4,775,584 Hits in 6.2 sec


2009 Advances in Complex Systems  
We propose an alternative notion of predictive models in terms of a hidden Markov model capable of generating the underlying stochastic process.  ...  Here, the corresponding so-called ε-machine encodes the mechanisms of prediction.  ...  They also thank the anonymous referees for their constructive comments which substantially improved the paper. Nihat Ay is affiliated with the Santa Fe Institute.  ... 
doi:10.1142/s0219525909002143 fatcat:e3ekqhzusrhmnijqnucx3ylc3e

Reading Times Predict the Quality of Generated Text Above and Beyond Human Ratings

Sina Zarrieß, Sebastian Loth, David Schlangen
2015 Proceedings of the 15th European Workshop on Natural Language Generation (ENLG)  
Typically, human evaluation of NLG output is based on user ratings. We collected ratings and reading time data in a simple, low-cost experimental paradigm for text generation.  ...  generated text.  ...  Many aspects of linguistic well-formedness and naturalness play a role for assessing the quality of an automatically generated text.  ... 
doi:10.18653/v1/w15-4705 dblp:conf/enlg/ZarriessLS15 fatcat:x4fnem4g3jbw5mxiffowau4iyy

Generating E-Commerce Product Titles and Predicting their Quality

José G. Camargo de Souza, Michael Kozielski, Prashant Mathur, Ernie Chang, Marco Guerini, Matteo Negri, Marco Turchi, Evgeny Matusov
2018 Proceedings of the 11th International Conference on Natural Language Generation  
The task of automatically generating these titles given noisy user provided titles is one way to achieve the goal.  ...  As such, we propose approaches that (i) automatically generate product titles, (ii) predict their quality.  ...  The observation that BLEU alone is not appropriate for evaluating natural language generation systems is not new and corroborates previous work on the field, most notably the recent work by Reiter (2018  ... 
doi:10.18653/v1/w18-6530 dblp:conf/inlg/SouzaKMCGNTM18 fatcat:54ewmhr3xjgiracdemml6wspmq

A Study on the Generalized Approximation Modeling Method Based on Fitting Sensitivity for Prediction of Engine Performance

Lin Lin, Fang Wang, Shisheng Zhong
2017 Discrete Dynamics in Nature and Society  
By investigating the effect of every model parameter on the prediction precision in experiments, we summarized the change regularities of the root-mean-square errors (RMSEs) varying with the model parameters  ...  In the prediction of engine performance, to address overfitting and underfitting problems with the approximation modeling technique, we derived a generalized approximation model that could be used to adjust  ...  Acknowledgments The authors are thankful for the support from the Key National Natural Science Foundation of China, China (no.  ... 
doi:10.1155/2017/5729786 fatcat:oo4tyoayhzcidax33oldynrozi

Coarse-grained Candidate Generation and Fine-grained Re-ranking for Chinese Abbreviation Prediction

Longkai Zhang, Houfeng WANG, Xu Sun
2014 Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP)  
Correctly predicting abbreviations given the full forms is important in many natural language processing systems.  ...  The candidate generation and coarse-grained ranking is totally unsupervised. The re-ranking phase can use a very small amount of training data to get a reasonably good result.  ...  Chinese Abbreviation Prediction System Chinese Abbreviation Prediction is the task of selecting representative characters from the long full form 1 .  ... 
doi:10.3115/v1/d14-1202 dblp:conf/emnlp/ZhangWS14 fatcat:c7vo23r6mnf7hkldejvxpetxmq

Inducing Transformer's Compositional Generalization Ability via Auxiliary Sequence Prediction Tasks

Yichen Jiang, Mohit Bansal
2021 Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing   unpublished
These automatically-generated sequences are more representative of the underlying compositional symbolic structures of the input data.  ...  Motivated by the failure of a Transformer model on the SCAN compositionality challenge (Lake and , which requires parsing a command into actions, we propose two auxiliary sequence prediction tasks as additional  ...  The views are those of the authors and not of the funding agency.  ... 
doi:10.18653/v1/2021.emnlp-main.505 fatcat:4xm6czoc6zbpvdtclpesv73zw4

Aspect Sentiment Quad Prediction as Paraphrase Generation

Wenxuan Zhang, Yang Deng, Xin Li, Yifei Yuan, Lidong Bing, Wai Lam
2021 Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing   unpublished
, pages 9209–9219 November 7–11, 2021. c 2021 Association for Computational Linguistics prediction performance hinges on the accuracy of of natural sentences, rather  ...  On one our PARAPHRASE method naturally facilitates the hand, the sentiment quads can be predicted in an knowledge transfer across related tasks with the end-to-end manner, alleviating  ... 
doi:10.18653/v1/2021.emnlp-main.726 fatcat:x22fbjfr2nchnpbfklvh32rzki

How Do Neural Sequence Models Generalize? Local and Global Cues for Out-of-Distribution Prediction

D. Anthony Bau, Jacob Andreas
2021 Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing   unpublished
We validate our models of gener- 3.1 Local and global models of generalization alization using real models trained on natural data We focus on two idealized models of the general- and explain them  ...  {(XG , XL )i }, a hypothesis p̃, and a trained To compute generalization model predictions model pLM , we compute the accuracy of the on natural  ... 
doi:10.18653/v1/2021.emnlp-main.448 fatcat:5ecowb2cvrfvfij2ijvw72wev4

Answering Ambiguous Questions through Generative Evidence Fusion and Round-Trip Prediction

Yifan Gao, Henghui Zhu, Patrick Ng, Cicero Nogueira dos Santos, Zhiguo Wang, Feng Nan, Dejiao Zhang, Ramesh Nallapati, Andrew O. Arnold, Bing Xiang
2021 Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)   unpublished
Therefore, a system needs to find possible interpretations of the question, and predict one or multiple plausible answers.  ...  Our model, named REFUEL, achieves a new state-of-the-art performance on the AMBIGQA dataset, and shows competitive performance on NQ-OPEN and Trivi-aQA.  ...  One straightforward way to generate more QA pairs is setting a minimum length of generation for the answer prediction model, and then go through the LM Verification process to drop the low-quality predictions  ... 
doi:10.18653/v1/2021.acl-long.253 fatcat:arwdfevcwzbznd2skk5hhqex3q

An Enhanced Electrolarynx with Automatic Fundamental Frequency Control based on Statistical Prediction

Kou Tanaka, Tomoki Toda, Graham Neubig, Sakriani Sakti, Satoshi Nakamura
2015 Proceedings of the 17th International ACM SIGACCESS Conference on Computers & Accessibility - ASSETS '15  
To make it possible to generate more natural excitation sounds, we have proposed a method to automatically control the fundamental frequency of the sounds generated by the electrolarynx based on a statistical  ...  prediction model, which predicts the fundamental frequency patterns from the produced EL speech in real-time.  ...  Consequently, quality of life of laryngectomees is significantly degraded. To generate more natural F0 patterns, we have proposed a method to control F0 based on the statistical F0 prediction [1] .  ... 
doi:10.1145/2700648.2811340 dblp:conf/assets/TanakaTNSN15 fatcat:uerrwnomnvhvpcengui4qlcska

DNN Architecture for High Performance Prediction on Natural Videos Loses Submodule's Ability to Learn Discrete-World Dataset [article]

Lana Sinapayen, Atsushi Noda
2019 arXiv   pre-print
We consider the theory that predictive coding is such a general rule, and falsify it for one specific neural architecture known for high-performance predictions on natural videos and replication of human  ...  itself is responsible for both the high performance on natural videos and the loss of performance on the GoL.  ...  This suggests that high performance on natural videos comes to the cost of generality for videos exhibiting different types of dynamics (even if these dynamics are totally deterministic and predictable  ... 
arXiv:1904.07969v1 fatcat:toxbq3irkzbmvpse6mvy2tbsme

Niche construction, sources of selection and trait coevolution

Kevin Laland, John Odling-Smee, John Endler
2017 Interface Focus  
We show how the perspective leads to testable predictions related to: (i) reduced variance in measures of responses to natural selection in the wild; (ii) multiple trait coevolution, including the evolution  ...  More generally, we submit that evolutionary biology would benefit from greater attention to the diverse properties of all sources of selection.  ...  We are grateful to Andrew Clark, Hilton Japyassú , Sally Street, Tobias Uller and two anonymous referees for helpful comments on earlier drafts.  ... 
doi:10.1098/rsfs.2016.0147 pmid:28839920 fatcat:zse67vbrxzev7eualllymgbzpq

Understanding and Predicting the Memorability of Outdoor Natural Scenes [article]

Jiaxin Lu, Mai Xu, Ren Yang, Zulin Wang
2020 arXiv   pre-print
Recent works have shed light on the visual features that make generic images, object images or face photographs memorable.  ...  However, these methods are not able to effectively predict the memorability of outdoor natural scene images.  ...  In summary, our Deep-NSM model outperforms the state-of-the-art generic methods on predicting outdoor natural scene memorability, making up the shortcomings of these generic image methods.  ... 
arXiv:1810.06679v4 fatcat:vmng3pgngzbpvgug7cqnh56aom

Towards Interpretable Natural Language Understanding with Explanations as Latent Variables [article]

Wangchunshu Zhou, Jinyi Hu, Hanlin Zhang, Xiaodan Liang, Maosong Sun, Chenyan Xiong, Jian Tang
2021 arXiv   pre-print
Experiments on two natural language understanding tasks demonstrate that our framework can not only make effective predictions in both supervised and semi-supervised settings, but also generate good natural  ...  Our framework treats natural language explanations as latent variables that model the underlying reasoning process of a neural model.  ...  Acknowledgments This project is supported by the Natural Sciences and Engineering Research Council (NSERC) Discovery Grant, the Canada CIFAR AI Chair Program, collaboration grants between Microsoft Research  ... 
arXiv:2011.05268v2 fatcat:szknr42dxbcmrdir3s5ntbyr3i

Flaws in evaluation schemes for pair-input computational predictions

Yungki Park, Edward M Marcotte
2012 Nature Methods  
To the Editor: Computational prediction methods that operate on pairs of objects by considering features of each (hereafter referred to as "pair-input methods") have been crucial in many areas of biology  ...  In this study we demonstrate that the paired nature of inputs has significant, though not yet widely perceived, implications for the validation of pair-input methods.  ...  on the manuscript.  ... 
doi:10.1038/nmeth.2259 pmid:23223166 pmcid:PMC3531800 fatcat:lrsjgqhlhrgibdwygg2pcjxlju
« Previous Showing results 1 — 15 out of 4,775,584 results