Design and Effects of Co-Learning in Human-AI Teams

Authors: Karel van den Bosch, Emma M. Van Zoelen, Tjeerd A.J. Schoonderwoerd, Anthia Solaki, Birgit Van der Stigchel, Ivana Akrum

JAIR 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Effects of LDPs on collaboration, knowledge and understanding, and team performance were experimentally evaluated using the testbed.
Researcher Affiliation Collaboration Karel van den Bosch EMAIL Human-Machine Teaming Organization for Applied Scientific Research, TNO Emma M. van Zoelen EMAIL Intelligent Systems Department Delft University of Technology Human-Machine Teaming Organization for Applied Scientific Research, TNO
Pseudocode No The paper does not contain any explicitly labeled pseudocode or algorithm blocks. Figure 4 presents a decision diagram, which is a flowchart, not pseudocode.
Open Source Code No The paper mentions using the MATRX software library (https://matrx-software.com/), but this is a third-party tool that the authors used, not the source code for their own methodology described in the paper. There is no explicit statement or link provided for the code implementation of their experiments.
Open Datasets No The paper describes the creation of a virtual USAR task and testbed (
Dataset Splits No The paper describes participant runs as 'a tutorial by means of a video-clip; then two training sessions with two learning challenges; a test session including all learning challenges', which refers to experimental sessions for human participants rather than dataset splits for model training. There is no mention of splitting a machine learning dataset into training, validation, or test sets.
Hardware Specification No The paper mentions that "The task was presented using a laptop" but does not provide specific details such as CPU, GPU, or memory models.
Software Dependencies Yes A simulation of the USAR task (see section 3.1) was developed using the Python programming language (version 3.7) and the MATRX (multi-agent teaming rapid experimentation)2 2.3.0 software library.
Experiment Setup Yes The paper provides extensive details on the experimental setup, including the design of the USAR task, robot behavior, learning opportunities, and learning activities (LDPs) in Sections 3.1, 3.2, 3.3, and 3.4. Furthermore, Appendix E and F explicitly detail the implementation and determination of 'Timing Thresholds for Agent Behavior' which are specific settings for the experiment logic.