scholarly journals Probabilistic Top-Down Parsing and Language Modeling

2001 ◽  
Vol 27 (2) ◽  
pp. 249-276 ◽  
Author(s):  
Brian Roark

This paper describes the functioning of a broad-coverage probabilistic top-down parser, and its application to the problem of language modeling for speech recognition. The paper first introduces key notions in language modeling and probabilistic parsing, and briefly reviews some previous approaches to using syntactic structure for language modeling. A lexicalized probabilistic top-down parser is then presented, which performs very well, in terms of both the accuracy of returned parses and the efficiency with which they are found, relative to the best broad-coverage statistical parsers. A new language model that utilizes probabilistic top-down parsing is then outlined, and empirical results show that it improves upon previous work in test corpus perplexity. Interpolation with a trigram model yields an exceptional improvement relative to the improvement observed by other models, demonstrating the degree to which the information captured by our parsing model is orthogonal to that captured by a trigram model. A small recognition experiment also demonstrates the utility of the model.

With the Internet and the World Wide Web revolution, large corpora in variety of forms are germinating ceaselessly that can be manifested as big data. One obligatory area for the usage of such large corpora is language modeling for large vocabulary continuous speech recognition. Language modeling is an indispensable module in speech recognition architecture, which plays a vital role in reducing the search space during the recognition process. Additionally, the language model that is contiguous to the domain of the speech can dwindle the search space and escalate the recognition accuracy. In this paper, an efficient searching mechanism for domain-specific document retrieval from the large corpora has been elucidated using Elasticsearch which is a distributed and an efficient search engine for big data. This assisted us in tuning the language model in accordance with the domain and also by reducing the search time by more than 90% in comparison to conventional search and retrieval mechanism used in our earlier work. A word level and a phrase level retrieval process for creating domain-specific language model has been implemented. The evaluation of the system is performed on the basis of word error rate (WER) and perplexity (PPL) of the speech recognition system. The results shows nearly 10% decrease on WER and a major reduction in the PPL that helped in boosting the performance of the speech recognition process. From the results, it can be consummated that Elasticsearch is an efficient mechanism for domain specific document retrieval from large corpora rather than using topic modeling toolkits


2019 ◽  
Vol 2019 ◽  
pp. 1-8 ◽  
Author(s):  
Edvin Pakoci ◽  
Branislav Popović ◽  
Darko Pekar

Serbian is in a group of highly inflective and morphologically rich languages that use a lot of different word suffixes to express different grammatical, syntactic, or semantic features. This kind of behaviour usually produces a lot of recognition errors, especially in large vocabulary systems—even when, due to good acoustical matching, the correct lemma is predicted by the automatic speech recognition system, often a wrong word ending occurs, which is nevertheless counted as an error. This effect is larger for contexts not present in the language model training corpus. In this manuscript, an approach which takes into account different morphological categories of words for language modeling is examined, and the benefits in terms of word error rates and perplexities are presented. These categories include word type, word case, grammatical number, and gender, and they were all assigned to words in the system vocabulary, where applicable. These additional word features helped to produce significant improvements in relation to the baseline system, both for n-gram-based and neural network-based language models. The proposed system can help overcome a lot of tedious errors in a large vocabulary system, for example, for dictation, both for Serbian and for other languages with similar characteristics.


2019 ◽  
Author(s):  
Hendi

Perkembangan studi linguistik di dalam penafsiran teks kitab suci memang masih lambat dan dalam tahap perkembangan. Tulisan ini adalah suatu penelitian untuk mengembangkan studi linguistik di dalam penafsiran teks kitab suci di Indonesia. Penulis memilih studi linguistik dengan pendekatan analisis wacana. Model analisis wacana yang digunakan adalah analisis colon yang diperkenalkan oleh Johanes P. Louw. Sampel teks yang digunakan adalah surat Filemon. Hasil penelitian ini akan mendapatkan struktur dan tema surat Filemon. Pendekatan analisis wacana menekankan semantik. Arti atau makna di dalam teks melampaui kata, frasa, dan kalimat (struktur mikro teks) sehingga fokus analisis sampai kepada keseluruhan wacana (struktur makro teks). Struktur makro teks melingkupi arti dari struktur mikro teks. Seorang penulis menulis teks mulai dari ide wacana yang kemudian secara sadar membangun ide tersebut dari struktur mikro teks yang dipilihnya. Dalam analisis colon, struktur makro teks yang terpenting adalah paragraf yang merupakan satu unit semantis yang dibangun dari beberapa kelompok colon (cluster) dan atau colons. Arti kata, frasa, klausa, dan kalimat tidak lepas dari isi semantis paragraf yang mewadahinya. Sementara, penafsir-penafsir lain lebih memprioritaskan penafsiran struktur mikro teks daripada makro teks. Wacana dianalisis mulai dari paragraf sampai frasa dan kata (top down). Unit semantis dalam bentuk kata, frasa, dan klausa akan dianalisis dengan kategori semantis, pengelompokkan kata (grouping of words atau immediate constituents), dan transformasi struktur luar (surface structure) ke dalam struktur dalam (deep structure). Unit semantis dalam bentuk paragraf akan dianalisis dengan metode analisis colon. Di dalam analisis colon, ada beberapa langkah yang akan diuraikan yaitu pertama, membuat struktur colon (syntactic structure) dari setiap paragraf dan terjemahan literal. Pengelompokkan kata akan terlihat di dalam struktur colon. Kedua, mencari isi semantis dari setiap colon atau kelompok (cluster) dengan menganalisis kata, frasa, dan colon. Ketiga, mencari hubungan semantis di antara colon atau kelompok di dalam satu paragraf yang sama. Keempat, menentukan tema atau ide utama (the pivot point) dari setiap paragraf. Berdasarkan analisis colon, ide utama atau tema surat ini adalah permohonan Rasul Paulus kepada Filemon untuk mengembalikan atau menerima kembali Onesimus sebagai saudara di dalam Kristus. Tema wacana ini menentukan struktur makro dan mikro teks ditulis oleh Rasul Paulus. Rasul Paulus mulai menulis dengan pembukaan yaitu sapaan dan salam kepada Filemon dan seluruh jemaatnya. Kemudian, Rasul Paulus menuliskan dasar permohonannya yaitu iman dan kasih Filemon yang selama ini sudah didengar olehnya. Lalu, Rasul Paulus menuliskan permohonannya bahwa Filemon bisa menerima kembali Onesimus. Terakhir, Rasul Paulus menuliskan penutup yaitu salam dan doa berkat kepada seluruh jemaat. Implikasi pastoral atau teologis yang bisa dipelajari adalah cara iman dan kasih diterapkan secara nyata di dalam persekutuan dan kehidupan seperti pengampunan dan rekonsiliasi relasi dengan orang lain yang sudah berbuat dosa. Pengalaman jatuh ke dalam dosa dan dipulihkan oleh Allah adalah pengalaman yang tidak mungkin dipisahkan dalam hidup ini. Oleh karena itu, persekutuan sesama orang percaya menjadi wadah atau alat anugerah bagi setiap orang percaya menghadapi berbagai godaan dosa. Secara khusus, penulis mengucap terima kasih kepada para mahasiswa STT Soteria Purwokerto terutama mereka yang sudah mengikuti kelas Studi dan Exegesis Perjanjian Baru. Mereka adalah orang (pembaca) pertama yang bersama penulis menggumuli teks ini selama 1 semester. Penulis juga mengucap terima kasih kepada isteri, Rina Mansyur, dan puteri, Filipe File Cendekia atas dukungan yang tiada taranya. Terakhir, penulis mengucapkan terima kasih kepada penerbitan Leutikaprio yang sudah bersedia mengedit dan menerbitkan buku ini.


Author(s):  
Zhong Meng ◽  
Sarangarajan Parthasarathy ◽  
Eric Sun ◽  
Yashesh Gaur ◽  
Naoyuki Kanda ◽  
...  

2021 ◽  
Vol 11 (6) ◽  
pp. 2866
Author(s):  
Damheo Lee ◽  
Donghyun Kim ◽  
Seung Yun ◽  
Sanghun Kim

In this paper, we propose a new method for code-switching (CS) automatic speech recognition (ASR) in Korean. First, the phonetic variations in English pronunciation spoken by Korean speakers should be considered. Thus, we tried to find a unified pronunciation model based on phonetic knowledge and deep learning. Second, we extracted the CS sentences semantically similar to the target domain and then applied the language model (LM) adaptation to solve the biased modeling toward Korean due to the imbalanced training data. In this experiment, training data were AI Hub (1033 h) in Korean and Librispeech (960 h) in English. As a result, when compared to the baseline, the proposed method improved the error reduction rate (ERR) by up to 11.6% with phonetic variant modeling and by 17.3% when semantically similar sentences were applied to the LM adaptation. If we considered only English words, the word correction rate improved up to 24.2% compared to that of the baseline. The proposed method seems to be very effective in CS speech recognition.


2005 ◽  
Author(s):  
Chuang-Hua Chueh ◽  
To-Chang Chien ◽  
Jen-Tzung Chien

2013 ◽  
Author(s):  
Haşim Sak ◽  
Yun-hsuan Sung ◽  
Françoise Beaufays ◽  
Cyril Allauzen

2004 ◽  
Author(s):  
Dimitra Vergyri ◽  
Katrin Kirchhoff ◽  
Kevin Duh ◽  
Andreas Stolcke

Sign in / Sign up

Export Citation Format

Share Document