Alone or by team of two, select one journal article and study it. Use all resources available (previous lectures and web based resources) to understand it.
Papers for journal club:
cognitive_exercise_for_persons_with_alzheimers_disease_and_related_dementia_using_a_social_robot.pdf
how_social_robots_influence_peoples_trust_in_critical_situations.pdf
humanintheloopoptimizationhri.pdf
socialrobotsconversationalcatalysts.pdf
socialrobotseffectonlearning.pdf
socialrobotsfor2ndlanguagelearning.pdf
uncannyvalley_robotsnonverbalbehaviour.pdf
icubrobotsfaceharsherjudgments.pdf
impact_of_active_vs_passive_robot_behavior.pdf
On the Journal article that you selected, the group (or individual) will be asked 6 questions to test your understanding. 3 questions will be precise on the text content [1pt per question], and 3 other questions will be open questions [1pt per question]. You will not have any notes or any help during the evaluation. For 2 more points you can summarize the journal article in front of the class [2pts].
(For a total of 8pts) (the total can change according to the other course, the proportion of points attributed will be the same).
Example of questions (using an unrelated journal article):
Bertacchini, F., Demarco, F., Scuro, C., Pantano, P., & Bilotta, E. (2023). A social robot connected with chatGPT to improve cognitive functioning in ASD subjects. Frontiers in Psychology, 14, 1232177.
PRECISE QUESTIONS:
- The article presents two distinct interaction scenarios with Pepper. What are the key differences in purpose and structure between the informal interaction scenario and the structured problem-solving scenario?
- The article expose some key points regarding the use of social robotics in the context of ASD. Give two examples of those.
- Constructing a specific architecture that links Pepper to OpenAI involves technical expertise in robotics, software development, and familiarity with the OpenAI platform. The authors propose an explanation of how to practically connect these systems, in order to simplify the process. Summarize very briefly the steps.
OPEN QUESTIONS:
- Patients might become overly dependent on the robot. To what extent should they be used as a “substitute” versus a “tool among other tools” in therapy. In you opinion who should draw the line?
- The two scenarios presented are described as simulated and toy examples. How far the experimental conditions are from a real world context and how could they validate the system clinically?
- Beyond ASD, could conversational AI be adapted enough to support the diversity of human profiles and interaction (from ADHD to language disorder for example), or does the heterogeneity of the possible interactions make a unified technological approach unrealistic?
Lukas Navarra will present his own work with LLms and the Pepper robot.