Abstract | Using parse accuracy in a simple reranking strategy for self-monitoring, we find that with a state-of-the-art averaged perceptron realization ranking model, BLEU scores cannot be improved with any of the well-known Treebank parsers we tested, since these parsers too often make errors that human readers would be unlikely to make. |
Abstract | Moreover, via a targeted manual analysis, we demonstrate that the SVM reranker frequently manages to avoid vicious ambiguities, while its ranking errors tend to affect fluency much more often than adequacy. |
Introduction | To do so—in a nutshell—we enumerate an n-best list of realizations and rerank them if necessary to avoid vicious ambiguities, as determined by one or more automatic parsers. |
Introduction | Consequently, we examine two reranking strategies, one a simple baseline approach and the other using an SVM reranker (J oachims, 2002). |
Introduction | Our simple reranking strategy for self-monitoring is to rerank the realizer’s n-best list by parse accuracy, preserving the original order in case of ties. |
Abstract | We propose a robust answer reranking model for non-factoid questions that integrates lexical semantics with discourse information, driven by two representations of discourse: a shallow representation centered around discourse markers, and a deep one based on Rhetorical Structure Theory. |
Approach | The proposed answer reranking component is embedded in the QA framework illustrated in Figure 1. |
Approach | CQA: In this scenario, the task is defined as reranking all the user-posted answers for a particular question to boost the community-selected best answer to the top position. |
Approach | These answer candidates are then passed to the answer reranking component, the focus of this work. |
Introduction | We propose a novel answer reranking (AR) model that combines lexical semantics (LS) with discourse information, driven by two representations of discourse: a shallow representation centered around discourse markers and surface text information, and a deep one based on the Rhetorical Structure Theory (RST) discourse framework (Mann and Thompson, 1988). |
Related Work | First, most NF QA approaches tend to use multiple similarity models (information retrieval or alignment) as features in discriminative rerankers (Riezler et al., 2007; Higashinaka and Isozaki, 2008; Verberne et al., 2010; Surdeanu et al., 2011). |
Related Work | (2011) extracted 47 cue phrases such as because from a small collection of web documents, and used the cosine similarity between an answer candidate and a bag of words containing these cue phrases as a single feature in their reranking model for non-factoid why QA. |
Related Work | This classifier was then used to extract instances of causal relations in answer candidates, which were turned into features in a reranking model for J apanense why QA. |
Abstract | LSH accounts for neighbor candidate pruning, while ITQ provides an efficient and effective reranking over the neighbor pool captured by LSH. |
Document Retrieval with Hashing | In this section, we first provide an overview of applying hashing techniques to a document retrieval task, and then introduce two unsupervised hashing algorithms: LSH acts as a neighbor-candidate filter, while ITQ works towards precise reranking over the candidate pool returned by LSH. |
Document Retrieval with Hashing | Hamming Reranking |
Document Retrieval with Hashing | In this framework, LSH accounts for neighbor candidate pruning, while ITQ provides an efficient and effective reranking over the neighbor pool captured by LSH. |
Experiments | Another crucial observation is that with ITQ reranking , a small number of LSH hash tables is needed in the pruning step. |
Experiments | Since the LSH pruning time can be ignored, the search time of the two-stage hashing scheme equals to the time of hamming distance reranking in ITQ codes for all candidates produced from LSH pruning step, e.g., LSH(48bits, 4 tables) + |
Experiments | 2 (f) shows the ITQ data reranking percentage for different LSH bit lengths and table numbers. |
Morphology-based Vocabulary Expansion | Reranking Models Given that the size of the expanded vocabulary can be quite large and it may include a lot of over-generation, we rerank the expanded set of words before taking the top n words to use in downstream processes. |
Morphology-based Vocabulary Expansion | We consider four reranking conditions which we describe below. |
Morphology-based Vocabulary Expansion | Reranked Expansion |
Experimental Setup | We also compare our model against a discriminative reranker . |
Experimental Setup | The reranker operates over the |
Experimental Setup | We then train the reranker by running 10 epochs of cost-augmented MIRA. |
Features | Global Features We used feature shown promising in prior reranking work Chamiak and Johnson (2005), Collins (2000) and Huang (2008). |
Introduction | They first appeared in the context of reranking (Collins, 2000), where a simple parser is used to generate a candidate list which is then reranked according to the scoring function. |
Introduction | Our method provides a more effective mechanism for handling global features than reranking , outperforming it by 1.3%. |
Related Work | The first successful approach in this arena was reranking (Collins, 2000; Charniak and J ohn-son, 2005) on constituency parsing. |
Related Work | Reranking can be combined with an arbitrary scoring function, and thus can easily incorporate global features over the entire parse tree. |
Related Work | Its main disadvantage is that the output parse can only be one of the few parses passed to the reranker . |
Results | 4The MST parser is trained in projective mode for reranking because generating top-k list from second-order non-projective model is intractable. |
Experiment | We group parsing systems into three categories: single systems, reranking systems and semi-supervised systems. |
Experiment | Our N0nlocal&Cluster system further improved the parsing F1 to 86.3%, and it outperforms all reranking systems and semi-supervised systems. |
Experiment | *Huang (2009) adapted the parse reranker to CTB5. |
Joint POS Tagging and Parsing with Nonlocal Features | But almost all previous work considered nonlocal features only in parse reranking frameworks. |
Related Work | However, almost all of the previous work use nonlocal features at the parse reranking stage. |
Experiments | We used the Charniak parser (Charniak et al., 2005) for our experiment, and we used the proposed algorithm to train the reranking feature weights. |
Experiments | For comparison, we also investigated training the reranker with Perceptron and MIRA. |
Experiments | There are around V = 1.33 million features in all defined for reranking, and the n-best size for reranking is set to 50. |
Introduction | There have been nonlocal approaches as well, such as tree-substitution parsers (Bod, 1993; Sima’an, 2000), neural net parsers (Henderson, 2003), and rerankers (Collins and Koo, 2005; Charniak and Johnson, 2005; Huang, 2008). |
Other Languages | it does not use a reranking step or post-hoc combination of parser results. |
Other Languages | 5 Their best parser, and the best overall parser from the shared task, is a reranked product of “Replaced” Berkeley parsers. |