On memory limitations in natural language processing

摘要: This paper proposes a welcome hypothesis: computationally simple device is sufficient for processing natural language. Traditionally it has been argued that language syntax requires very powerful machinery. Many engineers have come to this rather grim conclusion; almost all working parsers are actually Turing Machines (TM). For example, Woods specifically designed his Augmented Transition Networks (ATN''s) be Equivalent. If the problem really as hard appears, then only solution grin and bear it. Our own position parsing acceptable sentences simpler because there constraints on human performance drastically reduce computational requirements (time space bounds). Although ideal linguistic competence complex, observation may not apply directly real such parsing. By including factors, possible simplify computation. We will propose two limitations, bounded memory deterministic control, which incorporated in new parser YAP.