Details
| Original language | English |
|---|---|
| Article number | 9321731 |
| Pages (from-to) | 3037-3054 |
| Number of pages | 18 |
| Journal | IEEE Transactions on Pattern Analysis and Machine Intelligence |
| Volume | 43 |
| Issue number | 9 |
| Publication status | Published - 1 Sept 2021 |
| Externally published | Yes |
Abstract
Automated machine learning (AutoML) supports the algorithmic construction and data-specific customization of machine learning pipelines, including the selection, combination, and parametrization of machine learning algorithms as main constituents. Generally speaking, AutoML approaches comprise two major components: a search space model and an optimizer for traversing the space. Recent approaches have shown impressive results in the realm of supervised learning, most notably (single-label) classification (SLC). Moreover, first attempts at extending these approaches towards multi-label classification (MLC) have been made. While the space of candidate pipelines is already huge in SLC, the complexity of the search space is raised to an even higher power in MLC. One may wonder, therefore, whether and to what extent optimizers established for SLC can scale to this increased complexity, and how they compare to each other. This paper makes the following contributions: First, we survey existing approaches to AutoML for MLC. Second, we augment these approaches with optimizers not previously tried for MLC. Third, we propose a benchmarking framework that supports a fair and systematic comparison. Fourth, we conduct an extensive experimental study, evaluating the methods on a suite of MLC problems. We find a grammar-based best-first search to compare favorably to other optimizers.
Keywords
- Automated machine learning, Bayesian optimization, hierarchical planning, multi-label classification
ASJC Scopus subject areas
- Computer Science(all)
- Software
- Computer Science(all)
- Computer Vision and Pattern Recognition
- Computer Science(all)
- Computational Theory and Mathematics
- Computer Science(all)
- Artificial Intelligence
- Mathematics(all)
- Applied Mathematics
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
In: IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 43, No. 9, 9321731, 01.09.2021, p. 3037-3054.
Research output: Contribution to journal › Review article › Research › peer review
}
TY - JOUR
T1 - AutoML for Multi-Label Classification
T2 - Overview and Empirical Evaluation
AU - Wever, Marcel
AU - Tornede, Alexander
AU - Mohr, Felix
AU - Hullermeier, Eyke
N1 - Publisher Copyright: © 1979-2012 IEEE.
PY - 2021/9/1
Y1 - 2021/9/1
N2 - Automated machine learning (AutoML) supports the algorithmic construction and data-specific customization of machine learning pipelines, including the selection, combination, and parametrization of machine learning algorithms as main constituents. Generally speaking, AutoML approaches comprise two major components: a search space model and an optimizer for traversing the space. Recent approaches have shown impressive results in the realm of supervised learning, most notably (single-label) classification (SLC). Moreover, first attempts at extending these approaches towards multi-label classification (MLC) have been made. While the space of candidate pipelines is already huge in SLC, the complexity of the search space is raised to an even higher power in MLC. One may wonder, therefore, whether and to what extent optimizers established for SLC can scale to this increased complexity, and how they compare to each other. This paper makes the following contributions: First, we survey existing approaches to AutoML for MLC. Second, we augment these approaches with optimizers not previously tried for MLC. Third, we propose a benchmarking framework that supports a fair and systematic comparison. Fourth, we conduct an extensive experimental study, evaluating the methods on a suite of MLC problems. We find a grammar-based best-first search to compare favorably to other optimizers.
AB - Automated machine learning (AutoML) supports the algorithmic construction and data-specific customization of machine learning pipelines, including the selection, combination, and parametrization of machine learning algorithms as main constituents. Generally speaking, AutoML approaches comprise two major components: a search space model and an optimizer for traversing the space. Recent approaches have shown impressive results in the realm of supervised learning, most notably (single-label) classification (SLC). Moreover, first attempts at extending these approaches towards multi-label classification (MLC) have been made. While the space of candidate pipelines is already huge in SLC, the complexity of the search space is raised to an even higher power in MLC. One may wonder, therefore, whether and to what extent optimizers established for SLC can scale to this increased complexity, and how they compare to each other. This paper makes the following contributions: First, we survey existing approaches to AutoML for MLC. Second, we augment these approaches with optimizers not previously tried for MLC. Third, we propose a benchmarking framework that supports a fair and systematic comparison. Fourth, we conduct an extensive experimental study, evaluating the methods on a suite of MLC problems. We find a grammar-based best-first search to compare favorably to other optimizers.
KW - Automated machine learning
KW - Bayesian optimization
KW - hierarchical planning
KW - multi-label classification
UR - http://www.scopus.com/inward/record.url?scp=85099549846&partnerID=8YFLogxK
U2 - 10.1109/TPAMI.2021.3051276
DO - 10.1109/TPAMI.2021.3051276
M3 - Review article
C2 - 33439834
VL - 43
SP - 3037
EP - 3054
JO - IEEE Transactions on Pattern Analysis and Machine Intelligence
JF - IEEE Transactions on Pattern Analysis and Machine Intelligence
SN - 0162-8828
IS - 9
M1 - 9321731
ER -