Integrating CAT and MT in AnglaBhart-II architecture.

摘要: Machine translation (MT) is a complex and difficult task. It not possible to achieve human competing performance with the present state of technology. Automating process natural languages requires number knowledge sources their appropriate invocation in engine. A practical machine system limited resources cannot embody all that beings use. However, an MT can be considerably improved if automated integrated supporting modules provide synergy for arriving at correct translations. The computer assisted tools (CAT) identify limitations clues cope up them, constitute important module enhancing performance. This paper presents details AnglaBharti-II architecture highlighting role CAT system. translating English Indian languages. AnglaBharti primarily rule-based (RBMT). input sentence transformed pseudo-interlingual structure called PLIL (Pseudo Lingua Languages) using CFG like pattern directed rule-base. RBMT its own dealing real-life texts. We have tried overcome some these by integrating additional modules. These are basically incorporating memory, raw generalized example-bases, interactive pre-editing, paraphrasing, failure analysis heuristics attempt deal variety constructs frequently encountered real life text.