58 citations · 174 across the 28 of their papers we have counts for
8 papers · 1 filter
Rethinking Label Smoothing on Multi-hop Question Answering
Zhangyue Yin, Yuxin Wang, Xiannian Hu +6
Multi-Hop Question Answering (MHQA) is a significant area in question answering, requiring multiple reasoning components, including document retrieval, supporting sentence predicti…
Pre-training for Information Retrieval: Are Hyperlinks Fully Explored?
Jiawen Wu, Xinyu Zhang, Yutao Zhu +7
Recent years have witnessed great progress on applying pre-trained language models, e.g., BERT, to information retrieval (IR) tasks. Hyperlinks, which are commonly used in Web page…
Multi-Modal Experience Inspired AI Creation
Qian Cao, Xu Chen, Ruihua Song +3
AI creation, such as poem or lyrics generation, has attracted increasing attention from both industry and academic communities, with many promising models proposed in the past few…
Enhancing User Behavior Sequence Modeling by Generative Tasks for Session Search
Haonan Chen, Zhicheng Dou, Yutao Zhu +3
Users' search tasks have become increasingly complicated, requiring multiple queries and interactions with the results. Recent studies have demonstrated that modeling the historica…
Coarse-to-Fine: Hierarchical Multi-task Learning for Natural Language Understanding
Zhaoye Fei, Yu Tian, Yongkang Wu +9
Generalized text representations are the foundation of many natural language understanding tasks. To fully utilize the different corpus, it is inevitable that models need to unders…
RetroMAE: Pre-Training Retrieval-oriented Language Models Via Masked Auto-Encoder
Shitao Xiao, Zheng Liu, Yingxia Shao +1
Despite pre-training's progress in many important NLP tasks, it remains to explore effective pre-training strategies for dense retrieval. In this paper, we propose RetroMAE, a new…
Hyperlink-induced Pre-training for Passage Retrieval in Open-domain Question Answering
Jiawei Zhou, Xiaoguang Li, Lifeng Shang +10
To alleviate the data scarcity problem in training question answering systems, recent works propose additional intermediate pre-training for dense passage retrieval (DPR). However,…
KMIR: A Benchmark for Evaluating Knowledge Memorization, Identification and Reasoning Abilities of Language Models
Daniel Gao, Yantao Jia, Lei Li +6
Previous works show the great potential of pre-trained language models (PLMs) for storing a large amount of factual knowledge. However, to figure out whether PLMs can be reliable k…