- Listening Chat System End Design nishio20190729_0218
1> Let’s talk. 0> What would you like to see happen in this conversation? 1> I would like to clarify how to proceed after a user enters “I see”. 0> Is there anything else about that user? 1> Users are looking for “I see”. 0> What type of user is this user? 1> First listen. 0> Yes 1> After typing “I see”, confirm with “What just happened?” and then, regardless of the answer, the next question is “Is this a good place to end? If the answer is “Yes”, the system will return to the original mode with “What would you like to see happen now? 0> Mode 1> I feel like this is the right thing to do. 0> qi 1> The end 0> What would you like to see happen in this conversation? 1> I wanted to clarify how to proceed after a user enters “I see”, but I think I’ve already clarified. 0> How do you know the user? 1> The good flow can now be detected by the user entering “I see”, but the bad flow has not been detected. 0> Is there anything else about that process? 1> Now “NGKW” is a command to delete target keywords, and there is a separate “NG” which is ambiguous in meaning, and the person who created it does not understand the meaning of the command correctly after 5 years since he created it, maybe he should get this right. 0> What type of command is that? 1> A command for a human to ambiguously express “not good enough”. Maybe it’s just a marker for later machine learning and has no function as a command, or maybe it has a function to remove the bot from the target to be read and transformed by the bot to note what it thought was bad separately from the story. 0> Is there anything else about that command? 1> Maybe there is a function to reduce the score of a target keyword in a one-off, or maybe we should detect the deterioration of the user’s mental state by the accumulation of “NG has occurred many times”. 0> What do you value? 1> Interesting exchanges should continue. We want to increase the density of interesting exchanges by allowing uninteresting exchanges to be easily terminated. 0> What is the relationship between that interesting exchange and the user? 1> The user’s state of mind improves when interesting exchanges continue and new ideas are generated. 0> Is that user the same as an interesting exchange? Is it different? 1> No. 0> What happens in an interesting exchange when a user 1> NG 0> What happens to the user during an interesting exchange? 1> The user’s brain is activated by the stimulus of the exchange. Given a small unit task of a question, a high level of concentration on it occurs, and the user enters a flow state. 0> What is between that user and an interesting exchange? 1> There’s a period of aiding and abetting before the interesting exchange switches on. 0> What is the relationship between that aid and the user? 1> The user’s mental state changes during the helping period. 0> What is the relationship between its users and their mental state? 1> Good mental state is necessary for the device, the user, to be used efficiently 0> What is the relationship between the user and the device? 1> Same 0> Is a user the same as a mental state? Is it different? 1> No. 0> What happens to your mental state when you are a user? 1> NG 0> What happens to the user when in a mental state? 1> NG 0> What is the connection between the user and the mental state? 1> (What would be… the train arrived at the office. 1> Bye-bye.
This page is auto-translated from /nishio/「なるほど」の後の進め方 using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.