System and method for argument retrieval

A method for identifying in a corpus of documents at least one argument relevant to an identified topic, comprising:

• identifying a plurality of topic-related sentences relevant to the identified topic in a plurality of documents of the corpus of documents;
• producing a plurality of synthetic documents, each comprising a plurality of sentences wherein a first sentence of the plurality of sentences of each respective synthetic document is a respective one of the identified plurality of topic-related sentences, extracted and copied from a respective one of the plurality of documents and wherein other sentences of the plurality of sentences of said each respective synthetic document are an identified amount of other sentences extracted from the respective document from which the topic-related sentence was extracted;
• identifying t least one synthetic document, among the plurality of produced synthetic documents, containing an argument relevant to the identified topic, by analyzing each of the plurality of synthetic documents by at least one machine learning model trained to identify an argument in response to an inputted document;
• and
• outputting the at least one identified synthetic document;
• wherein said at least one argument is a structure of language which comprises one or more premises, a claim, and a conclusion.

An argument generation system, comprising at least one hardware processor adapted to:

• receiving an identified topic;
• accessing a corpus of documents;
• identifying in the corpus of documents at least one argument relevant to the identified topic by:
• identifying a plurality of topic-related sentences relevant to the identified topic in a plurality of documents of the corpus of documents;
• producing a plurality of synthetic documents, each comprising a plurality of sentences, wherein a first sentence of the plurality of sentences of each respective synthetic document is a respective one of the identified plurality of topic-related sentences, extracted and copied from a respective one of the plurality of documents and wherein other sentences of the plurality of sentences of said each respective synthetic document are an identified amount of other sentences extracted from the respective document from which the topic-related sentence was extracted;
• identifying at least one synthetic document, among the plurality of produced synthetic documents, containing an argument relevant to the identified topic, by analyzing each of the plurality of synthetic documents by at least one machine learning model trained to identify an argument in response to an inputted document;
• and
• outputting the at least one identified synthetic document;
• and
• providing the at least one identified synthetic document to at least one software object executed by the at least one hardware processor for the purpose of performing at least one argument related task;
• wherein said at least one argument is a structure of language which comprises one or more premises, a claim, and a conclusion.