Towards Exploiting Background Knowledge for Building Conversation Systems

By Nikita Moghe, Siddharth Arora, Suman Banerjee, Mitesh M. Khapra. 2018

174 citations recorded by the source corpus.

Existing dialog datasets contain a sequence of utterances and responses without any explicit background knowledge associated with them. This has resulted in the development of models which treat conversation as a sequence-to-sequence generation task (i.e., given a sequence of utterances generate the response sequence). This is not only an overly simplistic view of conversation but it is also emphatically different from the way humans converse by heavily relying on their background knowledge about the topic (as opposed to simply relying on the previous sequence of utterances). For example, it is common for humans to (involuntarily) produce utterances which are copied or suitably modified from background articles they have read about the topic. To facilitate the development of such natural conversation models which mimic the human process of conversing, we create a new dataset containing movie chats wherein each response is explicitly generated by copying and/or modifying sentences from unstructured background knowledge such as plots, comments and reviews about the movie. We establish baseline results on this dataset (90K utterances from 9K conversations) using three different models: (i) pure generation based models which ignore the background knowledge (ii) generation based models which learn to copy information from the background knowledge when required and (iii) span prediction based models which predict the appropriate response span in the background knowledge.

Subfields: Artificial Intelligence, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Natural Language Processing

Citations in the corpus

Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks (2020, background): “Such tasks include open-domain question answering [5, 25], fact checking [50], fact completion [42], long-form question answering [12], Wikipedia article generation [32], dialogue [36, 59, 9, 13], translation [16], and language modeling [17, 23].”
Wizard of Wikipedia: Knowledge-Powered Conversational agents (2018, background): “and Transformers which have been shown to be on-par or superior to RNN encoder-decoders (Vaswani et al., 2017), and develops an architecture that combines these approaches. Concurrently with our work Moghe et al. (2018) proposed a dataset based on the closed domain of movie chats. Our paper shows mo”
Can You Put it All Together: Evaluating Conversational Agents’ Ability to Blend Skills (2020, background): “…gauging and improving performance of open-domain conversational agents along speciﬁc axes such as how knowledgeable they are (Dinan et al., 2019b; Moghe et al., 2018; Qin et al., 2019), how well they can display empathy (Rashkin et al., 2019; Lin et al., 2019) or talk about their personal…”
The Dialogue Dodecathlon: Open-Domain Knowledge and Image Grounded Conversational Agents (2019, background): “For example, there are additional knowledge tasks (Qin et al., 2019; Moghe et al., 2018; Gopalakrishnan et al., 2019) and image-based datasets (Das et al., 2017) one could use.”
Challenges in Building Intelligent Open-domain Dialog Systems (2019, background): “Recently, there is a significant burst in constructing document or knowledge grounded dialog corpora [17, 25, 40, 86, 87, 104, 156, 191], which will be described in Section 7 in details.”
Conversing by Reading: Contentful Neural Conversation with On-demand Machine Reading (2019, methodology): “Moghe et al. (2018) explored directly using a span-prediction QA model for conversation.”
Knowledge-Grounded Dialogue Generation with Term-level De-noising (2021, methodology): “Holl-E Moghe et al. (2018) also made use of MTurk workers to create an annotated dataset that focuses on movies as two workers talk with each other about a chosen movie.”
Sequential Latent Knowledge Selection for Knowledge-Grounded Dialogue (2020, methodology): “In our experiments, we also evaluate on Holl-E (Moghe et al., 2018) as another dataset for knowledge-grounded dialogue, after collecting clearer labels of knowledge sentences.”
TicketTalk: Toward human-level performance with end-to-end, transaction-based dialog systems (2020, methodology): “The TicketTalk movie ticketing dataset was created using the self-dialog collection method (Krause et al., 2017; Moghe et al., 2018; Byrne et al., 2019) in which a paid crowd-sourced worker writes both sides of the dialog (i.e. both customer and ticketing agent turns) based on a particular scenario…”
A Controllable Model of Grounded Response Generation (2020, background): “Moghe et al. (2018) and Zhou, Prabhumoye, and Black (2018) collect movie discussion datasets via crowdsourcing.”
Plug-and-Play Conversational Models (2020, background): “, 2019), knowledge grounding (Dinan et al., 2018; Gopalakrishnan et al., 2019; Ghazvininejad et al., 2018; Moghe et al., 2018; Wu et al., 2020), emotions (Li et al.”
Approximation of Response Knowledge Retrieval in Knowledge-grounded Dialogue Generation (2020, background): “Holl-E (Moghe et al., 2018) comprises dialogues between two Amazon Mturk workers 2 about a selected movie, supported by selected sources of background knowledge: movie plots, reviews, comments, and the fact tables related to the movie.”
Exploring Prompt-based Few-shot Learning for Grounded Dialog Generation (2021, background): “Since these data is usually not directly available, existing GDG datasets are mostly collected via crowdsourcing (Zhang et al., 2018; Moghe et al., 2018; Dinan et al., 2018; Gopalakrishnan et al., 2019; Liu et al., 2021b), which is costly and time-consuming.”
Knowledge Aware Conversation Generation with Explainable Reasoning over Augmented Graphs (2019, methodology): “We implement the knowledge aware generation model based on the code of GTTP 4 released by (Moghe et al., 2018).”
A Survey of Knowledge-enhanced Text Generation (2020, background): “Background Based Conversation (BBC, also known as grounded conversation), as an emerging topic, has been proposed for generating informative responses that are grounded on one or multiple background documents [11, 100].”
ProsocialDialog: A Prosocial Backbone for Conversational Agents (2022, background): “Holl-E is also a knowledge-grounded dialogue dataset about various movie information (e.g., plots, comments, reviews; Moghe et al., 2018).”
On Knowledge distillation from complex networks for response prediction (2019, methodology): “Such complex span prediction models are also being adapted for other NLP tasks such as dialog response prediction (Moghe et al., 2018), which is the focus of this work.”
Topical-Chat: Towards Knowledge-Grounded Open-Domain Conversations (2019, background): “[7] released a similar dataset of conversations about movies, where the knowledge base comprises Wikipedia articles, reviews and comments mined from the web about ∼1K movies.”