Filters








2,421 Hits in 4.5 sec

Dynamic Prefix-Tuning for Generative Template-based Event Extraction [article]

Xiao Liu, Heyan Huang, Ge Shi, Bo Wang
2022 arXiv   pre-print
We consider event extraction in a generative manner with template-based conditional generation.  ...  In this paper, we propose a generative template-based event extraction method with dynamic prefix (GTEE-DynPref) by integrating context information with type-specific prefixes to learn a context-specific  ...  Acknowledgments We thank the anonymous reviewers for their valuable comments and suggestions. This work is supported by National Natural Science Foundation of China (No. U19B2020 and No. 62106010).  ... 
arXiv:2205.06166v1 fatcat:la5vzkmjofhlld2wdwczlrx2q4

Dynamic Prefix-Tuning for Generative Template-based Event Extraction

Xiao Liu, Heyan Huang, Ge Shi, Bo Wang
2022 Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)   unpublished
We consider event extraction in a generative manner with template-based conditional generation.  ...  In this paper, we propose a generative templatebased event extraction method with dynamic prefix (GTEE-DYNPREF) by integrating context information with type-specific prefixes to learn a context-specific  ...  Acknowledgments We thank the anonymous reviewers for their valuable comments and suggestions. This work is supported by National Natural Science Foundation of China (No. U19B2020 and No. 62106010).  ... 
doi:10.18653/v1/2022.acl-long.358 fatcat:36v6j46phbeqndhrqyhgubhfdi

Efficient Online Log Parsing with Log Punctuations Signature

Shijie Zhang, Gang Wu
2021 Applied Sciences  
The core of LogPunk is a novel log signature method based on log punctuations and length features. According to the signature, we can quickly find a small set of candidate templates.  ...  As the first and foremost step of typical log analysis, many data-driven methods have been proposed for automated log parsing.  ...  A mechanism for automated parameters tuning can greatly improve this situation. (2) Punctuation table generation. The punctuation table (cf.  ... 
doi:10.3390/app112411974 fatcat:dq7lxnbpnvfw5pbgh5qju6vbqe

Recent Advances in Natural Language Processing via Large Pre-Trained Language Models: A Survey [article]

Bonan Min, Hayley Ross, Elior Sulem, Amir Pouran Ben Veyseh, Thien Huu Nguyen, Oscar Sainz, Eneko Agirre, Ilana Heinz, Dan Roth
2021 arXiv   pre-print
We also present approaches that use pre-trained language models to generate data for training augmentation or other purposes.  ...  We present a survey of recent work that uses these large language models to solve NLP tasks via pre-training then fine-tuning, prompting, or text generation approaches.  ...  Acknowledgments We would like to thank Paul Cummer for his insightful comments on this work.  ... 
arXiv:2111.01243v1 fatcat:4xfjkkby2bfnhdrhmrdlliy76m

HealthPrompt: A Zero-shot Learning Paradigm for Clinical Natural Language Processing [article]

Sonish Sivarajkumar, Yanshan Wang
2022 arXiv   pre-print
Prompt-based learning is an emerging ZSL technique where we define task-based templates for NLP tasks.  ...  In this technique, rather than fine-tuning a Pre-trained Language Model(PLM), the task definitions are tuned by defining a prompt template.  ...  Li et al. 32 experimented with prefix tuning and showed that these prompts could improve generalization on Natural Language Generation (NLG) tasks. "[X].  ... 
arXiv:2203.05061v1 fatcat:avus5stdqrcsji4fw5n7ghedim

Continuous Detection, Rapidly React: Unseen Rumors Detection based on Continual Prompt-Tuning [article]

Yuhui Zuo, Wei Zhu, Guoyong Cai
2022 arXiv   pre-print
To avoid forgetting, we optimize and store task-special soft-prompt for each domain.  ...  Since open social platforms allow for a large and continuous flow of unverified information, rumors can emerge unexpectedly and spread quickly.  ...  EANN: Wang et al. (2018) uses adversarial networks to extract event invariant features for generalization when new events arrive.  ... 
arXiv:2203.11720v1 fatcat:he33j35wg5dmtboxiphianzkya

Translating Place-Related Questions to GeoSPARQL Queries [article]

Ehsan Hamzei, Martin Tomko, Stephan Winter
2022 arXiv   pre-print
Next, we use a dynamic template-based approach to generate fully executable GeoSPARQL queries from the logical representations.  ...  Yet, to enable question answering based on linked knowledge bases, natural language questions must first be re-formulated as formal queries.  ...  The dynamic approach uses templates for partial GeoSPARQL constructs, e.g., defining a concept, rather than a template for the whole query.  ... 
arXiv:2205.03067v1 fatcat:xkmnxpf3qzf3heuyezmct4lswm

Contrastive Demonstration Tuning for Pre-trained Language Models [article]

Xiaozhuan Liang, Ningyu Zhang, Siyuan Cheng, Zhen Bi, Zhenru Zhang, Chuanqi Tan, Songfang Huang, Fei Huang, Huajun Chen
2022 arXiv   pre-print
Concretely, the demonstration examples are crucial for an excellent final performance of prompt-tuning.  ...  Recent works have focused on automatically searching discrete or continuous prompts or optimized verbalizers, yet studies for the demonstration are still limited.  ...  Additionally, propose LM-BFF-better few-shot fine-tuning of language models, which utilizes a generation model to obtain templates and a refined strategy for dynamically and selectively incorporating  ... 
arXiv:2204.04392v2 fatcat:oczxkefaz5fw3mo3moachgb4dq

Human Effort and Machine Learnability in Computer Aided Translation

Spence Green, Sida I. Wang, Jason Chuang, Jeffrey Heer, Sebastian Schuster, Christopher D. Manning
2014 Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP)  
We describe a new translator interface, extensive modifications to a phrasebased MT system, and a novel objective function for re-tuning to human corrections.  ...  However, re-tuning the MT system to interactive output leads to larger, statistically significant reductions in HTER versus re-tuning to post-edit.  ...  Acknowledgements We thank TAUS for access to their data repository. We also thank John DeNero, Chris Dyer, Alon Lavie, and Matt Post for helpful conversations.  ... 
doi:10.3115/v1/d14-1130 dblp:conf/emnlp/GreenWCHSM14 fatcat:iw2tuyv4cfaoznusc2hwwtib2a

DISco: a Distributed Information Store for network Challenges and their Outcome [article]

Sylvain Martin, Laurent Chiarello, Guy Leduc
2012 arXiv   pre-print
We propose to use state-of-art peer-to-peer publish/subscribe and distributed storage as core building blocks for the DISco service.  ...  Shared keys to establish the secure connection for the lookup request are generated iteratively and piggy-backed on BGP prefix advertisement 4 .  ...  The use of dynamic aggregation and filtering as events are forwarded is a fundamental feature of our solution, for which we propose an OMNet++ implementation based on the OverSim package.  ... 
arXiv:1201.3073v1 fatcat:nolg5lpowbgynjye3nkrgiza54

BioEve Search: A Novel Framework to Facilitate Interactive Literature Search

Syed Toufeeq Ahmed, Hasan Davulcu, Sukru Tikves, Radhika Nair, Zhongming Zhao
2012 Advances in Bioinformatics  
We developed a novel framework (named "BioEve") that seamlessly integrates Faceted Search (Information Retrieval) with Information Extraction module to provide an interactive search experience for the  ...  An automated extraction system coupled with a cognitive search and navigation service over these document collections would not only save time and effort, but also pave the way to discover hitherto unknown  ...  Kanwar for valuable discussions, ideas, and help with writing this paper. They also wish to thank the researchers who provided an initial user review and gave them valuable feedback.  ... 
doi:10.1155/2012/509126 pmid:22693501 pmcid:PMC3368157 fatcat:om4m3xco7vd7rkjmnpvhtkysoq

Neural factoid geospatial question answering

Haonan Li, Ehsan Hamzei, Ivan Majic, Hua Hua, Jochen Renz, Martin Tomko, Maria Vasardani, Stephan Winter, Timothy Baldwin
2021 Journal of Spatial Information Science  
Compared with current template-based approaches, our method generalises to a broader range of questions, including those with complex syntax and semantics.  ...  task, and the graph generation of the query as a semantic dependency parsing task.  ...  Generally, IR-based QA queries the web or other document collection for relevant documents, from which it extracts answers.  ... 
doi:10.5311/josis.2021.23.159 fatcat:mvba3par2zb7zoy7kjv3optryu

Pretrained Language Models for Text Generation: A Survey [article]

Junyi Li, Tianyi Tang, Wayne Xin Zhao, Jian-Yun Nie, Ji-Rong Wen
2022 arXiv   pre-print
Finally, we highlight the future research directions which will further improve these PLMs for text generation.  ...  We also include a summary of various useful resources and typical text generation applications based on PLMs.  ...  In contrast to full-parameter fine-tuning, which requires storing a tuned copy of the model for each text generation task, prefix-tuning only optimizes the prefix for each text generation task.  ... 
arXiv:2201.05273v4 fatcat:pnffabspsnbhvo44gbaorhxc3a

A Directed Acyclic Graph Approach to Online Log Parsing [article]

Pinjia He, Jieming Zhu, Pengcheng Xu, Zibin Zheng, Michael R. Lyu
2018 arXiv   pre-print
Drain can automatically generate a directed acyclic graph for a new system and update the graph according to the incoming log messages.  ...  To address this problem, we propose an online log parsing method, namely Drain, based on directed acyclic graph, which encodes specially designed rules for parsing.  ...  With respect to system runtime, each logging statement generates log messages of the same log event (i.e., log template), which describes a specific system operation.  ... 
arXiv:1806.04356v1 fatcat:shwr276o7bhx5h2jbkzlznuuy4

Multiple Features Based Approach to Extract Bio-molecular Event Triggers Using Conditional Random Field

Amit Majumder
2012 International Journal of Intelligent Systems and Applications  
Proposed Approach for Event Word Extraction This, as in HMMs, can be obtained eff iciently by In this section we describe our proposed approach for dynamic programming.  ...  template to U33:%x[0,50] generate model file.  ... 
doi:10.5815/ijisa.2012.12.06 fatcat:elq62iki2bfwpd2yc2b6gzztgy
« Previous Showing results 1 — 15 out of 2,421 results