Skip to main content

IBM Brazil Researcher Heloisa Candello gives Guest Lecture on User Evaluation of Multi-party Conversational Systems

Posted November 9, 2018
Heloisa Candello presenting at RPI
CISL hosts a talk by IBM Brazil Researcher Heloisa Candello on her research in user evaluation of multi-party conversational systems.

Recent advances in artificial intelligence, natural language processing, and mobile computing, together with the rising popularity of chat and messaging environments, have enabled a boom in the deployment of interactive systems based on conversation and dialogue. This talk explores the design and evaluation of conversational interfaces, and it is focused on design and evaluation methods which address specific challenges of interfaces based on multi-party dialogue. I will show two projects. Café com os Santiagos, an artwork where visitors conversed with three chatbots portraying characters from a book in a scenographic space recreating a 19th-century coffee table. It was accessed by more than 10.000 users in a public space resulting in insights to improve even more the conversation system field.  And an experiment with a cognitive investment adviser called Finch. Finch interface was able to make a state-of-art artificial conversational governance system accessible for regular users to assist in financial decisions. 

 

Speaker: Heloisa Candello, Ph.D. is a researcher and an interaction designer at the Conversational Intelligence group at IBM Research laboratory in Brazil. She holds a PhD. in Computer Science, Interactive Technologies from University of Brighton – UK. Prior to joining IBM, Heloisa was a lecturer and a researcher in mobile interaction technologies. Nowadays, at IBM, Heloisa led and designed novel and extraordinary user experiences and research in the context of multi-bot systems. Her research resulted in several publications in leading conferences (CHI, CSCW, DRS, DUXU) and recognition in the HCI and Design field.  Heloisa was also a co-chair of several workshops and conferences in this field. https://researcher.watson.ibm.com/researcher/view.php?person=br-heloisacandelloAbstract