Multi-document Summarization Using Adaptive Composite Differential Evolution

In the current paper, a system for multi-document summarization (MLDS) is developed which simultaneously optimized different quality measures to obtain a good summary. These measures include anti-redundancy, coverage, and, readability. For optimization, multi-objective binary differential evolution (MBDE) is utilized which is an evolutionary algorithm. MBDE consists of a set of solutions and each solution represents a subset of sentences to be selected in the summary. Generally, MBDE uses a single DE variant, but, here, an ensemble of two different DE variants measuring diversity among solutions and convergence towards the global optimal solution, respectively, is employed for efficient search. Three versions of the proposed model are developed varying the syntactic/semantic similarity between sentences and DE parameters selection strategy. Results are evaluated on the standard DUC datasets using ROUGE-2 measure and significant improvements of \(15.5\%\) and \(4.56\%\) are attained by the proposed approach over the two exiting techniques.

[1]  Matt J. Kusner,et al.  From Word Embeddings To Document Distances , 2015, ICML.

[2]  Kalyanmoy Deb,et al.  A fast and elitist multiobjective genetic algorithm: NSGA-II , 2002, IEEE Trans. Evol. Comput..

[3]  Pushpak Bhattacharyya,et al.  Extractive single document summarization using multi-objective optimization: Exploring self-organized differential evolution, grey wolf optimizer and water cycle algorithm , 2019, Knowl. Based Syst..

[4]  Sriparna Saha,et al.  Multiobjective-Based Approach for Microblog Summarization , 2019, IEEE Transactions on Computational Social Systems.

[5]  Rasim M. Alguliyev,et al.  Formulation of document summarization as a 0-1 nonlinear programming problem , 2013, Comput. Ind. Eng..

[6]  Rasim M. Alguliyev,et al.  Multiple documents summarization based on evolutionary optimization algorithm , 2013, Expert Syst. Appl..

[7]  Rasim M. Alguliyev,et al.  CDDS: Constraint-driven document summarization models , 2013, Expert Syst. Appl..

[8]  Rasim M. Alguliyev,et al.  DESAMC+DocSum: Differential evolution with self-adaptive mutation and crossover parameters for multi-document summarization , 2012, Knowl. Based Syst..

[9]  Enrique Herrera-Viedma,et al.  A New Memetic Algorithm for Multi-document Summarization Based on CHC Algorithm and Greedy Search , 2014, MICAI.

[10]  Minrui Fei,et al.  A Modified Binary Differential Evolution Algorithm , 2010, LSMS/ICSEE.

[11]  Sriparna Saha,et al.  Textual Entailment--Based Figure Summarization for Biomedical Articles , 2020, ACM Trans. Multim. Comput. Commun. Appl..

[12]  Jeffrey Dean,et al.  Efficient Estimation of Word Representations in Vector Space , 2013, ICLR.

[13]  Qingfu Zhang,et al.  Differential Evolution With Composite Trial Vector Generation Strategies and Control Parameters , 2011, IEEE Transactions on Evolutionary Computation.

[14]  Rasim M. Alguliyev,et al.  Sentence selection for generic document summarization using an adaptive differential evolution algorithm , 2011, Swarm Evol. Comput..