Case-Based Translation: First Steps from a Knowledge-Light Approach Based on Analogy to a Knowledge-Intensive One

Abstract

This paper deals with case-based machine translation. It is based on a previous work using a proportional analogy on strings, i.e., a quaternary relation expressing that "String A is to string B as string C is to string D". The first contribution of this paper is the rewording of this work in terms of case-based reasoning: a case is a problem-solution pair (A, A) where A is a sentence in an origin language and A , its translation in the destination language. First, three cases (A, A), (B, B), (C, C) such that "A is to B as C is to the target problem D" are retrieved. Then, the analogical equation in the destination language "A is to B as C is to x" is solved and D = x is a suggested translation of D. Although it does not involve any linguistic knowledge, this approach was effective and gave competitive results at the time it was proposed. The second contribution of this work aims at examining how this prior knowledge-light case-based machine translation approach could be improved by using additional pieces of knowledge associated with cases, domain knowledge, retrieval knowledge, and adaptation knowledge, and other principles or techniques from case-based reasoning and natural language processing.

The French Correction: When Retrieval Is Harder to Specify than Adaptation

“…The rst application of the analogy on strings dened in [4] is case-based machine translation: a case is a pair (x s , y s ), where x s is a sentence in a natural language and y s is a translation of x s in another natural language. The approach proposed in [4] and studied in [5] at the light of the CBR methodology consists rst in nding 3 source cases (x a , y a ), (x b , y b ) and (x c , y c ) such that x a :x b ::x c :x tgt holds and then in solving the analogical equation y a :y b ::y c :y: a solution y of this equation is a candidate solution y tgt of x tgt . Now, this idea could be reused for TFC: this would mean that the problems would be sentences in an incorrect French language and the solutions would be sentences in a correct French language.…”

An Approach to Case-Based Reasoning Based on Local Enrichment of the Case Base

“…An analogy on S can be defined likewise. Therefore, using the problem variation space and the solution variation space, an analogy on P and an analogy on S can be built, and thus, the approach to CBR based on the following principle (called extrapolation in [13] and used in [12] and [10]) can be applied:…”

“…However, the principle of the approach still holds. These applications show how case-based translation [10] can be actually performed thanks to the approach of local enrichment of the case base. The results are demonstrative and promising.…”