meta data for this page
Differences
This shows you the differences between two versions of the page.
| Both sides previous revision Previous revision | |||
|
mimicry_of_prompt [2018/08/21 11:30] 127.0.0.1 external edit |
mimicry_of_prompt [2019/08/08 13:17] (current) lisa.illgen_concentrix.com Added Anchor Links |
||
|---|---|---|---|
| Line 1: | Line 1: | ||
| ==== Mimicry of Prompts ==== | ==== Mimicry of Prompts ==== | ||
| **// When designing grammars, take advantage of the tendency of callers to mimic the prompt they just heard //**\\ | **// When designing grammars, take advantage of the tendency of callers to mimic the prompt they just heard //**\\ | ||
| - | Designers and researchers have noted the tendency of callers to mimic the wording of prompts (either initial prompts or the example prompting that supports open-ended prompts for systems using natural language understanding (NLU) and statistical language modeling (SLM)). For example, Knott, Bushey, and Martin (2004) reported that participants in their experiment had a strong tendency to mimic the examples presented in the prompts. From a prompt design perspective, this underscores the importance of selecting highly relevant examples and using optimal wording in prompts. From the grammar design perspective, this illustrates the need to ensure that grammars (whether FSG or SLM -- see [[Directed Dialog vs. SLM]]) are able to understand what callers are likely to say after having heard a prompt. | + | Designers and researchers have noted the tendency of callers to mimic the wording of prompts (either initial prompts or the example prompting that supports open-ended prompts for systems using natural language understanding (NLU) and statistical language modeling (SLM)). For example, [[references#knott|Knott, Bushey, and Martin]] (2004) reported that participants in their experiment had a strong tendency to mimic the examples presented in the prompts. From a prompt design perspective, this underscores the importance of selecting highly relevant examples and using optimal wording in prompts. From the grammar design perspective, this illustrates the need to ensure that grammars (whether FSG or SLM -- see [[Directed Dialog vs. SLM]]) are able to understand what callers are likely to say after having heard a prompt. |
| ==== References ==== | ==== References ==== | ||
| Knott, B. A., Bushey, R. R., & Martin, J. M. (2004). Natural language prompts for an automated call router: Examples increase the clarity of user responses. In Proceedings of the Human Factors and Ergonomics Society 48th annual meeting (pp. 736–739). Santa Monica, CA: Human Factors and Ergonomics Society. | Knott, B. A., Bushey, R. R., & Martin, J. M. (2004). Natural language prompts for an automated call router: Examples increase the clarity of user responses. In Proceedings of the Human Factors and Ergonomics Society 48th annual meeting (pp. 736–739). Santa Monica, CA: Human Factors and Ergonomics Society. | ||