Contemporary Machine Learning (ML) often focuses on large existing and labeled datasets and metrics around accuracy and performance. In pervasive online systems, conditions change constantly and there is a need for systems that can adapt. In Machine Teaching (MT) a human domain expert is responsible for the knowledge transfer and can thus address this. In my work, I focus on domain experts and the importance of, for the ML system, available features and the space they span. This space confines the, to the ML systems, observable fragment of the physical world. My investigation of the feature space is grounded in a conducted study and related theories. The result of this work is applicable when designing systems where domain experts have a key role as teachers.
[1]
Hanghang Tong,et al.
PC-Fairness: A Unified Framework for Measuring Causality-based Fairness
,
2019,
NeurIPS.
[2]
David Maxwell Chickering,et al.
Machine Teaching: A New Paradigm for Building Machine Learning Systems
,
2017,
ArXiv.
[3]
W. B. Roberts,et al.
Machine Learning: The High Interest Credit Card of Technical Debt
,
2014
.
[4]
Xiaojin Zhu,et al.
Machine Teaching: An Inverse Problem to Machine Learning and an Approach Toward Optimal Education
,
2015,
AAAI.
[5]
Been Kim,et al.
Towards A Rigorous Science of Interpretable Machine Learning
,
2017,
1702.08608.