Table IV summarizes the performance of the stack decoding algorithm with a segmenting and a time-synchronous acoustic processor on three tasks of varying perplexity. The Raleigh task has been described earlier in the paper. The Laser task i…[14] L. R. Bahl and F. Jelinek, "Decoding for channels with insertions, deletions, and substitutions with applications to speech recognition," IEEE Trans. Inform. Theory, vol. IT-21, pp. 404-411, July 1975.
[15] F. Jelinek and R. L. Mercer…
Raimo Bakis
engineer · 2 mentions across 1 reading
In this course
Bakis appears as a co-author on foundational work in speech recognition and decoding algorithms, particularly within the IBM research sphere where stack decoding and acoustic modeling were being developed in the 1970s. His citation in these excerpts situates him within the early machine learning pipeline—the technical apparatus for converting acoustic signals into language—which the course frames as a crucial precursor to contemporary AI systems' ability to process and interpret human communication. This work exemplifies how cybernetic feedback loops between signal processing and probabilistic language models became embedded in the infrastructure of AI long before "machine learning" became a household term.
Mentioned in 1 reading
Appears alongside
People mentioned in the same passages — sorted by co-occurrence weight.