Our innovations — many of them covered by pending patent applications — span the fields of interactive simulations, interactive multimedia instruction, intelligent tutoring systems, artificial intelligence, speech recognition, human-computer interaction, advanced software architectures, and authoring and production.

Rich interactive simulations with "socially intelligent virtual humans"

Our immersive, interactive 3D video games simulate real-life communication, allowing users to role play with animated "socially intelligent virtual humans" that recognize the user's speech, intent, gestures and behavior.

We promote extensive levels of engagement, motivation and practice by providing "free form" storylines with very wide ranges of game-play paths, interactive dialogs and action options. For instance, in our Mission to Iraq language course if the user speaks and behaves correctly, the virtual Iraqis become trustful and cooperative, and provide information that users need to advance. Otherwise, the virtual Iraqis are uncooperative and prevent learners from "winning" the game.

Our Cultural Puppets™ technology drives virtual humans to exhibit behavior that is appropriate to the simulation's contexts of storyline, culture, task, user actions, and the personality, emotions and intent specific to each virtual human. User communication towards virtual humans might elicit different responses. For example, in one scene in the Mission to Iraq course, the user's disrespect towards two virtual Iraqis elicits a passive response from one and insults from the other.

Cultural protocols involve cultural knowledge, sensitivity and awareness — including non-verbal gestures, etiquette, and norms of politeness — that are critical for successful communication. In another scene of the Mission to Iraq course, the user can speak perfect Arabic and still offend his virtual Iraqi host by not asking him to enter the house first or by later talking to his wife without observing Iraqi cultural protocols.

Our automated speech recognizer is optimized to understand native and non-native speakers of the language. it tolerates pronunciation errors while supporting pronunciation training.
» See sidebar

Intelligent tutoring systems

We guide our work by adoption-based research in the learning sciences, which is predicated upon and contributes to the successful adoption of effective learning systems. We use our learning and simulation platforms as both versatile and evolving content delivery systems, and testbeds for evaluation of new learning methods. We continually collect data from field use of our learning systems, analyze the data, publish evaluation results, and use the results to inform further research and development.

Game-based learning is an important new area for intelligent learning environments, and we are at the forefront of this trend.

We are also playing a leading role in the development of pedagogical agents, which are animated agents that promote learning by interacting with learners at both cognitive and affective levels. For example, we continue to conduct fundamental research on the politeness effect, whereby learning systems that adhere to social norms of politeness in human-computer interaction promote better learning than learning systems that violate those norms. These principles guide the behavior of the virtual tutors and socially intelligent virtual humans in our intelligent tutoring systems.

Our studies show that the manner in which systems provide tutorial feedback has an effect on learning outcomes. Feedback that is encouraging and sensitive to the learner's sense of self-esteem leads to better learning than simply telling learners when their responses are right or wrong.

Corrective feedback is most effective when it is embedded in the game, instead of being in tutorial critiques. For example, if the learner inadvertently is rude to a virtual Iraqi in the Mission to Iraq course, he may call the learner a "son of a dog". This very effectively gets the learner's attention without damaging the learner's self-esteem.

Authoring and production

Our Authoring and Production Framework provides methods and web-based tools to create quickly and economically courses and simulations that incorporate our technologic and pedagogic innovations.

The three top priorities of our Framework are:

• Support the effective collaboration among multidisciplinary, geographically disperse teams.
• Remove the involvement of engineers and researchers so instructional designers, content specialists and animators can produce courses and simulations themselves without mastering the intricacies of our underlying technology.
• Provide a rich environment for managing the full authoring and production life-cycle — from initial concept formulation, to content authoring and testing, to product release and post-release maintenance.

Architecture and platforms

Our Alelo Architecture serves all stakeholders: authors, artists, producers, engineers, instructors, researchers and, of course, users.

Authors, artists, producers and engineers access central repositories of content using web-based tools to create, produce and deploy simulations and courses.

Users get simulations and courses deployed on several platforms, including desktop and laptop computers, websites, and handheld devices like Windows Mobile computers and the Apple iPod. Our current PC-based games use the Unreal Engine by Epic Games, but we are extending our architecture to support other game engines so we may more easily integrate with other products and technologies, as well as find new delivery options and modes of play such as multi-user web-based environments.

Instructors and educators manage their students training on a SCORM-compliant learning management system (LMS). Researchers get access to usage and progress databases.