable and true to ourselves. We learned to ask for advice, share honestly if we had just failed to doit that month, and let ourselves be pushed toward the goals we alone had defined. The intelligent agent in SL has to be a compassionate, trustworthy, yet strong figure of accountability for the learner.

Human mentors should have time available, motivation to share experience and knowledge, willingness to learn, freedom to question and challenge, willingness to try new things, and patience. An intelligent agent should be programmed with the same functions to operate successfully with learners, especially at risk populations.
Human mentors can be role models, confidant, and nurterers of possibilities. What is possible for an agent we could create within SL?


About Intelligent Agents
Intelligent agents are computer programs within the system that act as autonomous entitites. An agent observes, and may learn or use knowledge to reach their predefined goals. they may be simple or complex.
Russell and Norvig define a rational agent.
ideal rational agent: "For each possible percept sequence, an ideal rational agent should do whatever action is expected to maximize its performance measure, on the basis of the evidence prvided by the percept sequence and whatever built-in knowledge the agent has."
To design the rational agent, the task environment must be specified.
Performance measure?
environment?
actuators?
sensors?

Other questions in designing a rational agent
Observable?
Deterministic?
Episodic?
Static?
Discrete?
Single-agent?

*Artificial Intelligence: A Modern Approach by Stuart Russell and Peter Norvig, 1995, Prentice-Hall, Inc.

Nikola Kasabov defines intelligent agents as "a systems should exhibit the following characteristics:
•accomodate new problem solving rules incrementally
•adapt online and in real time
•be able to analyze itself in terms of behavior, error and success
•learn and imporve through interaction with the environment (embodiment)
•learn quickly from large amounts of data
•have memory-based exemplar storage and retrieval capacities
•have parameters to represent short and long term memory, age, forgetting, etc.
*N. Kasabov, Introduction: Hybrid intelligent adaptive systems. International Journal of Intelligent Systems, Vol.6, (1998) 453–454.

Russell & Norvig (2003) group agents into five classes based on their degree of perceived intelligence and capability:[7]

simple reflex agents
model-based reflex agents
goal-based agents
utility-based agents
learning agents
Russell, Stuart J.; Norvig, Peter (2003), Artificial Intelligence: A Modern Approach (2nd ed.), Upper Saddle River, New Jersey: Prentice Hall, ISBN 0-13-790395-2, chpt. 2

An intelligent agent to mentor learners learning dimensions would be an automated online assistant, that would function to perceive the needs of learners in situ to perform individualized mentoring (supplying suggestions of alternative actions, supplying resources, experts and peers to contact, etc.) I see this agent in SL as an avatar that appears as a friendly support to learners.