Any theory of speech recognition must account for both data-driven and conceptually driven processing.

Question

Quizplus · Accepted Answer

Both data-driven (bottom-up) and conceptually driven (top-down) processing are necessary for speech recognition. Data-driven processing involves the analysis of incoming acoustic signals, while conceptually driven processing involves using knowledge and context to inform the interpretation of those signals.

Any Theory of Speech Recognition Must Account for Both Data-Driven