A fast, scalable and memory-efficient method for static decoding graph construction is presented. As an alternative to the traditional transducer-based approach, it is based on incremental composition. Memory efficiency is achieved by combining composition, determinization and minimization into a single step, thus eliminating large intermediate graphs. We have previously reported the use of incremental composition limited to grammars and left cross-word context [1]. Here, this approach is extended to n-gram models with explicit å arcs and right cross-word context.

s/h4> M. Novak and V. Bergl, “Memory efficient decoding graph compilation with wide cross-word acoustic context”, In Proceedings of Interspeech 2004, 281-284, Seul, South Korea, 2004.