Filters








3 Hits in 2.0 sec

BeliefBank: Adding Memory to a Pre-Trained Language Model for a Systematic Notion of Belief [article]

Nora Kassner, Oyvind Tafjord, Hinrich Schütze, Peter Clark
2021 arXiv   pre-print
Our approach is to embed a PTLM in a broader system that also includes an evolving, symbolic memory of beliefs -- a BeliefBank -- that records but then may modify the raw PTLM answers.  ...  This is significant as it is a first step towards PTLM-based architectures with a systematic notion of belief, enabling them to construct a more coherent picture of the world, and improve over time without  ...  The authors of this work take full responsibility for its content.  ... 
arXiv:2109.14723v1 fatcat:b5wgdj66gbb5zk3zpfkb4fi3q4

BeliefBank: Adding Memory to a Pre-Trained Language Model for a Systematic Notion of Belief

Nora Kassner, Oyvind Tafjord, Hinrich Schütze, Peter Clark
2021 Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing   unpublished
Our approach is to embed a PTLM in a broader system that also includes an evolving, symbolic memory of beliefs -a BeliefBank -that records but then may modify the raw PTLM answers.  ...  This is significant as it is a first step towards PTLM-based architectures with a systematic notion of belief, enabling them to construct a more coherent picture of the world, and improve over time without  ...  The authors of this work take full responsibility for its content.  ... 
doi:10.18653/v1/2021.emnlp-main.697 fatcat:e2vauuoc6nehri54i75vtag574

Enriching a Model's Notion of Belief using a Persistent Memory [article]

Nora Kassner, Oyvind Tafjord, Hinrich Schutze, Peter Clark
2021 arXiv   pre-print
Our approach is to add a memory component -- a BeliefBank -- that records a model's answers, and two mechanisms that use it to improve consistency among beliefs.  ...  This is significant as it is a first step towards endowing models with an evolving memory, allowing them to construct a more coherent picture of the world.  ...  Introduction How might we ascribe a notion of belief to a model?  ... 
arXiv:2104.08401v2 fatcat:lm46ijzvbrcphbhxoc3krpcq5y