ALOHA From New York, GBDA Student Helps Advance Conversational AI Systems

Tuesday, May 19, 2020

Julia Sprague confrence standing beside project summary

ALOHA From New York, GBDA Student Helps Advance Conversational AI Systems

Global Business and Digital Arts student Julia Sprague, alongside her Computation Health and Informatics Lab (CHIL) team, recently celebrated the publication of their paper, “ALOHA: Artificial Learning of Human Attributes for Dialogue Agents.” Led by Professor Jesse Hoey, the team’s paper was published in Google Scholar and accepted to the Association for the Advancement of Artificial Intelligence (AAAI) conference in New York, where the group presented their findings in February.

"Our paper was about using movie data to make conversational AI systems converse more like humans (have personality and emotion) so that they can establish a greater connection with individuals," said Sprague in an email.

By combining detailed Human Level Attribute (HLA) data that were based on tropes–characteristics of fictional personalities that are observed recurrently and determined by viewers' impressions–with dialogue data for specific characters, we present a dataset that models character profiles and gives dialogue agents the ability to learn characters' language styles through their Human Level Attributes.

The AAAI conference is one of the top Artificial Intelligence (AI) conferences in the world. Of the over 8,000 papers submitted, Sprague's was one of 1,500 accepted. The team had the opportunity to present in front of industry, researchers and AI enthusiasts at the conference in February.

This achievement shows current and future Global Business and Digital Arts Students that they can contribute to highly technical projects, Sprague said. It shows the "breadth of what you are able to do with our degree and that the possibilities are endless." 

Link to paper:https://arxiv.org/abs/1910.08293

  1. 2020 (14)
    1. October (2)
    2. September (3)
    3. June (3)
    4. May (3)
    5. April (1)
    6. February (2)
  2. 2019 (6)
    1. November (1)
    2. September (2)
    3. June (1)
    4. April (1)
    5. January (1)
  3. 2018 (7)
    1. December (1)
    2. September (1)
    3. May (2)
    4. April (2)
    5. January (1)
  4. 2017 (13)
  5. 2016 (17)
  6. 2015 (8)
  7. 2014 (23)
  8. 2013 (37)
  9. 2012 (9)