National Institute of Standards and Technology
Manufacturer’s Suggested Retail Price
The mode is the value that appears most often in a set of data.
Meeting Asistant – Phonexia research project
Median – Value separating higher half of data sample from lower half.
Mean Absolute Error – In statistics, the mean absolute error (MAE) is a quantity used to measure how close forecasts or predictions are to the eventual outcomes.
Large Vocabulary Continuous Speech Recognition
Log-Likelihood Ratio – Logarithmic function of LR (le)
Likelihood Ratio – Result from statistical test for two models comparation. It gives back number which expresses how many times more likely the data are under one model than the other. LR meets numbers in interval <-∞;+∞>
Language Print Archive – pack of language prints from the recordings spoken in the same language/dialect. Used for the language identification in LID comparison.
Language Print – output data from LID technology
Language Model (“vocabulary” in STT technology)
Phonexia Language Identification, multiple generations available marked by version like
Law Enforcement Agency
Linguistic Data Consortium (Fisher English etc.) – Group of libraries, universities,… working with speech data.
Phonexia Keyword Spotting – acoustics based several technologies possible, language dependent
Interactive Voice Response – Process of caller authentication (typically banking segment)
iVector – Unique value representing specific group of voice tracks. Consist of 600 values, usually lowered to low dimensional vector (200 values)
Intelligence Support Systems for lawfull interception, criminal investigations, and intelligence gathering. Really important International conference, invitation only.
High Perfomance Computing
Horizon 2020 – EU program for inovations and developement
Graphical User Interface
Global system for mobile communications – Mobile phone protocol
Gausian mixture model – Statistical model for representing values.
Phonexia Gender Identification
Faster than the Real Time – processing speed on 1 instance of the speech technology. Can be described also as “X min of audio archive to be processed in Y min of the real time”.
False Rejection Ratio – Values, which were not accepted, when they were supposed to be accepted, aka ‘misses’
Features – FEA is optional output from KWS technology. Looking for keywords in FEA is faster than in original recording.