Towards interactive Machine Learning (iML): Applying Ant Colony Algorithms to Solve the Traveling Salesman Problem with
Most Machine Learning (ML) researchers focus on automatic Machine Learning (aML) where great advances have been made, for example, in speech recognition, recommender systems, or autonomous vehicles. Automatic approaches greatly benefit from the availabili
- PDF / 684,495 Bytes
- 15 Pages / 439.37 x 666.142 pts Page_size
- 71 Downloads / 184 Views
4
Holzinger Group HCI-KDD, Institute for Medical Informatics, Statistics & Documentation, Medical University Graz, Graz, Austria {a.holzinger,m.plass,k.holzinger}@hci-kdd.org 2 Vasile Alecsandri University of Bacˇ au, Bacˇ au, Romania [email protected] 3 Technical University of Cluj-Napoca, Cluj-Napoca, Romania [email protected] Faculty of Engineering, Environment and Computing, Coventry University, Coventry, UK [email protected]
Abstract. Most Machine Learning (ML) researchers focus on automatic Machine Learning (aML) where great advances have been made, for example, in speech recognition, recommender systems, or autonomous vehicles. Automatic approaches greatly benefit from the availability of “big data”. However, sometimes, for example in health informatics, we are confronted not a small number of data sets or rare events, and with complex problems where aML-approaches fail or deliver unsatisfactory results. Here, interactive Machine Learning (iML) may be of help and the “human-in-the-loop” approach may be beneficial in solving computationally hard problems, where human expertise can help to reduce an exponential search space through heuristics. In this paper, experiments are discussed which help to evaluate the effectiveness of the iML-“human-in-the-loop” approach, particularly in opening the “black box”, thereby enabling a human to directly and indirectly manipulating and interacting with an algorithm. For this purpose, we selected the Ant Colony Optimization (ACO) framework, and use it on the Traveling Salesman Problem (TSP) which is of high importance in solving many practical problems in health informatics, e.g. in the study of proteins. Keywords: interactive Machine Learning · Human-in-the-loop Traveling Salesman Problem · Ant Colony Optimization
c IFIP International Federation for Information Processing 2016 Published by Springer International Publishing Switzerland 2016. All Rights Reserved F. Buccafurri et al. (Eds.): CD-ARES 2016, LNCS 9817, pp. 81–95, 2016. DOI: 10.1007/978-3-319-45507-5 6
·
82
1
A. Holzinger et al.
Introduction and Motivation for Research
Automatic Machine Learning (aML) is increasingly making big theoretical as well as practical advances in many application domains, for example, in speech recognition [1], recommender systems [2], or autonomous vehicles [3]. The aML-approaches sometimes fail or deliver unsatisfactory results, when being confronted with complex problem. Here interactive Machine Learning (iML) may be of help and a “human-in-the-loop” may be beneficial in solving computationally hard problems, where human expertise can help to reduce, through heuristics, an exponential search space. We define iML-approaches as algorithms that can interact with both computational agents and human agents and can optimize their learning behaviour through these interactions [4]. To clearly distinguish the iML-approach from a classic supervised learning approach, the first question is to define the human’s role in this loop (see Fig. 1), [5].
Fig. 1. The iML human
Data Loading...
