EMNLP Findings2023
Context Quality Matters in Training Fusion-in-Decoder for Extractive Open-Domain Question Answering
Abstract
Retrieval-augmented generation models augment knowledge encoded in a language model by providing additional relevant external knowledge (context) during generation. Although it has been shown that the quantity and quality of context impact the performance of retrieval-augmented generation models during inference, limited research explores how these characteristics affect model training. This paper explores how context quantity and quality during model training affect the performance of Fusion-in-Decoder (FiD), the state-of-the-art retrieval-augmented generation model, in extractive open-domain question answering tasks. Experimental results suggest that FiD models overfit to context quality during training and show suboptimal performance when evaluated on different context quality. Through the experimental results, we also reveal FiD models trained with different context quality have different cross-attention distribution patterns. Specifically, as context quality during training increases, FiD models tend to attend more uniformly to each passage in context. Finally, based on these observations, we propose a method to mitigate overfitting to specific context quality by introducing bias to the cross-attention distribution, which we demonstrate to be effective in improving the performance of FiD models on different context quality.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
Lineage
- Improving Language Models by Retrieving from Trillions of Tokens cites Retrieval Augmentation Reduces Hallucination in Conversation
- Improving Language Models by Retrieving from Trillions of Tokens cites Language Models are Few-Shot Learners
- Improving Language Models by Retrieving from Trillions of Tokens cites Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
- When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memories cites Improving Language Models by Retrieving from Trillions of Tokens
- When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memories cites Retrieval Augmentation Reduces Hallucination in Conversation
- When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memories cites Language Models are Few-Shot Learners
- When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memories cites Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
- Retrieval Augmentation Reduces Hallucination in Conversation cites Language Models are Few-Shot Learners
- Retrieval Augmentation Reduces Hallucination in Conversation cites Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
- Language Models are Few-Shot Learners cites Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
Builds on16
- Language Models are Few-Shot LearnersTom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah et al.NeurIPS 2020 · 64,255 citations
- Retrieval-Augmented Generation for Knowledge-Intensive NLP TasksPatrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni et al.NeurIPS 2020 · 19,162 citations
- Improving Language Models by Retrieving from Trillions of TokensSebastian Borgeaud, Arthur Mensch, Jordan Hoffmann, Trevor Cai et al.ICML 2022 · 1,629 citations
- Retrieval Augmentation Reduces Hallucination in ConversationKurt Shuster, Spencer Poff, Moya Chen, Douwe Kiela et al.EMNLP Findings 2021 · 579 citations
- When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric MemoriesAlex Mallen, Akari Asai, Victor Zhong, Rajarshi Das et al.ACL 2023 · 233 citations
Related papers
- Optimizing Retrieval-augmented Reader Models via Token EliminationMoshe Berchansky, Peter Izsak, Avi Caciularu, Ido Dagan et al.EMNLP 2023 · 4 citations
- REANO: Optimising Retrieval-Augmented Reader Models through Knowledge Graph GenerationJinyuan Fang, Zaiqiao Meng, Craig MacDonaldACL 2024 · 7 citations
- Pre-computed memory or on-the-fly encoding? A hybrid approach to retrieval augmentation makes the most of your computeMichiel de Jong, Yury Zemlyanskiy, Nicholas FitzGerald, Joshua Ainslie et al.ICML 2023 · 20 citations
- Making Retrieval-Augmented Language Models Robust to Irrelevant ContextOri Yoran, Tomer Wolfson, Ori Ram, Jonathan BerantICLR 2024 · 361 citations
- KG-FiD: Infusing Knowledge Graph in Fusion-in-Decoder for Open-Domain Question AnsweringDonghan Yu, Chenguang Zhu, Yuwei Fang, Wenhao Yu et al.ACL 2022 · 108 citations
