RESUMEN
Parallel texts represent a very valuable resource in many applications of natural language processing. The fundamental step in creating parallel corpus is the alignment. Sentence alignment is the issue of finding correspondence between source sentences and their equivalent translations in the target text. A number of automatic sentence alignment approaches were proposed including neural networks, which can be divided into length-based, lexicon-based, and translation-based. In our study, we used five different aligners, namely Bilingual sentence aligner (BSA), Hunalign, Bleualign, Vecalign, and Bertalign. We evaluated both, the performance of the Bertalign in terms of accuracy against the up to now employed aligners as well as among each other in the language pair English-Sovak. We created our custom corpus consisting of texts collected in 2021 and 2022. Vecalign and Bertalign performed statistically significantly best and BSA the worst. Hunalign and Bleualign achieved the same performance in terms of F1 score. However, Bleualign achieved the most diverse results in terms of performance.
Asunto(s)
Lenguaje , Procesamiento de Lenguaje Natural , Eslovaquia , Redes Neurales de la ComputaciónRESUMEN
PURPOSE: To determine the rate of opacification of hydrophilic MemoryLens U940A intraocular lenses (IOLs) (Mentor Ophthalmics, Inc.) in the given cohort and perform a histopathological and spectrophotometer analysis of 2 explanted opacified IOLs. SETTING: Ophthalmology Department, Faculty Hospital, Nitra, Slovakia. METHOD: This retrospective study comprised 182 patients (205 eyes) who had implantation of a MemoryLens U940A IOL from June 1997 to June 2000. The patients were examined using a slitlamp to detect the presence of IOL opacification. In 4 cases, the lenses were explanted because of significant opacification and patient-reported problems; 2 lenses were provided for further analysis. One unused reference MemoryLens U940A IOL was also evaluated. All IOL were stained with von Kossa to determine the presence of calcium in the opacification. To confirm the components presence of an ultraviolet (UV) absorber, the IOLs were examined with an Avatar 330 Fourier transfer infrared (IR) spectroscope and a UV visible spectrophotometer (Philips). The IR spectrums for the IOL were identified using an IR spectrum atlas. The opacified IOLs, reference IOL, and the IOL packaging were further examined to determine the presence of silicone. RESULTS: Various amounts of opacification were found on the MemoryLens U940A IOL in 30 eyes (30 patients) (14.63%). Two explanted IOLs were positive for von Kossa staining, proving the presence of calcium deposits; the reference lens staining was negative. Spectrophotometry showed that the reference IOL and opacified IOLs were of the same polymer. The presence of the UV absorber on the benzophenone base was seen in the reference lens but not the opacified IOLs. In contrast, an increased concentration of low-molecular-weight components generated during the degradation of the polymer was present in the opacified lenses. The white cover pf the IOL is of polydimethyl siloxane, a silicone rubber. However, no silicone rubber was present in any examined lens, perhaps because the IOLs were in contact with alcohol during the histopathologic examination. CONCLUSIONS: The results indicate opacification of the hydrophilic MemoryLens U940A was caused by premature consumption of the UV absorber in the polymer component of the IOLs optic, with a subsequent degradation of the polymer. Whether silicone from the white cover led to the IOL opacification, as reported with other types of hydrophilic IOLs, could not be confirmed.