Prosodic Feature-Based Discriminatively Trained Low Resource Speech Recognition System

Speech recognition has been an active field of research in the last few decades since it facilitates better human–computer interaction. Native language automatic speech recognition (ASR) systems are still underdeveloped. Punjabi ASR systems are in their infancy stage because most research has been conducted only on adult speech systems; however, less work has been performed on Punjabi children’s ASR systems. This research aimed to build a prosodic feature-based automatic children speech recognition system using discriminative modeling techniques. The corpus of Punjabi children’s speech has various runtime challenges, such as acoustic variations with varying speakers’ ages. Efforts were made to implement out-domain data augmentation to overcome such issues using Tacotron-based text to a speech synthesizer. The prosodic features were extracted from Punjabi children’s speech corpus, then particular prosodic features were coupled with Mel Frequency Cepstral Coefficient (MFCC) features before being submitted to an ASR framework. The system modeling process investigated various approaches, which included Maximum Mutual Information (MMI), Boosted Maximum Mutual Information (bMMI), and feature-based Maximum Mutual Information (fMMI). The out-domain data augmentation was performed to enhance the corpus. After that, prosodic features were also extracted from the extended corpus, and experiments were conducted on both individual and integrated prosodic-based acoustic features. It was observed that the fMMI technique exhibited 20% to 25% relative improvement in word error rate compared with MMI and bMMI techniques. Further, it was enhanced using an augmented dataset and hybrid front-end features (MFCC + POV + Fo + Voice quality) with a relative improvement of 13% compared with the earlier baseline system.

Assessing Signed Language Skills in Bi-/Multilingual, Deaf and Hard of Hearing Children

Providing appropriate and evidence-based services for linguistically diverse, bilingual, and multimodal deaf and/or hard-of-hearing (D/HH) children is an important aspect of education and intervention programs. The increasing linguistic diversity of D/HH children creates challenges for professionals who need to understand and document these children’s linguistic systems through appropriate and informative assessment of children’s speech, and spoken language, and/or sign language skills. However, assessment can prove challenging, particularly when a child’s language environment contains more than one language in one or more modality. In this chapter, areas of assessment relevant to bilingual D/HH children are discussed to guide practitioners’ decisions on selecting and using appropriate assessment materials and approaches. Special focus is given to assessment considerations that can inform practice when assessment resources for a particular language are not available.

Factors for developing children's speech through didactic games

The article focuses on the development of speech skills, thinking skills, vocabulary through didactic games in preschool and family atmosphere. we should not forget about the modern words that enter the children's speech and vocabulary under the influence of cartoons, movies and television. In the process of working on the lexicon, it is necessary to control the compliance of children's speech with the form of pure literary language, the absence of foreign words in the speech.

Developmental change in children’s speech processing of auditory and visual cues: An eyetracking study

This study investigates how children aged two to eight years (N = 129) and adults (N = 29) use auditory and visual speech for word recognition. The goal was to bridge the gap between apparent successes of visual speech processing in young children in visual-looking tasks, with apparent difficulties of speech processing in older children from explicit behavioural measures. Participants were presented with familiar words in audio-visual (AV), audio-only (A-only) or visual-only (V-only) speech modalities, then presented with target and distractor images, and looking to targets was measured. Adults showed high accuracy, with slightly less target-image looking in the V-only modality. Developmentally, looking was above chance for both AV and A-only modalities, but not in the V-only modality until 6 years of age (earlier on /k/-initial words). Flexible use of visual cues for lexical access develops throughout childhood.

Lexical and Grammatical Errors in Developmentally Language Disordered and Typically Developed Children: The Impact of Age and Discourse Genre

Persistent lexical and grammatical errors in children’s speech are usually recognized as the main evidence of language delay or language disorder. These errors are usually treated as a sign of a deficit in language competence. On the other hand, some studies have revealed the same kinds of grammatical errors in children with developmental language disorder (DLD) and in typically developed (TD) children. Quite often, DLD children use grammatical markers properly, but sometimes they do this erroneously. It has been suggested that the main area of the limitations in DLD children is language performance but not language competence. From the perspective of the resource deficit model, the error rate in DLD children should be influenced by the cognitive demands of utterance and text production. We presume that different genres of discourse demand a different number of cognitive resources and, thus, should differently impact the error rate in children’s speech production. To test our hypothesis, we carried out an error analysis of two corpora of child discourse. The first corpus contained longitudinal data of discourse (personal narratives, fictional stories, chats, and discussions) collected from 12 children at four age points (4 years 3 months., 4 years 8 months., 5 years 3 months., and 5 years 9 months. years). Another corpus contained discourse texts (fictional stories and discussions) collected in the framework of a cross-sectional study from 6-year-old TD and DLD children; the DLD children had language expression but not comprehension difficulties. A comparative analysis between different discourse genres evidenced that the genre of discourse and age of assessment impacted the error distribution in the DLD and TD children. Such variables as the lexical and morphological error rates were impacted the most significantly. The results of the two studies confirmed our hypothesis regarding the probabilistic nature of lexical and grammatical errors in both DLD and TD children and the relationship between a cognitive loading of the genre and the error rate.

Spontaneous speech intelligibility: early cochlear implanted children versus their normally hearing peers at seven years of age

Speaking intelligibly is an important achievement in children’s language development. How far do congenitally severe-to-profound hearing-impaired children who received a cochlear implant (CI) in the first two years of their life advance on the path to intelligibility in comparison to children with typical hearing (NH)? Spontaneous speech samples of children with CI and children with NH were orthographically transcribed by naïve transcribers. The entropy of the transcriptions was computed to analyze their degree of uniformity. The same samples were also rated on a continuous rating scale by another group of adult listeners. The transcriptions of the NH children’s speech were more uniform, i.e., had significantly lower entropy, than those of the CI children, suggesting that the latter group displayed lower intelligibility. This was confirmed by the ratings on the continuous scale. Despite the relatively restricted age ranges, older children reached better intelligibility scores in both groups.

Feature generalization in Dutch–German bilingual and monolingual children’s speech production

Dutch and German employ voicing contrasts, but Dutch lacks the ‘voiced’ dorsal plosive /ɡ/. We exploited this accidental phonological gap, measuring the presence of prevoicing and voice onset time durations during speech production to determine (1) whether preliterate bilingual Dutch–German and monolingual Dutch-speaking children aged 3;6–6;0 years generalized voicing to /ɡ/ in Dutch; and (2) whether there was evidence for featural cross-linguistic influence from Dutch to German in bilingual children, testing monolingual German-speaking children as controls. Bilingual and monolingual children’s production of /ɡ/ provided partial evidence for feature generalization: in Dutch, both bilingual and monolingual children either recombined Dutch voicing and place features to produce /ɡ/, suggesting feature generalization, or resorted to producing familiar /k/, suggesting segment-level adaptation within their Dutch phonological system. In German, bilingual children’s production of /ɡ/ was influenced by Dutch although the Dutch phoneme inventory lacks /ɡ/. This suggests that not only segments but also voicing features can exert cross-linguistic influence. Taken together, phonological features appear to play a crucial role in aspects of bilingual and monolingual children’s speech production.

Heart-moving stories about kitties in Runet in the paradigm of the genres of volunteer discourse

The article deals with heart-moving stories about kitties published on the Internet by the Moscow adoption center for cats “Murkosha”, which are studied in connection with the dominants of the volunteer discourse. The article focuses on speech genre and narrative characteristics of these stories. It is shown that their specific language, text, and other features are secondary to the main goal – to influence the reader, to induce him either to take the cat directly or to help the adoption center financially. Accordingly, the means that make it possible to enhance the impact come to the fore: expression, different types of direct appeal to moral imperatives, playing with values; at the syntax level – an abundance of direct and indirect directives. The author shows some speechgenre connections of the heart-moving stories about kitties with other genres of volunteer and non-volunteer (advertising) discourse: they are united by the role of indirect communication in text-building and (usually not directly named) practical illocutionary goal; but they demostrate a different attitude to the values that the author of the heart-moving stories about kitties operates (among them the main one is the increase in the amount of love and goodness in the world). The cat is presented in a humanized form (the method of personification) (hence, and not only from “advertising” intentionality, there is a lot of indirect communication, including metaphors, pastiche of various “human” genres), most often as a child. Hence – many “children’s” genres, imitation of the features of children’s speech and speech of adults in communication with children. The method of personification determines the greatest variety of expressive means when describing the “most metaphorical” components of a cat’s image (frame actants): the “character” of the cat and the “communication” (friendship, love) of the cat and the owner. The article analyzes illocutionary types of heart-moving stories about kitties, identified by the authors themselves and marked with smilies (emergency message for help; the story of a cat that entered the adoption center and is ready to be handed over to the future owner; a message about the need for especially careful treatment of a cat with physical or psychological problems; an adoption center that found a family; “letter from home” from new owners) and narrative types (narrative with partially expressed authorship; mixed (authorship) narratives; pastiche of the “dialogue” of a cat with an adoption center employee; “narrative” on behalf of a cat, etc.). A separate micro-study deals with the heart-moving stories frame structure, where the actants / slots are distinguished: a cat (external data, diseases, and other physical and / or psychological problems, “psychology” of a cat); a past owner of a cat, a new or future owner of a cat; street, street life, homelessness, dangers; adoption center for cats “Murkosha” and its staff. There is a characteristic of the use and distribution of linguistic means (primarily expressive: metaphors, especially – conceptual metaphors, definitions, including applications, epithets, etc.) by frame actants and slots. In particular, it is shown that the image of a cat is formed by three meaningful dominants (concepts): love-friendship (as an indissoluble unity) (hence the image of communication happiness), orphanhood and doing good. The latter corresponds to the dominant of the volunteer discourse. Of the two remaining, orphanhood is well combined with it (the targeting of doing good is emphasized), but the latter rather contradicts it (it is more likely mutually beneficial cooperation, even exchange, than disinterested service) and can probably be explained by the focus on the diversity of reaching a heterogeneous audience, where the motives to take a cat from a shelter can also be different.