270 likes | 417 Views
Atgentive. Choices and constraints for Atgentive’s conceptual framework. Aims. Intelligent support for attention in the context of collaborative learning environments (DoW, p.18) Why the need to support attention in learning environments? What aspects of attention are relevant?
E N D
Atgentive Choices and constraints for Atgentive’s conceptual framework
Aims • Intelligent support for attention in the context of collaborative learning environments(DoW, p.18) • Why the need to support attention in learning environments? • What aspects of attention are relevant? • Where do we go from here? Claudia Roda - AUP - Atgentive WP1
Attention in learning environments • Support of learning not limited to lecture based (experiential, project based, context based) Choice Large amount of information available to students Autonomy Problems formulationIssues identificationSolution recognition Claudia Roda - AUP - Atgentive WP1
Learners guidance Choice Large amount of information available to students Autonomy Problems formulationIssues identificationSolution recognition Attention guidance, minimisation of distractors Claudia Roda - AUP - Atgentive WP1
Research questions * System Attention mng ** System Attention mngwith emb.agts *** System Claudia Roda - AUP - Atgentive WP1
Research questions * System Attention mng ** System Attention mngwith emb.agts *** System What do we measure? Claudia Roda - AUP - Atgentive WP1
Research questions Perceived Objective * • Achievements (e.g. user has learned X, user has solved Y, …) • Performance (e.g user has learned X in time T, user has found and efficient solution to Y) ** *** What do we measure? Claudia Roda - AUP - Atgentive WP1
Research questions Perceived Objective * • Achievements (e.g. user has learned X, user has solved Y, …) • Performance (e.g user has learned X in time T, user has found and efficient solution to Y) • Pleasurableness • ??? ** *** What do we measure? Claudia Roda - AUP - Atgentive WP1
Research questions Perceived Objective * • Achievements (e.g. user has learned X, user has solved Y, …) • Performance (e.g user has learned X in time T, user has found and efficient solution to Y) • Pleasurableness • ??? ** *** What do we measure? Claudia Roda - AUP - Atgentive WP1
Research questions • How are Perceived / Objective achievements, performance, pleasure, etc. related to lower level indicators such as: • Cognitive load • Memory load • Interaction dynamics • … * Attention as a tool for cognitive (and physical?) resources allocation ** *** What do we measure? Claudia Roda - AUP - Atgentive WP1
What do we know about attention • Attention is the set of processes enabling and guiding the selection of incoming perceptual information in order to limit the external stimuli processed by our bounded cognitive system and to avoid overloading it (Posner 1982; Lavie and Tsal 1994; Chun and Wolfe 2001; Driver 2001). Claudia Roda - AUP - Atgentive WP1
How is attention controlled? • Exogenous, bottom-up, stimulus-driven attention. • Attention captured by external event • Different degrees of power • Automatic attention drive • Rapid, transient time course • Endogenous, top-down, goal-driven attention. • Attention is controlled voluntarily • Effortful • Slow (sustained) time course Claudia Roda - AUP - Atgentive WP1
How is attention controlled? • Exogenous, bottom-up, stimulus-driven attention. • Attention captured by external event • Different degrees of power • Automatic attention drive • Rapid, transient time course Sensory mechanisms supporting attention • Endogenous, top-down, goal-driven attention. • Attention is controlled voluntarily • Effortful • Slow (sustained) time course Motivational mechanisms supporting attention Claudia Roda - AUP - Atgentive WP1
How is attention controlled? • Exogenous, bottom-up, stimulus-driven attention. • Attention captured by external event • Different degrees of power • Automatic attention drive • Rapid, transient time course Sensory mechanisms supporting attention Interaction, e.g. what I am looking for may determine what I can ignore • Endogenous, top-down, goal-driven attention. • Attention is controlled voluntarily • Effortful • Slow (sustained) time course Motivational mechanisms supporting attention Claudia Roda - AUP - Atgentive WP1
Which information, at what time, and in which format is best suited for presentation to the learner Information about user Information about environment Atgentive Sensory mechanisms supporting attention Motivational mechanisms supporting attention Claudia Roda - AUP - Atgentive WP1
Detect current user’s attentional state Determine possible alternative foci Evaluate cost/benefits of attentional shifts Decide if / when / how to intervene Atgentive We know that switching attention comes at a cost Which information, at what time, and in which format is best suited for presentation to the learner Claudia Roda - AUP - Atgentive WP1
Atgentive Sensory mechanisms supporting attention Motivational mechanisms supporting attention Note: we may want to maintain focus rather than switching focus Detect current user’s attentional state Determine possible alternative foci Evaluate cost/benefits of attentional shifts Decide if / when / how to intervene Claudia Roda - AUP - Atgentive WP1
What do we need to know • What are the essential stimuli selection processes? • How stimuli’s relevance is evaluated • Why sometime there are “strange” attentional behaviours? (Interference) • How do higher level (motivational) processes interact with lower level sensory processes? • Do emotions play a role in all this? • Are there “special cases” where we know more on how attention intervenes to reduce cognitive load? • … Stimuli selection Stimuli relevance Interference Top-down/bottom-up Emotion Task specific Claudia Roda - AUP - Atgentive WP1
Stimuli selection • How the selection of exogenous stimuli may take place ? • First hypothesis - two stage processing of external stimuli: • parallel preattentive process filtering non-relevant stimuli ( simple physical properties). • serial, attentive stage - limited processing encoding more abstract properties of the attended stimuli. (Broadbent 1958) • Later theories: all stimuli are analysed, but only pertinent stimuli are selected for awareness and memorisation. • Recent theories: although not all the stimuli are analysed, non attended stimuli are not completely filtered out either, stimuli impact depends upon their relevance to the environment or personal experience . Stimuli selection Stimuli relevance Interference Top-down/bottom-up Emotion Task specific Claudia Roda - AUP - Atgentive WP1
Relevance evaluation Learning involves the creation of a set of expectation with respect to external stimuli. These expectations allow us to focus on the expected data. It is the resonance between expectations and received input that brings certain stimuli to the conscious state and generates learning Research on the influence of top-down processes on bottom-up ones (Executive Control Processes, ACT_R, Adaptive Resonance Theory, …) Stimuli selection Stimuli relevance Interference Top-down/bottom-up Emotion Task specific Claudia Roda - AUP - Atgentive WP1
Interference • Sometime unable to disregard irrelevant stimuli • Lavie and Tsal theory on capacity • Sometime we miss noticing something obviously present • Change blindness • Distractors may delay attentional processes (time and relation w/t target stimuli) • Other interferences • Negative priming, Learned inattentions Stimuli selection Stimuli relevance Interference Top-down/bottom-up Emotion Task specific Claudia Roda - AUP - Atgentive WP1
Top-down / bottom-up • Motivation affects attention both at the conscious and unconscious level (nonconscious cues) • Positive/negative valence associated to stimuli and to goals at motivational level (attaining/avoiding something) • Automatic vigilance terminates when goal is attained Stimuli selection Stimuli relevance Interference Top-down/bottom-up Emotion Task specific Claudia Roda - AUP - Atgentive WP1
Emotions • Hypothesis: “happier moods promote a greater focus on the forest and sadder moods a greater focus on the trees”. (Gasper & Clore 2002) • Global attention versus local attention Stimuli selection Stimuli relevance Interference Top-down/bottom-up Emotion Task specific Claudia Roda - AUP - Atgentive WP1
Task specific • Cognitive Load Theory: working memory size is a fundamental bottleneck • Information for task support divided in supportive and procedural (theory/routine) - presentation before/during task execution. Avoid split attention. Stimuli selection Stimuli relevance Interference Top-down/bottom-up Emotion Task specific Claudia Roda - AUP - Atgentive WP1
Towards the Atgentive framework Sensory mechanisms supporting attention Information about user Information about environment Motivational mechanisms supporting attention Claudia Roda - AUP - Atgentive WP1
Sensory mechanisms supporting attention Motivational mechanisms supporting attention Towards the Atgentive framework Information about environment Information about user System: AtGentSchoolAtGentNet Claudia Roda - AUP - Atgentive WP1
Sensory mechanisms supporting attention Motivational mechanisms supporting attention Towards the Atgentive framework • Information about environment • Information about content • Required cognitive load • Local/global • Relation to other content • Relation to task (supporting/procedural) • Relation to goals (positive/negative valence) • … • Information about user • Current concerns • Ignored and pursued stimuli • Goals (achieved, suspended, failed, impossible) • Tasks • Relations to other users • … System: AtGentSchoolAtGentNet Claudia Roda - AUP - Atgentive WP1