Citation: Daggumati, S.; Revesz, P.Z. Convolutional Neural Networks Analysis Reveals Three Possible Sources of Bronze Age Writings between Greece and India. Information 2023, 14, 227. https:// doi.org/10.3390/info14040227 Academic Editor: Xin Ning Received: 7 February 2023 Revised: 4 April 2023 Accepted: 4 April 2023 Published: 7 April 2023 Copyright: © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). information Article Convolutional Neural Networks Analysis Reveals Three Possible Sources of Bronze Age Writings between Greece and India † Shruti Daggumati and Peter Z. Revesz * School of Computing, College of Engineering, University of Nebraska-Lincoln, Lincoln, NE 68588, USA; sdaggumati@unl.edu (S.D.); revesz@cse.unl.edu (P.Z.R.) * Correspondence: revesz@cse.unl.edu; Tel.: +1-402-421-6990 † This paper is an extended version of our paper published in the 23rd International Database Engineering and Applications Symposium, IDEAS 2019, Athens, Greece, 10–12 June 2019. Abstract: This paper analyzes the relationships among eight ancient scripts from between Greece and India. We used convolutional neural networks combined with support vector machines to give a numerical rating of the similarity between pairs of signs (one sign from each of two different scripts). Two scripts that had a one-to-one matching of their signs were determined to be related. The result of the analysis is the finding of the following three groups, which are listed in chronological order: (1) Sumerian pictograms, the Indus Valley script, and the proto-Elamite script; (2) Cretan hieroglyphs and Linear B; and (3) the Phoenician, Greek, and Brahmi alphabets. Based on their geographic locations and times of appearance, Group (1) may originate from Mesopotamia in the early Bronze Age, Group (2) may originate from Europe in the middle Bronze Age, and Group (3) may originate from the Sinai Peninsula in the late Bronze Age. Keywords: classification; epigraphy; neural networks; script family; support vector machine 1. Introduction In this paper, we use data mining methods to analyze the relationships among eight Bronze Age scripts from between Greece and India, namely the Brahmi script [1], Cretan hieroglyphs [2], the Greek alphabet [3], the Indus Valley script [4–8], the Linear B syl- labary [9], the Phoenician alphabet [10–12], the proto-Elamite script [13,14], and Sumerian pictographs [15]. We are interested in testing the hypothesis that these eight scripts had a single origin. This is probable given that the eight scripts originate from geographic locations along an east–west line between India and Greece, as is shown in Figure 1. We are going to test this hypothesis by applying data mining to the scripts. The data mining method that we have chosen for this study is a convolutional neural networks analysis. Convolutional neural networks have previously been applied to the recognition of various signs, including alphabets, but they have not been used in a multiscript analysis. The novel idea in our approach is to first train separate convolutional neural networks to recognize various scripts (see Section 5.1 for a review of works that are related to this first phase). Then, in the second phase, we pass one script’s signs into another’s convolutional neural network. The sign ‘recognized’ by the convolutional neural network can be considered the closest to the input sign. If the two scripts are related to each other, then a one-to-one mapping may be found between the signs of the two scripts. If the two scripts are not related to each other, then there will be no one-to-one mapping. Our study is motivated by a desire to contribute to the decipherment of ancient, Bronze Age scripts, especially the Indus Valley script [8]. Decipherment can be greatly facilitated by understanding the precise relationships among these ancient scripts. A one-to-one Information 2023, 14, 227. https://doi.org/10.3390/info14040227 https://www.mdpi.com/journal/information