CGSC170 Study Guide - Midterm Guide: Signal-To-Noise Ratio, Vocal Tract, Vocal Folds

• School: University of Delaware
• Department: Cognitive Science
• Course: CGSC170
• Professor: Andreasen Robin

Document Summary

It is presented with written letters of the alphabet. Its output is the correct pronunciation of the sounds represented by the letters. What is noisy or ambiguous about the input data: a: difficult because: for most letters, there is more than one phoneme associated with that letter. Input, output & layers: input, hidden, and output. Input: the system is presented with a text seven letters at a time. Reads the text one letter at a time. For each pass, it focuses on the middle (4th) letter. Output: systems interpretation of phoneme associated with 4th letter in the 7-letter stream. Articulatory features: properties of speech sound characterized in terms of the interaction of various physiological structures (such as tongue, vocal tract, and lips): ex. Bilabial: temporarily blacks air flow by bringing lips together (p, b, m: ex. Stop: complete blockage of air flow followed by sudden release of air. Voiced: vocal cords vibrate (m, b: ex.