Virtual environments MOOs and Virtual agents Virtual Environments

Virtual environments, MOOs and Virtual agents

Virtual Environments n Readings for this week: n n Peterson 1998 (VLE) Peterson 2004 MOO Morton and Jack 2005 Virtual agents Development of technology

Virtual environments n Learning environment (Peterson 1998) n n Very familiar one these days Does not now seem novel

Construction of the learning environment n n n Select a learning theory Cognitive processing model (Bialystock) Identify learner needs n n Needs analysis (questionnaire) (? ? ) Choose website design tools n n Netscape Navigator Gold Browser/editor Hand-coding Dreamweaver etc.

Construction of the learning environment n Instructional design/HCI (human-computer interface) issues n n Choice of number of links, font type and size, use of colour, arrangement of the page Links -- page 1 n n Cutting edge CALL Resources Sch. MOOze University Online English Grammar ESL Café

Construction of the learning environment n Links -- page 2 Technical Writing Page n n n Bilingual English-Japanese Online dictionary Online Writing Lab Online Thesaurus The Elements of Style etc. Links -- page 3 Presentation Resources n n n The Virtual Presentation Assistant Briefing Notes on Giving Short Talks Giving a Scientific Talk

Virtual Learning Environment n Site Evaluation n Student feedback n n n Lost in space -- Frames-based site More interactive materials needed More visual metaphors for navigation Online feedback link (email) Wider range of sites Site redesign

Many VLEs available n n n Individual sites, like Peterson’s CMS sites (Course Management System) Moodle, Web CT Intuto. com -- local online learning company

Virtual Learning Environment n n Too static ? ? Should be possible to create an individualised VLE n n Student types in requirements Web-page is generated based on those requirements

MOO n n MOOs and MUDs MOO -- multi-user object-oriented domain MOOs are virtual environments designed to facilitate synchronous textbased communication among users More permanent than chat rooms

Sch. MOOze University n n n http: //schmooze. hunter. cuny. edu/ Users log on (to a virtual domain such as a university) Create a nickname (and adopt an online persona ? ? ) Users then interact, navigate and manipulate virtual objects Series of virtual rooms and objects

Advantages of MOOs n n Increased communication Reduced stress n n n Anonymous user New persona Easy to make a contribution

Chatbots n n n Original program -- Eliza Conversation with a psychiatrist (Rogerian type psychiatry) Designed to show that dumb programs could appear to be intelligent Eliza and chatbots http: //www. cmr. fuberlin. de/~mck/courses/lv 00 ss/Pe. KMan /team 7/eliza. html

Chatbots n n Turing test -- a test to see if a computer is intelligent. Loebner prize -- annual competition for chatbots

Chatbot plus voice n n http: //www. daden. co. uk/chatbots/pages/ 000067. html http: //www. alicebot. org/

Visual agents n n n Morton & Jack reading Avatars -- virtual beings -- extensions of humans in the virtual world. An avatar may be an virtual “you” Visual agents -- other beings in a virtual world

Spoken Electronic Language Learning n n SPELL -- Morton & Jack Includes speech recognition n How good is speech recognition? How good is it with language learners Goal in this system is not to improve pronunciation, but to understand what the student says

Semantic representations n My guess is that the system uses representations of meaning based on verbs and their arguments: n n Eat (I, hamburger) Want (I, (Eat (I, hamburger)) Want (I, (Eat (I, ? ? )) See (I, You)

Semantic representations n Dialogue n n n Speech recognition has to decode the speech well enough to recognise hamburger or pizza etc. and create the meaning representation n Question: What do you want to eat? Learner: I’d like a pizza Want (I, (Eat (I, pizza)) This can then be used to continue with the dialogue -- what kind of pizza would you like Is the goal to have a dialogue or give feedback? ?

Desirable characteristics of speech interactive CALL n Wachowicz and Scott 1999 n Adopted by SPELL

Interactions in SPELL n n n Learner and computer interact -- no learner input, no dialogue Constrained environment -- so that the learner contribution can be understood Scenarios n n n Observational scenario One-to-one scenario Interactive scenario

Observational scenario

Observational scenario n n n Clear situation Students listen to the interaction Subtitles available Stop/start/replay the dialogue Access to other materials

One-to-one

One-to-one n Virtual tutor agent asks the learner some questions n n n About themselves About the dialogue What foods did the virtual people like? What foods does the learner like? Agents use pre-recorded audio files

Interactive scenario

Interactive scenario n n Learner enters the scene If he orders water, the waitress will bring water. Constrained environment limits what the learner is likely to say Recognition grammar -- range of utterances that the customer might use

Interactive scenario n n n Recognition grammar developed for each stage of the dialogue Possible learner “errors” are added to the recognition grammar The grammar is loaded into the program for each stage

Interactive scenario n For each stage, there assumed to be four types of response

Recasts

Recognition Grammar

Help for the learner -reformulation

Error analysis n Errors for each learner are logged

Prototype system n n Technical developments -- speech recognition of pronunciation Students want more “affective” behaviour from the visual agents n (Eliza effect)

Virtual Reality n MOOs are VR environments n n Text-based Active Worlds -http: //www. activeworlds. com/ n Education programs

Active Worlds

MOOs, Avatars, CMC n n Where is the learning? Issues?