Keyword verification considering the correlation of succeeding feature vectors

Junkawitsch, Jochen; Höge, Harald

doi:10.1109/icassp.1998.674407

Cited by 3 publications

(2 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A search of the traditional ROC literature [Swets 1988;Egan 1975;Swets and Pickett 1982] shows no mention of this formulation. It does appear, without comment or justification in the wordspotting literature [James and Young 1994;Jeanrenaud et al 1994;Lippmann et al 1994;Junkawitsch et al 1998;Dharanipragada and Roukos 1998], where it is usually, but not always, referred to as a ROC curve. We speculated that false alarms per unit time might be a surrogate measure for percent false alarms under the assumption that spoken word rates are approximately constant (at least compared to Internet traffic rates) across many speakers or passages and we were concerned that there might be assumptions associated with the word-spotting usage that would not hold for intrusion detection.…”

Section: Errors Per Unit Timementioning

confidence: 99%

Testing Intrusion detection systems

McHugh

2000

ACM Trans. Inf. Syst. Secur.

1,035

272

View full text Add to dashboard Cite

In 1998 and again in 1999, the Lincoln Laboratory of MIT conducted a comparative evaluation of intrusion detection systems (IDSs) developed under DARPA funding. While this evaluation represents a significant and monumental undertaking, there are a number of issues associated with its design and execution that remain unsettled. Some methodologies used in the evaluation are questionable and may have biased its results. One problem is that the evaluators have published relatively little concerning some of the more critical aspects of their work, such as validation of their test data. The appropriateness of the evaluation techniques used needs further investigation. The purpose of this article is to attempt to identify the shortcomings of the Lincoln Lab effort in the hope that future efforts of this kind will be placed on a sounder footing. Some of the problems that the article points out might well be resolved if the evaluators were to publish a detailed description of their procedures and the rationale that led to their adoption, but other problems would clearly remain.

show abstract

Section: Errors Per Unit Timementioning

confidence: 99%

Testing Intrusion detection systems

McHugh

2000

ACM Trans. Inf. Syst. Secur.

1,035

272

View full text Add to dashboard Cite

show abstract

“…A number of research results have already been published, e.g. in language identification [8], multilingual recognition [2,3], speaker verification [16] and general acoustic-phonetic modelling and adaptation for different environments and tasks [4,5,9,10,11,12,14,15]. Apart from this, the SpeechDat databases also represent a valuable collection of dialects and speakers for corpus-based linguistic and phonetic studies [1,6,7,13].…”

Section: Ev a Luation O F Th E Use O F The Speech D A T Databasesmentioning

confidence: 99%

Linguistic phrase spotting in a simple application spoken dialogue system

Hoege

Draxler²,

Heuvel³

et al. 1999

6th European Conference on Speech Communication and Technology

View full text Add to dashboard Cite

The goal of the SpeechDat project is to develop spoken language resources for speech recognisers suited to realise voice driven teleservices. SpeechDat created speech databases for all official languages of the European Union and some major dialectal varieties and minority languages. The size of the databases ranges between 500 and 5000 speakers. In total 20 databases are recorded over the fixed telephone network, 5 databases over the cellular network, and 3 databases are designed for speaker verification. To date the project has successfully reached its end. This paper briefly describes the project, addresses the validation of the databases, their availability to consortium members and third parties, publicity and awareness, and the spin-off of the project in speech recognition research.

show abstract