OSU TacoBot Team won the third-place honor in the inaugural Alexa Prize TaskBot Challenge
CSE is very delighted to announce that the OSU TacoBot team led by Prof. Huan Sun won third place in the first Alexa Prize TaskBot Challenge. The team consists of six students from Prof. Sun’s research group, two from Prof. Yu Su’s and one from Prof. Mike White’s.
According to Amazon Science and Prof. Sun, the TaskBot Challenge was launched in March 2021 and selected ten university teams to develop bots that assist customers with multi-step cooking or do-it-yourself (DIY) home improvement tasks. Success requires the teams to address many difficult artificial intelligence (AI) challenges, from knowledge representation and inference, and commonsense and causal reasoning, to language understanding and generation.
Only five teams advanced to the finals and OSU TacoBot is the only US team among the top-3 performers. The team will receive a $50,000 prize for the honor.
“Our developed bot was deployed in the Alexa virtual assistant to interact with real customers and receive feedback from them, ”said Prof. Sun. She has been dedicated to developing natural language interfaces and conversational AI systems in various domains. “It is a very meaningful process with real impact and we learnt a lot from this challenge.”
“We are excited about the possibility that future TaskBots can closely collaborate with users to address their needs and questions via knowledgeable, contextualized, and engaging multi-modal interactions,” said Shijie Chen, one student lead. “We believe that TaskBots can potentially make technologies more accessible. Instead of browsing through many web pages, users can have a multi-modal conversation with TaskBots to find instructions for their ongoing tasks and answers to their questions. With the upcoming SimBot challenge, it’d be even more exciting if the TaskBot could be embodied in a moving robot that executes complex tasks in the environment for users.” Ron Chen, the other student lead.
“The development of Conversational AI technologies is at a tipping point. On the one hand, there is growing consensus on the potential of conversational AI as the next-generation universal human-machine interface. On the other hand, the current technologies are still very limited in their coverage, contextualizability, and adaptivity. The fast feedback loop with real users provided by the Alexa Prize challenges is an important driver for developing next-generation conversational AI technologies.” Said Prof. Su, faculty co-advisor for the TacoBot team and faculty advisor for the SalsaBot team in the upcoming SimBot challenge. (Yes, TacoBot and SalsaBot will be two sister bots from OSU!)
More information can be found here.