Notation-based Semantification

The scientific community produces a large number of mathematical papers (approximately 108.000 new papers per year [arX]), which raises the importance of machine based processing of such documents. Unfortunately, the most popular formats in which these papers are found (for instance, LTEX) do not contain much information that would allow the computers to infer the human-understandable knowledge contained within a paper. Since, at this point, changing these formats is not practically possible, the other solution is to add a semantic flavor to the existing documents by translating them into a more suitable format, for instance, Content MathML. The current scientific community publication went through a series of evolutions in search of the best method of writing scientific documents. This process was highly influenced by the invention and spreading of the internet. Scientists understood the necessity of a standard that could help them write and exchange their findings in an efficient way. A lot of effort has gone into translating books into digital documents. The next step in this evolution is translating digital documents into knowledge-rich digital documents. This next step can only happen if a new feasible way of transition appears, which MathSemantifier attempts to become. Ambiguity is one of the main reasons that makes semantification complex. Mathematical documents are not a simple collection of symbols. The main use of these documents emerges only when the intended semantics of a document is accessible. However, humans tend to be lazy in writing down the whole graph, but instead rely on implicit human knowledge to decipher these documents. This is where ambiguity comes into play, when the author relies on the ability of the human to use the context of document in order to pinpoint the actual meaning an expression. Ambiguities can be largely divided into two: structural and idiomatic ambiguities. A simple example that demonstrates the concept of structural ambiguities can be “sin x / 2”. It can mean one of the following: