October 31, 2018
The use of connectionist approaches in conversational agents has been progressing rapidly due to the availability of large corpora. However current generative dialogue models often lack coherence and are content poor. This work proposes an architecture to incorporate unstructured knowledge sources to enhance the next utterance prediction in chit-chat type of generative dialogue models. We focus on Sequence-to-Sequence (Seq2Seq) conversational agents trained with the Reddit News dataset, and consider incorporating external knowledge from Wikipedia summaries as well as from the NELL knowledge base. Our experiments show faster training time and improved perplexity when leveraging external knowledge.
Publisher
EMNLP
December 07, 2023
Hakan Inan, Kartikeya Upasani, Jianfeng Chi, Rashi Rungta, Krithika Iyer, Yuning Mao, Davide Testuggine, Madian Khabsa
December 07, 2023
November 06, 2023
David Esiobu, Ellen Tan, Saghar Hosseini, Megan Ung, Yuchen Zhang, Jude Fernandes, Jane Yu, Eleonora Presani, Adina Williams, Eric Smith
November 06, 2023
October 27, 2023
Davis Liang, Hila Gonen, Yuning Mao, Rui Hou, Naman Goyal, Marjan Ghazvininejad, Luke Zettlemoyer, Madian Khabsa
October 27, 2023
August 06, 2023
Leo Adolphs, Tianyu Gao, Jing Xu, Kurt Shuster, Sainbayar Sukhbaatar, Jason Weston
August 06, 2023
Product experiences
Foundational models
Product experiences
Latest news
Foundational models