Lecture Notes in Artificial Intelligence
Refine
Document Type
- Part of a Book (5)
- Conference Proceeding (2)
Language
- English (7)
Has Fulltext
- yes (7)
Keywords
- Categories of PSMs (2)
- Computerlinguistik (2)
- Deutsch (2)
- Korpus <Linguistik> (2)
- Problem Solving Methods (2)
- Reuse (2)
- Similarities (2)
- Architectures (1)
- Automatische Sprachanalyse (1)
- Datenerfassung (1)
Publicationstate
- Veröffentlichungsversion (4)
- Postprint (1)
Reviewstate
- Peer-Review (3)
- (Verlags)-Lektorat (2)
Publisher
- Springer (7)
8105
For a number of languages, web crawling allows researchers to collect huge text samples to build corpora. However, only part of the material found on the internet is useful for Natural Language Processing, as e.g. parsers typically cannot handle lists and tables, or very short or very long sentences. There are methods (cf. e.g. [3]) for cleaning the downloaded data before adding it to a corpus collection – but even when these are applied, not all remaining textual material might be suitable for certain research requirements. This paper describes methods utilized to prepare deWaC, a freely available German web corpus of the WaCky project, for automatic processing up to the parsing level. It then discusses ways in which this corpus, called SdeWaC, has been used since its release.
8686
We examine the task of relation extraction in the food domain by employing distant supervision. We focus on the extraction of two relations that are not only relevant to product recommendation in the food domain, but that also have significance in other domains, such as the fashion or electronics domain. In order to select suitable training data, we investigate various degrees of freedom. We consider three processing levels being argument level, sentence level and feature level. As external resources, we employ manually created surface patterns and semantic types on all these levels. We also explore in how far rule-based methods employing the same information are competitive.
1867
This paper deals with different views of lexical semantics. The focus is on the relationship between lexical expressions and conceptual components. First the assumptions about lexicalization and decompositionality of concepts shared by the most semanticists are presented, followed by a discussion of the differences between two-level-semantics and one-level-semantics. The final part is concentrated on the interpretation of conceptual components in situations of communication.
10713
We present an approach for modeling German negation in open-domain fine grained sentiment analysis. Unlike most previous work in sentiment analysis, we assume that negation can be conveyed by many lexical units (and not only common negation words) and that different negation words have different scopes. Our approach is examined on a new dataset comprising sentences with mentions of polar expressions and various negation words. We identify different types of negation words that have the same scopes. We show that already negation modeling based on these types largely outperforms traditional negation models which assume the same scope for all negation words and which employ a window-based scope detection rather than a scope detection based on syntactic information.
10713
We present a method for detecting and reconstructing separated particle verbs in a corpus of spoken German by following an approach suggested for written language. Our study shows that the method can be applied successfully to spoken language, compares different ways of dealing with structures that are specific to spoken language corpora, analyses some remaining problems, and discusses ways of optimising precision or recall for the method. The outlook sketches some possibilities for further work in related areas.
1319
In the last years a common notion of a Problem-Solving Method (PSM) emerged from different knowledge engineering frameworks. As a generic description of the dynamic behaviour of knowledge based systems PSMs are favored subjects of reuse. Up to now, most investigations on the reuse of PSMs focus on static features and methods as objects of reuse. By this, they ignore a lot of information of how the PSM was developed that is, in principle, entailed in the different parts of a conceptual model of a PSM.
In this paper the information of the different parts of PSMs is reconsidered from a reuse process point of view. A framework for generalized problem-solving methods is presented that describes the structure of a category of methods based on family resemblances. These generalized methods can be used to structure libraries of PSMs and - in the process of reuse - as a means to derive an incarnation, i.e. a member of its family of PSMs.
For illustrating the ideas, the approach is applied to the task rsp. problem type of parametric design.
1621
A library of software components should be essentially more than just a juxtaposition of its items. For problem-solving methods the notion of a family is suggested as means to cluster the items and to provide partially a structure of the library. This paper especially investigates how the similar control flows of the members of such a family can be described in one framework.