Reproducibility Beyond Artifacts: Interactional Support for Collaborative Machine Learning
This work addresses reproducibility challenges for collaborative and interdisciplinary ML teams, offering a novel approach that goes beyond incremental artifact-based solutions.
The paper tackles the problem of machine learning reproducibility in collaborative projects by identifying that failures often stem from interactional breakdowns beyond just missing artifacts, and proposes a two-layer socio-technical system to address these issues through improved coordination and shared understanding.
Machine learning (ML) reproducibility is often framed as a problem of incomplete artifact recording. This framing leads to systems that prioritize capturing datasets, code, configurations, and execution environments.However, in collaborative and interdisciplinary ML projects, reproducibility failures often arise not only from missing artifacts but from difficulties in interpreting prior work, aligning evolving components, and reconstructing experimental intent over time. Drawing on a 19-month deployment of a data-centric ML management system in a clinical research project, we identify recurring interactional breakdowns that persist despite comprehensive structural traceability. Based on these findings, we propose a two-layer socio-technical ML management system combining lifecycle-aware artifact infrastructure with an interactional layer designed to mediate coordination, explanation, and shared understanding. We discuss how an AI-mediated semantic interface reframes reproducibility as an ongoing socio-technical accomplishment rather than a static property of recorded traces, and outline implications for human-centered ML infrastructure design.