Na področju govornih in jezikovnih tehnologij predstavlja avtomatsko razpoznavanje govora enega izmed ključnih gradnikov. V prispevku bomo predstavili razvoj avtomatskega razpoznavalnika slovenskega ...govora za domeno dnevnoinformativnih oddaj. Arhitektura sistema je zasnovana na globokih nevronskih mrežah. Pri tem smo ob upoštevanju razpoložljivih govornih virov izvedli modeliranje z različnimi aktivacijskimi funkcijami. V postopku razvoja razpoznavalnika govora smo preverili tudi, kakšen je vpliv izgubnih govornih kodekov na rezultate razpoznavanja govora. Za učenje razpoznavalnika govora smo uporabili bazi UMB BNSI Broadcast News in IETK-TV. Skupni obseg govornih posnetkov je znašal 66 ur. Vzporedno z globokimi nevronskimi mrežami smo povečali slovar razpoznavanja govora, ki je tako znašal 250.000 besed. Na ta način smo znižali delež besed izven slovarja na 1,33 %. Z razpoznavanjem govora na testni množici smo dosegli najboljšo stopnjo napačno razpoznanih besed (WER) 15,17 %. Med procesom vrednotenja rezultatov smo izvedli tudi podrobnejšo analizo napak razpoznavanja govora na osnovi lem in F-razredov, ki v določeni meri pokažejo na zahtevnost slovenskega jezika za takšne scenarije uporabe tehnologije.
Interpolation of a spatially continuous variable from point samples is an important field in spatial analysis and surface models for geosciences. In this study, spatial interpolation methods which ...are Inverse Distance Weighted, Ordinary Kriging (OK), Modified Shepard's (MS), Multiquadric Radial Basis Function (MRBF) and Triangulation with Linear (TWL), and Multi-Layer Perceptron (MLP) which is an Artificial Neural Networks (ANN) method were compared in order to predict height for different point distributions such as curvature, grid random and uniform on a Digital Elevation Model which is an USGS National Elevation Dataset (NED). Errors of different interpolations and ANN prediction were evaluated for different point distributions and three different cross-sections on the characteristic parts of the surface were selected and analyzed Generally, OK MS, MRBF and TWL gave promising results and were more effective in terms of characteristics of surface than MLP and IDW. Although MLP simplified the contours obtained from predicted heights, it was a satisfactory predictor for curvature, grid, random and uniform distributions.
Full text
Available for:
IZUM, KILJ, NUK, ODKLJ, PILJ, PNG, SAZU, UL, UM, UPUK
SI: Navigacijski sistemi običajno temeljijo na sprejemnikih GNSS. Članek predstavlja trenutno stanje sistemov GNSS, navigacijo s tehnologijami GNSS in opisuje GPS-navigacijske instrumente. Nekatere ...naloge zahtevajo neprekinjeno navigacijo, česar ne moremo zagotoviti samo z uporabo GNSS-navigacije. Sistem za neprekinjeno navigacijo dopolnjujejo inercialni navigacijski sistemi. Predstavljene so osnove inercialne navigacije, opisane so inercialne merilne enote,navedeni tipi IMU in tipični pogreški inercialnih senzorjev. Predstavljeni sta dve metodi obdelave podatkov združenih sistemov GNSS/INS, tradicionalni Kalmanov filter in umetne nevronske mreže, ki po nekaterih raziskavah dosegajo boljše rezultate kot Kalmanov filter. Omenjeni sta še dve dodatni možnosti za izboljšanje ali celo zagotovitev neprekinjene navigacije, psevdoliti in širokopasovni radijski valovi. EN: Navigation systems are commonly based on GNSS receivers. The article presents the current status of GNSS, discusses GNSS navigation and describes GPS navigation instruments. Some applications require seamless navigation which cannot be provided byGNSS itself. The system for uninterrupted navigation uses Inertial Navigation System besides GNSS. The bases of inertial navigation, Inertial Measurement Units, types of IMU and typical inertial sensorerrors are presented. The processing of data from an integrated GNSS/INS is usually performed either by a traditional Kalman Filter or an artificial neural network. According to some of the research, the latter performs better than the conventional Kalman Filter.Seamless navigation can be improved or even made possible by the use of pseudolites or Ultra-wide Band.
Landscape classification is a demanding task,mainly because of andscape’s holistic nature. Landscape experts are able to intuitively recognize landscape units which, regarding the morphological and ...physical factors, prove unified and homogeneous. But with such »gestalt « perception only those landscape units can be recognized that differ from the wider area and are evidently distinguishable. The classification of usually smooth and fuzzy passages between them often presents a problem. Another frequent approach to landscape classification is a clear and repeatable parametric approach. Landscape units are defined by simple combination of physical factors. But landscape is too complex system for such simplification. In search of new methods of landscape classification the usability of artificial neural networks was tested. Their application was based on landscape samples that experts identified in the area of »Karst landscape of the interior of Slovenia « and classified into 7 landscape types. Based on their known locations and on their spatial characteristics, artificial neural networks were able to learn the general rules of spatial occurrence of landscape types and use them for typological classification of the remaining territory. Artificial neural networks proved to be a very useful tool, mainly because of their ability to learn and to generalize.
PDF
