Filters








31,713 Hits in 15.1 sec

A recursive TF-ISF Based Sentence Retrieval Method with Local Context

Alen Doko, Maja Štula, Darko Stipaničev
2013 International Journal of Machine Learning and Computing  
However some newer sentence retrieval methods based on the language modeling framework successfully use some kind of context of sentences.  ...  Additional benefit of our method is the clear explicit model of the context that will allow us to automatically generate a document representation with context suitable for sentence retrieval which is  ...  This is identical to the improvements of including context into language modeling framework [1] .  ... 
doi:10.7763/ijmlc.2013.v3.301 fatcat:6c3dftsc2vcb7ioguxr3jd34my

Toward Making the Most of Context in Neural Machine Translation

Zaixiang Zheng, Xiang Yue, Shujian Huang, Jiajun Chen, Alexandra Birch
2020 Zenodo  
We argue that previous research did not make a clear use of the global context, and propose a new documentlevel NMT framework that deliberately models the local context of each sentence with the awareness  ...  We specifically design the model to be able to deal with documents containing any number of sentences, including single sentences.  ...  Segment-level Relative Attention Given the local representations of each sentences, we propose to extend the relative attention [Shaw et al., 2018] from token-level to segmentlevel to model the inter-sentence  ... 
doi:10.5281/zenodo.3923504 fatcat:s6jcn6vygrg2zonct5emgiq4iy

Towards Making the Most of Context in Neural Machine Translation [article]

Zaixiang Zheng, Xiang Yue, Shujian Huang, Jiajun Chen, Alexandra Birch
2020 arXiv   pre-print
We argue that previous research did not make a clear use of the global context, and propose a new document-level NMT framework that deliberately models the local context of each sentence with the awareness  ...  We specifically design the model to be able to deal with documents containing any number of sentences, including single sentences.  ...  Acknowledgements Shujian Huang is the corresponding author. This work was supported by the National Science Foundation of China (No. U1836221, 61772261, 61672277).  ... 
arXiv:2002.07982v2 fatcat:vzc5llfxzvazjp4j7owgt46dga

IRIT at INEX: Question Answering Task [chapter]

Liana Ermakova, Josiane Mothe
2012 Lecture Notes in Computer Science  
Sentence retrieval is based on standard TF-IDF measure enriched by named entity recognition, POS weighting and smoothing from local context.  ...  In this paper we describe an approach to tweet contextualization developed in the context of INEX QA track. The task is to provide a context up to 500 words to a tweet.  ...  The context was extended to two sentences in each direction and the target sentence weight was equal to 0.7. For NE we kept the weight equal to 1.0.  ... 
doi:10.1007/978-3-642-35734-3_19 fatcat:nbqs5qsoavgtph54as4gyqo4mu

End-to-End Convolutional Semantic Embeddings

Quanzeng You, Jiebo Luo, Zhengyou Zhang
2018 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition  
This could potentially limit the performance of the model due to the hierarchical structures in images and sentences.  ...  The experimental results show that our proposed textual CNN models with the new learning objective lead to better performance than the state-of-the-art approaches.  ...  Acknowledgment We thank the support of New York State through the Goergen Institute for Data Science, the corporate research sponsors Adobe and Snap Inc., and NSF Award #1704309.  ... 
doi:10.1109/cvpr.2018.00601 dblp:conf/cvpr/YouZL18 fatcat:aegpjpoze5ctvo6klzbxxoxwqe

Teacher-Student Networks with Multiple Decoders for Solving Math Word Problem

Jipeng Zhang, Roy Ka-Wei Lee, Ee-Peng Lim, Wei Qin, Lei Wang, Jie Shao, Qianru Sun
2020 Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence  
MWP models often simply fit this solution rather than truly understand or solve the problem. The generalization of models (to diverse word scenarios) is thus limited.  ...  To address this problem, this paper proposes a novel approach, TSN-MD, by leveraging the teacher network to integrate the knowledge of equivalent solution expressions and then to regularize the learning  ...  Acknowledgements Shujian Huang is the corresponding author. This work was supported by the National Science Foundation of China (No. U1836221, 61772261, 61672277).  ... 
doi:10.24963/ijcai.2020/551 dblp:conf/ijcai/ZhengYHCB20 fatcat:ee2hkv64vraivk2f2d4l224ssm

A Pooling Approach to Modelling Spatial Relations for Image Retrieval and Annotation [article]

Mateusz Malinowski, Mario Fritz
2015 arXiv   pre-print
Due to the complexity of the spatial language, we argue for a learning-based approach that acquires a representation of spatial relations by learning parameters of the pooling operator.  ...  Over the last two decades we have witnessed strong progress on modeling visual object classes, scenes and attributes that have significantly contributed to automated image understanding.  ...  Acknowledgement: We would like to thank Andrej Karpathy for his support and making his code available.  ... 
arXiv:1411.5190v2 fatcat:qy2ij4ljf5czfab2cpkdrdsbne

OMG: Observe Multiple Granularities for Natural Language-Based Vehicle Retrieval [article]

Yunhao Du, Binyu Zhang, Xiangning Ruan, Fei Su, Zhicheng Zhao, Hong Chen
2022 arXiv   pre-print
To tackle this issue, we propose a novel framework for the natural language-based vehicle retrieval task, OMG, which Observes Multiple Granularities with respect to visual representation, textual representation  ...  For the visual representation, target features, context features and motion features are encoded separately.  ...  Conclusion In this paper, we propose a novel framework for the natural language-based vehicle retrieval task, named OMG.  ... 
arXiv:2204.08209v2 fatcat:dymrkdwxwzew5fbgo3x4gyw6vm

On Broca, brain, and binding: a new framework

Peter Hagoort
2005 Trends in Cognitive Sciences  
Introduction In this article I will outline a new general framework for a model of the neural architecture of language, with a focus on the role of Broca's area.  ...  Broca's area extended Broca's area is a crucial component in all classical neurobiological models of language.  ... 
doi:10.1016/j.tics.2005.07.004 pmid:16054419 fatcat:nrrrgltwircipdidktqkygmliu

Learning 2D Temporal Adjacent Networks for Moment Localization with Natural Language [article]

Songyang Zhang, Houwen Peng, Jianlong Fu, Jiebo Luo
2020 arXiv   pre-print
Based on the 2D map, we propose a Temporal Adjacent Network (2D-TAN), a single-shot framework for moment localization.  ...  We address the problem of retrieving a specific moment from an untrimmed video by a query sentence.  ...  Acknowledgement We thank the support of NSF awards IIS-1704337, IIS-1722847, IIS-1813709, and the generous gift from our corporate sponsors.  ... 
arXiv:1912.03590v3 fatcat:owznntu63bfzbatngsewcwxuuu

Learning 2D Temporal Adjacent Networks for Moment Localization with Natural Language

Songyang Zhang, Houwen Peng, Jianlong Fu, Jiebo Luo
2020 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
Based on the 2D map, we propose a Temporal Adjacent Network (2D-TAN), a single-shot framework for moment localization.  ...  We address the problem of retrieving a specific moment from an untrimmed video by a query sentence.  ...  Acknowledgement We thank the support of NSF awards IIS-1704337, IIS-1722847, IIS-1813709, and the generous gift from our corporate sponsors.  ... 
doi:10.1609/aaai.v34i07.6984 fatcat:6imibis2ifb55gshcmsj5saq44

A Survey on Retrieval-Augmented Text Generation [article]

Huayang Li and Yixuan Su and Deng Cai and Yan Wang and Lemao Liu
2022 arXiv   pre-print
It firstly highlights the generic paradigm of retrieval-augmented generation, and then it reviews notable approaches according to different tasks including dialogue response generation, machine translation  ...  Compared with conventional generation models, retrieval-augmented text generation has remarkable advantages and particularly has achieved state-of-the-art performance in many NLP tasks.  ...  In this sense, it is urgent to consider some methods to trade off the retrieval memory size and retrieval efficiency, for example, data compression for the retrieval memory. Local vs.  ... 
arXiv:2202.01110v2 fatcat:vo6i5vq62raxtcry2xcbrm55be

WINGS:Writing with Intelligent Guidance and Suggestions

Xianjun Dai, Yuanchao Liu, Xiaolong Wang, Bingquan Liu
2014 Proceedings of 52nd Annual Meeting of the Association for Computational Linguistics: System Demonstrations  
The main strategies used by WINGS, including providing syntactically and semantically related words based on word vector representation and recommending contextually related sentences based on LDA, are  ...  In addition to supporting common Chinese input, WINGS mainly attempts to spark users' inspirations by recommending both word level and sentence level writing suggestions.  ...  Sentences Recommending By introducing the topic model LDA, the sentences recommended by WINGS are related to the topic of the local context.  ... 
doi:10.3115/v1/p14-5005 dblp:conf/acl/DaiLWL14 fatcat:6iyaeoosqbhwrgj2y6uayl3psi

Getting Started with Neural Models for Semantic Matching in Web Search [article]

Kezban Dilek Onal, Ismail Sengor Altingovde, Pinar Karagoz, Maarten de Rijke
2016 arXiv   pre-print
Recent advances in language technology have given rise to unsupervised neural models for learning representations of words as well as bigger textual units.  ...  This survey is meant as an introduction to the use of neural models for semantic matching. To remain focused we limit ourselves to web search.  ...  These similarity scores are included as additional features to represent session context in a context-aware learning to rank framework.  ... 
arXiv:1611.03305v1 fatcat:agdgj7allbczxcyteuomswn574

MixingBoard: a Knowledgeable Stylized Integrated Text Generation Platform [article]

Xiang Gao, Michel Galley, Bill Dolan
2020 arXiv   pre-print
To borrow advantages from different models, we implement strategies for cross-model integration, from the token probability level to the latent space level.  ...  A user interface for local development, remote webpage access, and a RESTful API are provided to make it simple for users to build their own demos.  ...  As an extended version of the one introduced by Gao et al. (2019c) , the current demo is grounded on knowledge and coupled with more advanced language modeling .  ... 
arXiv:2005.08365v2 fatcat:xc67c4ns3jblraxvy7ratgsy4u
« Previous Showing results 1 — 15 out of 31,713 results