Sagerer, G. and F. Kummert: "Knowledge Based Systems for Speech Understanding", in: Niemann, H., M. Lang, and G. Sagerer (eds.): "Recent Advances in Speech Understanding and Dialog Systems", NATO ASI Series F: Computer and Systems Sciences, vol 46, Springer-Verlag, 1988, pp. 421-458.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....Speech to speech real time translation system for the domain of conference registration. Language and dialogue: Uses acousticphonetic, linguistic and discourse knowledge. The architecture has massive parallel processing, a connectivistic network, multilevel parse trees, and marker passing. Ernest [Sagerer et al. 1988]: Train travel system. Representation: Uses semantic nets for the representation of linguistic knowledge. EVAR [Niemann et al. 1987] German intercity train inquiries. Recogniser: 3918 words, speaker (simulated) independent continuous speech. 36 of the 64 German phonemes are used. Language and ....

Sagerer, G. and F. Kummert: "Knowledge Based Systems for Speech Understanding", in: Niemann, H., M. Lang, and G. Sagerer (eds.): "Recent Advances in Speech Understanding and Dialog Systems", NATO ASI Series F: Computer and Systems Sciences, vol 46, Springer-Verlag, 1988, pp. 421-458.

....represent concept descriptions incorporating constraints arising from analysis, and adjacency matrices. The latter serve to describe well formed sequences of the concepts components. Furthermore, ERNEST enables a uniform representation of all knowledge that is needed for a linguistic analysis [8]. However, the description of linguistic knowledge is distinguished between various levels of abstraction. The hypothesis level forms the traditional interface between acoustic recognition and linguistic analysis. The syntactic level consists of concepts describing the structure of syntactic ....

G. Sagerer and F. Kummert. Knowledge based systems for speech understanding. In H. Niemann, M. Lang, and G. Sagerer, editors, Recent Advances in Speech Understanding and Dialog Systems, pages 421--458. NATO ASI Series F, Vol. 46, Springer-Verlag, Berlin, 1988.

....recognition process. This method of feedback is even more flexible and general than the use of various language models for e.g. different dialogue steps. 2 LINGUISTIC ANALYSIS The formalism of a semantic network is used to represent linguistic knowledge as specified by the ERNEST language [4]. Important features of ERNEST are the existence of modified concepts that represent concept descriptions incorporating constraints arising from analysis and adjacency matrices. The latter serve to describe the possible well formed sequences of a concept s components. Syntactic, semantic, and ....

.... data as well as on the expectations derived from the linguistic model [3] For a goal directed search of the best fitting interpretation the A algorithm with a judgment vector is used reflecting the compatibility, the quality, the reliability and the relevance of a partial interpretation [4]. 3 ACOUSTIC ANALYSIS The acoustic part of the analysis is covered by the ISADORA system [5] which provides highly flexible speech recognition based on Hidden Markov Models (HMM) ISADORA consists of modules for cepstral feature extraction, for optional hard or soft gaussian vector ....

G. Sagerer and F. Kummert. Knowledge based systems for speech understanding. In H. Niemann, M. Lang, and G. Sagerer, editors, Recent Advances in Speech Understanding and Dialog Systems, pages 421--458. NATO ASI Series F, Vol. 46, Springer-Verlag, Berlin, 1988.

....all knowledge sources within a semantic network makes an automatic derivation of language models possible. An algorithm for this task is presented in the following sections. 2 Structure of the Linguistic Knowledge Base The formalism of a semantic network is used to represent linguistic knowledge [5]. The representation language features three types of nodes (con cept, modified concept and instance) and three types of edges or links to connect nodes (part, specialization and concrete) A concept is used for the intensional description of an object, an event or some abstract conception. An ....

G. Sagerer and F. Kummert. Knowledge based systems for speech understanding. In H. Niemann, M. Lang, and G. Sagerer, editors, Recent Advances in Speech Understanding and Dialog Systems, pages 421--458. NATO ASI Series F, Vol. 46, SpringerVerlag, Berlin, 1988.

....5 provide the bottom up and top down propagation of constraints in the network. 3. 2 Semantic Network Representation of the Linguistic Knowledge For the representation of the linguistic knowledge a homogeneous hierarchical knowledge base using the above described system ERNEST was created (see [28]) An overview is given in Figure 5. Figure 5: Overview of the network for the interpretation of inquiries about German intercity train connections. Each block in this figure stands for a collection of concepts having an identical level of abstraction and an identical depth in the specialization ....

....in the context of linguistic processing is a collection of word and word chain hypotheses with a linguistic interpretation. Each of these hypotheses, represented in our system by a search tree node, has a judgement vector with the following components. Full details of the judgements are given in [21, 28, 27]. ffl Structural compatibility: This is a binary measure, which tests the linguistic restrictions, i.e. congruence of case, number, and gender in a noun group. That means: z(H) 8 : 1; if all restrictions are fulfilled by H 0; otherwise ffl Acoustic quality of the underlying word or ....

[Article contains additional citation context not shown here]

G. Sagerer and F. Kummert. Knowledge based systems for speech understanding. In H. Niemann, M. Lang, and G. Sagerer, editors, Recent Advances in Speech Understanding and Dialog Systems, pages 421--458. NATO ASI Series F, Vol. 46, Springer-Verlag, Berlin, 1988.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC