Citation Context

... computed also for the other four text representations previously introduced (B, Bg, D, and Dg). Counting the number of matched dependencies is essentially a simplified tree kernel for QA (e.g., see (=-=Moschitti et al., 2007-=-)) matching only trees of depth 2. Experiments with full dependency tree kernels based on several variants of the convolution kernels of Collins and Duffy (2001) did not yield improvements. We conject...

Citation Context

...omising results, e.g. [7, 20, 11, 32, 10, 21, 37, 17, 44]. More specifically, in question classification, tree kernels [43, 25] have shown accuracy comparable to the best models, e.g. [23]. Moreover, =-=[31, 28, 24]-=- have shown that shallow semantic information in the form of predicate argument structures (PASs) [14, 16] improves the automatic detection of correct answers to a target question. In particular, in [...

Citation Context

...rs beyond the bag-of-words paradigm by considering basic logical features in the text (i.e., AND, OR, NOT) as well as the existence of shallow grammatical features such as predicateargument structure(=-=Moschitti et al., 2007-=-) as well as semantic classes for words. Furthermore, it may be advantageous to expand upon the existing measures by applying machine learning techniques to create a hybrid decision system that would ...

Citation Context

...trees have to be identical to allow further parent matching, for which semantic representations cannot fit in well. To remedy the second issue, the Shallow Semantic Tree Kernel (SSTK) was proposed in =-=[14]-=-, where Predicate Argument Structures (PAS) are exploited to take dependencies into account. However, it was noted to be computational expensive for real world applications. In the remainder of this S...

Citation Context

...., 2005a; Basili et al., 2005b; Bloehdorn et al., 2006; Bloehdorn and Moschitti, 2007) or shallow semantic trees, (Giuglea and Moschitti, 2004; Giuglea and Moschitti, 2006; Moschitti and Bejan, 2004; =-=Moschitti et al., 2007-=-; Moschitti, 2008; Moschitti et al., 2008). ii) Design a new tree-based structures, which combines the information of both constituent and dependency parses. From dependency trees we can extract more ...

Citation Context

...ne grained task. It is worth nothing that when no similarity is applied: (i) BOW produces high accuracy, i.e. 88.8% but it is improved by STK (the current state-of-theart5 in QC (Zhang and Lee, 2003; =-=Moschitti et al., 2007-=-)); (ii) PTK applied to the same tree of STK produces a slightly lower value (non-statistically significant difference); (iii) interestingly, when PTK is instead applied to dependency structures, it i...

Citation Context

...sence of the words popcorn and vegan. machine learning algorithms applied to similarity scores between questions and answers would not be able to learn patterns such as the one above. In contrast, in =-=[23]-=-, we proposed to apply structural kernels [28] to the syntactic representation of question/answer passage pairs. However, such model could only take into account very short passages composed of only o...

Citation Context

.... (Shen and Lapata, 2007) has shown that shallow semantic information in the form of predicate argument structures (PASs) improves the automatic detection of correct answers to a target question. In (=-=Moschitti et al., 2007-=-), we proposed the Shallow Semantic Tree Kernel (SSTK) designed to encode PASs 1 in SVMs. 1 in PropBank format, (www.cis.upenn.edu/˜ace). In this paper, similarly to our previous approach, we design a...

Citation Context

... for natural language processing for any task in which structural information plays a role, e.g. (Collins and Duffy 2002; Zelenko et al. 2003; Giuglea and Moschitti 2006; Zanzotto and Moschitti 2006; =-=Moschitti et al. 2007-=-). Indeed, they have already been used in NLP to encode the type of structural information that plays a role in binding constraints (Yang et al. 2006); however, the methods used in this previous work ...