BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Chicago
X-LIC-LOCATION:America/Chicago
BEGIN:DAYLIGHT
TZOFFSETFROM:-0600
TZOFFSETTO:-0500
TZNAME:CDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0500
TZOFFSETTO:-0600
TZNAME:CST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20251016T135128Z
LOCATION:Grand Hall J
DTSTART;TZID=America/Chicago:20251016T141000
DTEND;TZID=America/Chicago:20251016T143000
UID:HFESAM_ASPIRE 2025_sess223_LECT254@linklings.com
SUMMARY:Inverse Counterfactual for AI-Assisted Decision Support: Enhancing
  Knowledge Elicitation for Capturing  Aircraft Pilot Decisions
DESCRIPTION:Jonay Ramon Alaman (Universite Laval); Daniel Lafond (Thales);
  Alexandre Marois (Université Laval, University of Central Lancashire); an
 d Sébastien Tremblay (Université Laval)\n\nIntegrating AI into decision-su
 pport systems (DSS) for safety-critical tasks like piloting poses challeng
 es for human-machine interaction. DSS must align with the mental model of 
 pilots and provide relevant information. One promising approach, the Cogni
 tive Shadow, addresses this challenge by continuously modelling each opera
 tor strategy and issuing alerts whenever its recommendation diverges from 
 the operator decision. We evaluated a new operator-driven knowledge-elicit
 ation procedure: the inverse counterfactual. This procedure aims to enhanc
 e human-AI model similarity. After choosing the optimal option, the user m
 odifies one factor, so the second-best option becomes preferable, providin
 g the AI with a contrasting case across the decision boundary.\nIn a simul
 ated adverse-weather avoidance task, forty-four participants solved 130 sc
 enarios and generated inverse counterfactuals for 20 additional cases. On 
 the final test phase, models trained with these paired cases demonstrated 
 comparable predictive accuracy to conventional models. However, when edits
  in counterfactual cases were minimal—i.e., closer to the decision boundar
 y— predictive accuracy improved and DSS advice was accepted more frequentl
 y. Larger edits degraded performance. The results suggest that well-guided
  counterfactual interactions may sharpen knowledge elicitation.\nThe prese
 ntation will introduce the methodology, discuss the results, and outline f
 uture research aimed at refining the method to better align with human cog
 nitive processes.\n\nTrack: Cognitive Engineering & Decision Making\n\nSes
 sion Chairs: Leia Stirling (University of Michigan) and Mansoor Nasir (For
 d Motor Company)\n\n
END:VEVENT
END:VCALENDAR