# Extending Neural Generative Conversational Model using External Knowledge Sources

arXiv: https://arxiv.org/abs/1809.05524 · HTML: https://mlsanity.com/paper/1809.05524/

## Key facts

- Cited by 16 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 14 Sep 2018.

## Built on by (most cited first, 16 of 16)

- [Few-Shot Bot: Prompt-Based Learning for Dialogue Systems](https://mlsanity.com/paper/2110.08118/) (2021)
- [Joint Biomedical Entity and Relation Extraction with Knowledge-Enhanced Collective Inference](https://mlsanity.com/paper/2105.13456/) (2021)
- [Enhanced Knowledge Selection for Grounded Dialogues via Document Semantic Graphs](https://mlsanity.com/paper/2206.07296/) (2022)
- [Human-like informative conversations: Better acknowledgements using conditional mutual information](https://mlsanity.com/paper/2104.07831/) (2021)
- [Generative Subgraph Retrieval for Knowledge Graph-Grounded Dialog Generation](https://mlsanity.com/paper/2410.09350/) (2024)
- [Viola: A Topic Agnostic Generate-and-Rank Dialogue System](https://mlsanity.com/paper/2108.11063/) (2021)
- [Improving Conversational Recommender System via Contextual and Time-Aware Modeling with Less Domain-Specific Knowledge](https://mlsanity.com/paper/2209.11386/) (2022)
- [Converse, Focus and Guess -- Towards Multi-Document Driven Dialogue](https://mlsanity.com/paper/2102.02435/) (2021)
- [Do Encoder Representations of Generative Dialogue Models Encode Sufficient Information about the Task ?](https://mlsanity.com/paper/2106.10622/) (2021)
- [Lexical Knowledge Internalization for Neural Dialog Generation](https://mlsanity.com/paper/2205.01941/) (2022)
- [Mind the Knowledge Gap: A Survey of Knowledge-enhanced Dialogue Systems](https://mlsanity.com/paper/2212.09252/) (2022)
- [Retrieval-Augmented Neural Response Generation Using Logical Reasoning and Relevance Scoring](https://mlsanity.com/paper/2310.13566/) (2023)
- [Unstructured Text Enhanced Open-domain Dialogue System: A Systematic Survey](https://mlsanity.com/paper/2411.09166/) (2024)
- [A Brief Study on the Effects of Training Generative Dialogue Models with a Semantic loss](https://mlsanity.com/paper/2106.10619/) (2021)
- [A Neural Conversation Generation Model via Equivalent Shared Memory Investigation](https://mlsanity.com/paper/2108.09164/) (2021)
- [Information for Conversation Generation: Proposals Utilising Knowledge Graphs](https://mlsanity.com/paper/2410.16196/) (2024)

## Abstract (arXiv)

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.