QML-AiNet
A QML-AiNet is an Qualitative Model Learning System based on a AiNet strategy.
- AKA: QML-AiNet System.
- Context:
- It was first introduced by Pang & Coghill (2015).
- It implements an QML-AiNet Algorithm to solve an QML-AiNet AINet-CM Task.
- Example(s):
- that described in Pang & Coghill (2015),
- …
- Counter-Example(s):
- See: Immune Network, Artificial Immune Network Topology, Neural Network, Artificial Immune Systems, Qualitative Differential Equation, Qualitative Reasoning, Clonal Selection System, Compartmental Model, Machine Learning System.
References
2015
- (Pang & Coghill, 2015) ⇒ Wei Pang, and George M. Coghill (2015). "Qml-ainet: an immune network approach to learning qualitative differential equation models". Applied soft computing, 27, 148-157. DOI: 10.1016/j.asoc.2014.11.008.
- QUOTE: Qualitative Reasoning (QR) [1] is a field devoted to reasoning about complex systems at a qualitative level when only imprecise data and incomplete knowledge are available. In QR research there exist a few subfields, for instance, qualitative simulation (QS) based on qualitative differential equations (QDEs) [2][3][4], qualitative process theory (QPT) [5] [6], QDE model learning (QML, and see [7] for a review), qualitative tree induction [8][9], and most recently, learning qualitative models by estimating partial derivatives [10] and learning QPT models [11] .
(...) we proposed immune-inspired approaches to QML as they have been proven to be effective in solving many problems with large-scale and multi-model search spaces. In previous work we first proposed a pilot system EQML [12][13], an evolutionary qualitative model learning framework. In EQML, CLONALG [14][15], an evolutionary and immune-inspired algorithm based on the clonal selection theory [16] in immunology, was adapted to QML and its performance was compared against a genetic algorithm. Experimental results obtained from EQML showed that immune inspired approaches were feasible, and had great potential to be applied to QML.
(...) from the literature of Artificial Immune Systems (AIS), we know that Opt-AiNet [17],[18], an immune network approach to optimisation problems, can more effectively deal with large-scale and multimodal search spaces compared to CLONALG. This is because apart from using the clonal selection mechanism, OptAiNet also introduced interactions between antibodies, that is, the antibody population will undergo a network suppression procedure, and those antibodies with lower fitness values among similar antibodies will be eliminated to make the search diverse. Furthermore, the behaviour of the intrinsic diversity mechanisms of Opt-AiNet has been experimentally studied and confirmed [19].
Based on the above considerations, in this paper we will investigate the application of Opt-AiNet to better address the issues of scalability and highly multimodal search spaces in QML. We proposed an immune network approach with modified mutation operator, which we termed QML-AiNet (MM), to qualitative model learning. Compared to existing QML systems, the advantages of our system is as follows:
- Compared to QML systems using deterministic search algorithms, such as backtracking and branch-and-bound search, the proposed algorithm is more scalable to large search spaces.
- Compared to QME, which uses a genetic algorithm as its model learning strategy, the proposed algorithm can better deal with search spaces with multimodal fitness landscapes.
- More importably, compared to previous immune-inspired QML systems, our algorithm is more scalable to extremely large search spaces. Experiments have shown that the proposed QML-AiNet (MM) is two to three orders of magnitude more efficient than our previous immune-inspired systems QML-CLONALG [20] and QML-AiNet (OO) [21], an earlier version of QML-AiNet using the original mutation operator.
- QUOTE: Qualitative Reasoning (QR) [1] is a field devoted to reasoning about complex systems at a qualitative level when only imprecise data and incomplete knowledge are available. In QR research there exist a few subfields, for instance, qualitative simulation (QS) based on qualitative differential equations (QDEs) [2][3][4], qualitative process theory (QPT) [5] [6], QDE model learning (QML, and see [7] for a review), qualitative tree induction [8][9], and most recently, learning qualitative models by estimating partial derivatives [10] and learning QPT models [11] .
- ↑ Price C., Trave-Massuyes L., Milne R., Ironi L., Forbus K., Bredeweg B., Lee M., Struss P., Snooke N., Lucas P., Cavazza M., Coghill G. Qualitative futures. Knowl. Eng. Rev. 2006;21:317–334.
- ↑ Kuipers B. MIT Press; Cambridge, MA: 1994. Qualitative Reasoning: Modeling and Simulation with Incomplete Knowledge.
- ↑ Shen Q., Leitch R. Fuzzy qualitative simulation. IEEE Trans. Syst. Man Cybern. 1993;23:1038–1061.
- ↑ Coghill G.M. Heriot-Watt University; 1996. Mycroft: A Framework for Constraint Based Fuzzy Qualitative Reasoning. (Ph.D. thesis)
- ↑ Forbus K.D. Qualitative process theory. Artif. Intell. 1984;24:85–168.
- ↑ Forbus K.D. The qualitative process engine. In: Weld D.S., de Kleer J., editors. Readings in Qualitative Reasoning About Physical Systems. Morgan Kaufmann Publishers Inc.; San Francisco, CA: 1990. pp. 220–235.
- ↑ Pang W., Coghill G.M. Learning qualitative differential equation models: a survey of algorithms and applications. Knowl. Eng. Rev. 2010;25:69–107.
- ↑ ˘Suc D., Bratko I. EMCL ’01 Proceedings of the 12th European Conference on Machine Learning. Springer-Verlag; London, UK: 2001. Induction of qualitative trees; pp. 442–453
- ↑ Bratko I., ˘Suc D. Learning qualitative models. AI Mag. 2003;24:107–119.
- ↑ Žabkar J., Mož ina M., Bratko I., Demšar J. Learning qualitative models from numerical data. Artif. Intell. 2011;175:1604–1619.
- ↑ Hinrichs T.R., Forbus K.D. Proceedings of the Twenty-Sixth AAAI Conference on Artificial Intelligence. AAAI Press; 2012. Learning qualitative models by demonstration; pp. 207–213.
- ↑ Pang W., Coghill G.M. Evolutionary approaches for learning qualitative compartment metabolic models. In: Wang X., Li R.F., editors. The 6th annual UK Workshop on Computational Intelligence. University of Leeds; Leeds, UK: 2006. pp. 11–16.
- ↑ Pang W., Coghill G.M. Fifth International Conference on Frontier of Computer Science and Technology. IEEE Computer Society; Changchun, China: 2010. Learning qualitative metabolic models using evolutionary methods; pp. 436–441.
- ↑ de Castro V., Zuben The clonal selection algorithm with engineering applications. Proceedings of GECCO, Workshop on Artificial Immune Systems and Their Applications, Morgan Kaufmann; Las Vegas, USA; 2000. pp. 36–39.
- ↑ de Castro, Zuben F.J.V. vol. 6. IEEE Press; 2002. Learning and optimization using the clonal selection principle; pp. 239–251. (IEEE Transactions on Evolutionary Computation, Special Issue on Artificial Immune Systems)
- ↑ Burnet F. Cambridge University Press; Cambridge: 1959. The Clonal Selection Theory of Acquired Immunity
- ↑ de Castro J., Timmis . Proceedings of IEEE Congress on Evolutionary Computation (CEC’02) IEEE Press; 2002. An artificial immune network for multimodal function optimization; pp. 674–699.
- ↑ Timmis J., Edmonds C. A comment on opt-AINet: an immune network algorithm for optimisation. In: Kalyanmoy D., editor. vol. 3102. Springer; Heidelberg: 2004. pp. 308–317. (Genetic and Evolutionary Computation (GECCO 2004), Lecture Notes in Computer Science).
- ↑ de França F.O., Coelho G.P., Zuben F.J.V. Proceedings of the IEEE Congress on Evolutionary Computation, CEC. IEEE; Barcelona, Spain: 2010. On the diversity mechanisms of opt-aiNet: a comparative study with fitness sharing; pp. 1–8.
- ↑ Pang W., Coghill G.M. Modified clonal selection algorithm for learning qualitative compartmental models of metabolic systems. In: Thierens D., editor. Genetic and Evolutionary Computation Conference (GECCO07) ACM Press; New York, NY, USA: 2007. pp. 2887–2894.
- ↑ Pang W., Coghill G.M. QML-AiNet: an immune-inspired network approach to qualitative model learning. In: Hart E., editor. LNCS 6209, Proceedings of 9th International Conference on Artificial Immune Systems (ICARIS 2010) Springer; Edinburgh, UK: 2010. pp. 223–236.