Titlebook: Empirical Methods in Natural Language Generation; Data-oriented Method Emiel Krahmer,Mariët Theune Book 2010 Springer-Verlag Berlin Heidelb

Detonate

Erich Dambacher,Oliver Schöffskiviour to noisy feedback from the current generation context. This policy is compared to several baselines derived from previous work in this area. The learned policy significantly outperforms all the prior approaches.

我的巨大

Learning Adaptive Referring Expression Generation Policies for Spoken Dialogue Systemsapt their choice of referring expressions online to different users, and that these policies are significantly better than hand-coded adaptive policies for this problem. The learned policies are consistently between 2 and 8 turns shorter than a range of different hand-coded but adaptive baseline REG policies.

inspiration

Natural Language Generation as Planning under Uncertainty for Spoken Dialogue Systemsviour to noisy feedback from the current generation context. This policy is compared to several baselines derived from previous work in this area. The learned policy significantly outperforms all the prior approaches.

我没有命令

A Flexible Approach to Class-Based Ordering of Prenominal Modifiersings for sets of modifiers with more flexible positional constraints, and lends itself to bootstrapping for the classification of previously unseen modifiers. The approach to modifier classification outlined here is useful for automated language generation tasks, and the proposed modifier classes may be useful within constraint-based grammars.

indigenous

Introducing Shared Tasks to NLG: The TUNA Shared Task Evaluation Challengesnd in 2009. While we discuss the role of the .s in yielding a substantial body of research on the . problem, which has opened new avenues for future research, our main focus is on the role of different evaluation methods in assessing the output quality of . algorithms, and on the relationship between such methods.

conflate

Generating Referring Expressions in Context: The GREC Task Evaluation Challengesscourse context, (ii) embedded within an application context, and (iii) informed by naturally occurring data. This paper provides an overview of our aims and motivations in this research programme, the data resources we have built, and the first three shared-task challenges, .’08, .’09 and .’09, we have run based on the data.

地牢

The First Challenge on Generating Instructions in Virtual Environmentshe Internet. We describe the design and results of GIVE-1 as well as the participating NLG systems, and validate the experimental methodology by comparing the results collected over the Internet with results from a more traditional laboratory-based experiment.

carbohydrate

subacute

hing problem (also known as the assignment problem), a well studied problem in graph theory. Using BLEU to measure performance on a string regeneration task, we demonstrate an improvement over standard language model baselines, illustrating the benefit of the spanning tree approach incorporating an argument satisfaction model.