Skip to content

the more I invoke an aws lex bot, the more it's trained ?

0

Does the AWS Lex bot get trained more with increased usage? I am asking because I am encountering inconsistent behavior. I'm working with a Lex bot deployed in two regions (Virginia and Frankfurt). Both bots are deployed identically and contain the same content, making them identical. However, I noticed that when I input a wrong input (neither an equivalent utterance nor a slot), the responses differ: in Virginia, it triggers a specific intent, while in Frankfurt, it triggers a different intent.

I am wondering if the more I use the AWS Lex bot, the more it gets trained. This is particularly relevant since I primarily work in Frankfurt for my implementations as with connect and tests.

asked 2 years ago444 views

1 Answer
6

Training Data Differences: Even though the bots are deployed identically, there could be differences in the training data used for each bot during the initial training phase. Even small differences in the training data can lead to variations in bot behavior.

Regional Differences: AWS services may have slight variations or differences in behavior across different regions. While efforts are made to maintain consistency, there may be subtle differences in how Lex operates in the Virginia and Frankfurt regions.

Configuration or Settings: Check if there are any differences in the configuration or settings between the bots deployed in the Virginia and Frankfurt regions. This includes settings related to intents, slots, utterances, and fulfillment logic.

Language or Locale Differences: Ensure that the language and locale settings are consistent across both bots. Differences in language or locale settings can impact how the bot interprets and responds to user input.

**Versioning: **Verify that both bots are using the same version of the bot configuration and code. If there have been updates or changes made to one bot but not the other, it could result in different behavior.

EXPERT

answered 2 years ago

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.