{ "paper_id": "W02-0207", "header": { "generated_with": "S2ORC 1.0.0", "date_generated": "2023-01-19T05:12:55.753685Z" }, "title": "Annotating Semantic Consistency of Speech Recognition Hypotheses", "authors": [ { "first": "Iryna", "middle": [], "last": "Gurevych", "suffix": "", "affiliation": { "laboratory": "European Media Laboratory GmbH Schlo\u00df", "institution": "", "location": { "addrLine": "Wolfsbrunnenweg 33", "postCode": "D-69118", "settlement": "Heidelberg", "country": "Germany" } }, "email": "gurevych@eml.villa-bosch.de" }, { "first": "Robert", "middle": [], "last": "Porzel", "suffix": "", "affiliation": { "laboratory": "European Media Laboratory GmbH Schlo\u00df", "institution": "", "location": { "addrLine": "Wolfsbrunnenweg 33", "postCode": "D-69118", "settlement": "Heidelberg", "country": "Germany" } }, "email": "porzel@eml.villa-bosch.de" }, { "first": "Michael", "middle": [], "last": "Strube", "suffix": "", "affiliation": { "laboratory": "European Media Laboratory GmbH Schlo\u00df", "institution": "", "location": { "addrLine": "Wolfsbrunnenweg 33", "postCode": "D-69118", "settlement": "Heidelberg", "country": "Germany" } }, "email": "strube@eml.villa-bosch.de" } ], "year": "", "venue": null, "identifiers": {}, "abstract": "Recent work on natural language processing systems is aimed at more conversational, context-adaptive systems in multiple domains. An important requirement for such a system is the automatic detection of the domain and a domain consistency check of the given speech recognition hypotheses. We report a pilot study addressing these tasks, the underlying data collection and investigate the feasibility of annotating the data reliably by human annotators.", "pdf_parse": { "paper_id": "W02-0207", "_pdf_hash": "", "abstract": [ { "text": "Recent work on natural language processing systems is aimed at more conversational, context-adaptive systems in multiple domains. An important requirement for such a system is the automatic detection of the domain and a domain consistency check of the given speech recognition hypotheses. We report a pilot study addressing these tasks, the underlying data collection and investigate the feasibility of annotating the data reliably by human annotators.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Abstract", "sec_num": null } ], "body_text": [ { "text": "The complete understanding of naturally occurring discourse is still an unsolved task in computational linguistics. Several large research efforts are underway to build multidomain and multimodal information systems, e.g. the DARPA Communicator Program 1 , the SmartKom research framework 2 (Wahlster et al., 2001) , the AT&T interactive speech and multimodal user interface program 3 .", "cite_spans": [ { "start": 291, "end": 314, "text": "(Wahlster et al., 2001)", "ref_id": "BIBREF5" }, { "start": 383, "end": 384, "text": "3", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "Dialogue systems which deal with complex dialogues require the interaction of multiple knowledge sources, e.g. domain, discourse and user model (Flycht-Eriksson, 1999) . Furthermore NLP systems have to adapt to different environments and applications. This can only be achieved if the system is able to determine how well a given speech recognition hypothesis (SRH) fits within the respective domain model and what domain should be considered by the system currently in focus.", "cite_spans": [ { "start": 144, "end": 167, "text": "(Flycht-Eriksson, 1999)", "ref_id": "BIBREF1" } ], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "The purpose of this paper is to develop an annotation scheme for annotating a corpus of SRH with information on semantic consistency and domain specificity. We investigate the feasibility of an automatic solution by first looking at how reliably human annotators can solve the task.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "The structure of the paper is as follows: Section 2 gives an overview of the domain modeling component in the SmartKom system. In Section 3 we report on the data collection underlying our study. A description of the suggested annotation scheme is given in Section 4. Section 5 presents the results of an experiment in which the reliability of human annotations is investigated.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "The SmartKom research project (a consortium of twelve academic and industrial partners) aims at developing a multi-modal and multidomain information system. Domains include cinema information, home electronic device control, etc. A central goal is the development of new computational methods for disambiguating different modalities on semantic and pragmatic levels.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Domain Modeling in SmartKom", "sec_num": "2" }, { "text": "The information flow in SmartKom is organized as follows: On the input side the parser picks an N-best list of hypotheses out of the speech recognizer's word lattice (Oerder and Ney, 1993) . This list is sent to the media fusion component and then handed over to the intention recognition component.", "cite_spans": [ { "start": 166, "end": 188, "text": "(Oerder and Ney, 1993)", "ref_id": "BIBREF3" } ], "ref_spans": [], "eq_spans": [], "section": "Domain Modeling in SmartKom", "sec_num": "2" }, { "text": "The main task of intention recognition in SmartKom is to select the best hypothesis from the N-best list produced by the parser. This is then sent to the dialogue management component for computing an appropriate action. In order to find the best hypothesis, the intention recognition module consults a number of other components involved in language, discourse and domain analysis and requests confidence scores to make an appropriate decision (s. Fig. 1 ).", "cite_spans": [], "ref_spans": [ { "start": 449, "end": 455, "text": "Fig. 1", "ref_id": null } ], "eq_spans": [], "section": "Domain Modeling in SmartKom", "sec_num": "2" }, { "text": "Tasks of the domain modeling component are:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Domain Modeling in SmartKom", "sec_num": "2" }, { "text": "\u2022 to supply a confidence score on the consistency of SRH with respect to the domain model; \u2022 to detect the domain currently in focus.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Domain Modeling in SmartKom", "sec_num": "2" }, { "text": "These tasks are inherently related to each other: It is possible to assign SRH to certain domains only if they are consistent with the domain model. On the other hand, a consistency score can only be useful when it is given with respect to certain domains.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 1. Information flow", "sec_num": null }, { "text": "We consider semantic consistency scoring and domain detection a classification task. The question is whether it is feasible to solve this task automatically. As a first step towards an answer we reformulate the problem: automatic classification of SRH is possible only if humans are able to do that reliably.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Data", "sec_num": "3" }, { "text": "In order to test the reliability of such annotations we collected a corpus of SRH. The data collection was conducted by means of a hidden operator test (Rapp and Strube, 2002) . In the test the SmartKom system was simulated. We had 29 subjects prompted to say certain inputs in 8 dialogues. 1479 turns were recorded. Each user-turn in the dialogue corresponded to a single intention, e.g. route request or sights information request.", "cite_spans": [ { "start": 152, "end": 175, "text": "(Rapp and Strube, 2002)", "ref_id": "BIBREF4" } ], "ref_spans": [], "eq_spans": [], "section": "Data Collection", "sec_num": "3.1" }, { "text": "The data obtained from the hidden operator tests had to be prepared for our study to compose a corpus with N-best SRH. For this pur-pose we sent the audio files to the speech recognizer. The input for the domain modeling component, i.e. N-best lists of SRH were recorded in log-files and then processed with a couple of Perl scripts. The final corpus consisted of ca. 2300 SRH. This corresponds to ca. 1.55 speech recognition hypotheses per user's turn.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Data Preprocessing", "sec_num": "3.2" }, { "text": "The SRH corpus was then transformed into a set of annotation files which could be read into MMAX, the annotation tool adopted for this task (Mueller and Strube, 2001 ).", "cite_spans": [ { "start": 140, "end": 165, "text": "(Mueller and Strube, 2001", "ref_id": "BIBREF2" } ], "ref_spans": [], "eq_spans": [], "section": "Data Preprocessing", "sec_num": "3.2" }, { "text": "For our study, a markable, i.e. an expression to be annotated, is a single SRH. The annotators as well as the domain modeling component in SmartKom currently do not take the dialogue context into account and do not perform context-dependent analysis. Hence, we presented the markables completely out of dialogue order and thus prevented the annotators from interpreting SRH contextdependently.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Annotation Scheme", "sec_num": "4" }, { "text": "In the first step, the annotators had to classify markables with respect to semantic consistency. Semantic consistency is defined as wellformedness of an SRH on an abstract semantic level. We differentiate three classes of semantic consistency: consistent, semi-consistent, or inconsistent. First, all nouns and verbs contained in the hypothesis are extracted and corresponding concepts are retrieved from a lemma-concept dictionary (lexicon) supplied for the annotators. The decision regarding consistency, semi-consistency and inconsistency has to be done on the basis of evaluating the set of concepts corresponding to the individual hypothesis.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Semantic Consistency", "sec_num": "4.1" }, { "text": "\u2022 Consistent means that all concepts are semantically related to each other, e.g. \"ich moechte die kuerzeste Route\" 4 is mapped to the concepts \"self\", \"wish\", \"route\" all of which are related to each other. Therefore the hypothesis is considered consistent. meaningful. For example, the hypothesis \"ich moechte das Video sind\" 5 is considered semi-consistent as the fragment \"ich moechte das Video\", i.e. a set of corresponding concepts \"self\", \"want\", \"video\" is semantically wellformed. \u2022 Inconsistent hypotheses are those whose conceptual mappings are not semantically related within the domain model. E.g. \"ich wuerde die Karte ja Wiedersehen\" 6 is conceptualized as \"self\", \"map\", \"parting\". This set of concepts does not semantically make sense and the hypothesis should be rejected.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Semantic Consistency", "sec_num": "4.1" }, { "text": "One of our considerations was that it is principally not always feasible to detect domains from an SRH. This is because the output of speech recognition is often corrupt, which may, in many cases, lead to false domain assignments. We argue that domain detection is dependent on the semantic consistency score. Therefore, according to our annotation scheme no domain analysis should be given to the semantically inconsistent SRH. If the hypothesis is considered either consistent or semi-consistent, certain domains will be assigned to it. The list of SmartKom domains for this study is finite and includes the following: route planning, sights information, cinema information, electronic program guide, home electronic device control, personal assistance, interaction management, small-talk and off-talk.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Domain Detection", "sec_num": "4.2" }, { "text": "In some cases multiple domains can be assigned to a single markable. The reason is that some domains are inherently so close to each other, e.g. cinema information and electronic program guide, that the distinction can only be made when the context is taken into account. As this is not the case for our study we allow for the specification of multiple domains per SRH.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Domain Detection", "sec_num": "4.2" }, { "text": "To measure the reliability of annotations we used the Kappa statistic (Carletta, 1996) .", "cite_spans": [ { "start": 70, "end": 86, "text": "(Carletta, 1996)", "ref_id": "BIBREF0" } ], "ref_spans": [], "eq_spans": [], "section": "The Kappa Statistic", "sec_num": "5.1" }, { "text": "The value of Kappa statistic (K) for semantic consistency in our experiment was 0.58, which shows that there was not a high level of agreement between annotators 7 . In the field of content analysis, where the Kappa statistic originated, K>0.8 is usually taken to indicate good reliability, 0.68Domain%Route planning33,1Sights info13,3Cinema info10,8Electr. Program guide15,9Home device control12,0Personal assistance1,1Interaction Management13,1Other0,7Figure 2. Distribution of Classes", "html": null, "type_str": "table", "num": null, "text": "." } } } }