Virtual assistants this sort of as Siri or Alexa are an inseparable section of our lives these days. Even so, they continue to struggle with some utterances. For occasion, digital assistants can’t reply to humorous requests, like a person asking to invest in a mind. A recent paper suggests utilizing scientific theories of humor to characterize these requests.
Picture credit history: Y2kcrazyjoker4 by way of Wikimedia (CC BY-SA four.)
A taxonomy of playful utterance kinds, driven by genuine-world info from Alexa, was introduced. The researchers centered on a unique section of taxonomy, the personification of the digital assistant as a human, for a person study. The design routinely identified personifying utterances. Manual answers were created if the design detected utterances that can’t be resolved by predefined answers. It was proven that the person encounter is enhanced when the digital assistant understands their playful utterances and responds accordingly.
Virtual assistants this sort of as Amazon’s Alexa, Apple’s Siri, Google Home, and Microsoft’s Cortana, are turning out to be ubiquitous in our daily lives and productively assist consumers in many daily duties, this sort of as building cellphone calls or taking part in music. Nonetheless, they continue to struggle with playful utterances, which are not meant to be interpreted virtually. Illustrations include things like jokes or absurd requests or issues this sort of as, “Are you scared of the dim?”, “Who enable the puppies out?”, or “Order a zillion gummy bears”. Today, digital assistants often return irrelevant answers to this sort of utterances, other than for tough-coded types resolved by canned replies.
To deal with the problem of routinely detecting playful utterances, we very first characterize the unique kinds of playful human-digital assistant interaction. We introduce a taxonomy of playful requests rooted in theories of humor and refined by examining genuine-world targeted visitors from Alexa. We then emphasis on 1 node, personification, the place consumers refer to the digital assistant as a human being (“What do you do for enjoyable?”). Our conjecture is that comprehending this sort of utterances will improve person encounter with digital assistants. We executed a Wizard-of-Oz person study and confirmed that endowing digital assistant s with the ability to detect humorous possibilities certainly has the potential to increase person fulfillment. We hope this perform will lead to the comprehending of the landscape of the challenge and inspire novel suggestions and methods in direction of the eyesight of providing digital assistants a sense of humor.
Research paper: Shani, C., Libov, A., Tolmach, S., Lewin-Eytan, L., Maarek, Y., and Shahaf, D., “”Alexa, what do you do for enjoyable?” Characterizing playful requests with digital assistants”, 2021. Hyperlink: https://arxiv.org/stomach muscles/2105.05571