Search

Recent work asked if visual cues facilitate word segmentation in adults and infants (Thiessen, 2010). While adults showed better word segmentation when presented with a regular visual cue (consistent mapping between words and objects), infants did not. This difference was attributed to infants' lack of understanding that objects have labels. Alternatively, infants' performance could reflect their difficulty with tracking and integrating multiple multimodal cues. We contrasted these two accounts by looking at the effect of visual cues on word segmentation in adults and across childhood (6-12 years). We found that older children (M age 10;7) benefitted from the regular visual cues, but younger children (M age 7;10), who already knew that objects have labels, did not. Knowing that objects have labels was not enough to use visual cues as an aid for segmentation. These findings show that the ability to integrate multimodal cues develops during childhood, and it is not yet adult-like in children., (Copyright © 2017 Cognitive Science Society, Inc.)

Cochlear implant (CI) users frequently achieve good speech understanding based on phoneme and word recognition. However, there is a significant variability between CI users in processing prosody. The aim of this study was to examine the abilities of an excellent CI user to segment continuous speech using intonational cues. A post-lingually deafened adult CI user and 22 normal hearing (NH) subjects segmented phonemically identical and prosodically different sequences in French such as 'l'affiche' (the poster) versus 'la fiche' (the sheet), both [lafiʃ]. All participants also completed a minimal pair discrimination task. Stimuli were presented in auditory-only and audiovisual presentation modalities. The performance of the CI user in the minimal pair discrimination task was 97% in the auditory-only and 100% in the audiovisual condition. In the segmentation task, contrary to the NH participants, the performance of the CI user did not differ from the chance level. Visual speech did not improve word segmentation. This result suggests that word segmentation based on intonational cues is challenging when using CIs even when phoneme/word recognition is very well rehabilitated. This finding points to the importance of the assessment of CI users' skills in prosody processing and the need for specific interventions focusing on this aspect of speech communication.

In an eye-tracking experiment we examined whether Chinese readers were sensitive to information concerning how often a Chinese character appears as a single-character word versus the first character in a two-character word, and whether readers use this information to segment words and adjust the amount of parafoveal processing of subsequent characters during reading. Participants read sentences containing a two-character target word with its first character more or less likely to be a single-character word. The boundary paradigm was used. The boundary appeared between the first character and the second character of the target word, and we manipulated whether readers saw an identity or a pseudocharacter preview of the second character of the target. Linear mixed-effects models revealed reduced preview benefit from the second character when the first character was more likely to be a single-character word. This suggests that Chinese readers use probabilistic combinatorial information about the likelihood of a Chinese character being single-character word or a two-character word online to modulate the extent of parafoveal processing.

The present study explored the processing units of high-proficiency second language (L2) Chinese learners in on-line reading in an eye-tracking experiment. The critical aim was to investigate how learners segment continuous characters into words without the aid of word boundary demarcations. Based on previous studies, the embedded words of 2- and 3-character incremental words were manipulated to be either plausible or implausible with the preceding verbs, while the incremental words themselves were always plausible. The results revealed an effect of the plausibility manipulation, which suggested that L2 Chinese learners activated embedded words first and integrated embedded words with previous sentence context as soon as they read them. [ABSTRACT FROM AUTHOR]

Prosody is known to scaffold the learning of language, and thus understanding prosodic development is vital for language acquisition. The present study explored the unfolding prosody model of prosodic development (proposed in Frota's et al. study in 2016) beyond early production data, to examine whether it predicted the development of early segmentation abilities. European Portuguese-learning infants aged between 5 and 17 months were tested in a series of word segmentation experiments. Developing prosodic structure was evidenced in word segmentation as proposed by the unfolding model: (i) a simple monosyllabic word shape crucially placed at a major prosodic edge was segmented first, before more complex word shapes under similar prosodic conditions; (ii) the segmentation of more complex words was easier at a major prosodic edge than in phrase-medial position; and (iii) the segmentation of complex words with an iambic pattern preceded the segmentation of words with a trochaic pattern. These findings demonstrated that word segmentation evolved with unfolding prosody, suggesting that the prosodic units developed in the unfolding process are used both as speech production planning units and to extract word-forms from continuous speech. Therefore, our study contributes to a better understanding of the mechanisms underlying word segmentation, and to a better understanding of early prosodic development, a cornerstone of language acquisition. [ABSTRACT FROM AUTHOR]

We reported a large-scale Internet-based experiment to investigate the impact of emotion information on Chinese word segmentation, in which participants completed an overlapping ambiguous string (OAS) segmentation task and the Chinese version of Beck Depression Inventory-II in a counterbalanced order. OAS is a three-character string (ABC) in which the middle character can form a distinct word with both the character on its left side (word AB) and the character on its right side (word BC). Participants were presented with isolated OASs and were asked to report the word they identified first. Emotional OAS was constructed by a combination of a neutral word and an emotional word, with the neutral and emotional words sharing character B. We orthogonally manipulated the valence of the emotional words (positive vs. negative) and their position in the OAS (left-side vs. right-side). The results showed that compared with neutral words, both positive and negative words were more likely to be segmented, and this segmentation outcome was not affected by readers with different depression tendencies. These findings suggest that emotion information can influence word segmentation, and that both positive and negative words take precedence over neutral words in the word segmentation process. This study provides a new perspective and evidence to understand the impact of emotion information on word processing. [ABSTRACT FROM AUTHOR]

Word segmentation is a crucial step in children's vocabulary learning. While computational models of word segmentation can capture infants' performance in small‐scale artificial tasks, the examination of early word segmentation in naturalistic settings has been limited by the lack of measures that can relate models' performance to developmental data. Here, we extended CLASSIC (Chunking Lexical and Sublexical Sequences in Children; Jones et al., 2021), a corpus‐trained chunking model that can simulate several memory and phonological and vocabulary learning phenomena to allow it to perform word segmentation using utterance boundary information, and we have named this extended version CLASSIC utterance boundary (CLASSIC‐UB). Further, we compared our model to the performance of children on a wide range of new measures, capitalizing on the link between word segmentation and vocabulary learning abilities. We showed that the combination of chunking and utterance‐boundary information used by CLASSIC utterance boundary allowed a better prediction of English‐learning children's output vocabulary than did other models. A one‐page Accessible Summary of this article in non‐technical language is freely available in the Supporting Information online and at https://oasis‐database.org [ABSTRACT FROM AUTHOR]

Characters segmentation is a necessity and the most critical stage in Arabic OCR system. It has attracted the interest of a wide range of researchers. However, the nature of the Arabic cursive script poses extra challenges that need further investigation. Therefore, having a reliable and efficient Arabic OCR system that is independent of font variations is highly required. In this paper, an indirect, font-in dependent word and character segmentation algorithm for printed Arabic text investigated. The proposed algorithm takes a binary line image as an input and produces a set of binary images consisting of one character or ligature as an output. The segmentation performed at two levels: a word segmentation performed in the first level, by employing a vertical projection at the input line image along with using Interquartile Range (IQR) method to differentiate between word gaps and within word gaps. A projection profile method used as a second level of segmentation along with a set of statistical and topological features, which are font-independent, to identify the correct segmentation points from all potential points. The APTI dataset used to test the proposed algorithm with a variety of font type, size, and style. The algorithm experimented on 1800 lines (approximately 24,816 words) with an average accuracy of 97.7% for words segmentation and 97.51% for characters segmentation. [ABSTRACT FROM AUTHOR]

One of the several challenges faced by neural machine translation systems is the lack of standard parallel corpora for several language pairs. Poor translation qualities often result from inadequate data. Aggravating this problem further are the issues of morphological complexity and agglutination, leading to unmanageable vocabulary size, rare words and data sparsity issues. Though this problem has been partly addressed by sub-word algorithms such as BPE, translation systems still lag in their ability to understand sentence and word structures associated with rich morphologies. This paper aims to address these issues by employing linguistically driven sub-word units into NMT systems. This approach is further enhanced by additional POS tag feature inputs. The proposed approach outperforms BPE driven machine translation models by several BLEU points and is also shown to have better recall measures from evaluation by ROUGE metric. The results have been evaluated upon a morphologically complex Dravidian language pair, Kannada and Telugu. [ABSTRACT FROM AUTHOR]

Aiming at the problem of ambiguity caused by boundary segmentation errors caused by word segmentation in Chinese relation extraction, and the problem of multiple relations involved in the overlapping of entity pairs, this paper proposed joint extraction method based on word mixing. Firstly, for the word segmentation boundary problem, the embedding layer combined word vectors on the basis of word vectors, and added position information to ensure the correct order between words. Secondly, the model introduced a hybrid expanded convolutional network for feature extraction with different granularity and longer distance. Finally, it used the hierarchical labeling method to mark the corresponding relationship and object entity through the obtained main entity information, and each main entity could correspond to multiple relations and object entities. Compared with other relation extraction methods on the same Chinese data set, the experimental results show that the extraction effect of this method is the best and also shows better stability. [ABSTRACT FROM AUTHOR]

Sanskrit processing has seen a surge in the use of data-driven approaches over the past decade. Various tasks such as segmentation, morphological parsing, and dependency analysis have been tackled through the development of state-of-the-art models despite working with relatively limited datasets compared to other languages. However, a significant challenge lies in the availability of annotated datasets that are lexically, morphologically, syntactically, and semantically tagged. While syntactic and semantic tags are preferable for later stages of processing such as sentential parsing and disambiguation, lexical and morphological tags are crucial for low-level tasks of word segmentation and morphological parsing. The Digital Corpus of Sanskrit (DCS) is one notable effort that hosts over 650,000 lexically and morphologically tagged sentences from around 250 texts but also comes with its limitations at different levels of a sentence like chunk, segment, stem and morphological analysis. To overcome these limitations, we look at alternatives such as Sanskrit Heritage Segmenter (SH) and Saṃsādhanī tools, that provide information complementing DCS’ data. This work focuses on enriching the DCS dataset by incorporating analyses from SH, thereby creating a dataset that is rich in lexical and morphological information. Furthermore, this work also discusses the impact of such datasets on the performances of existing segmenters, specifically the Sanskrit Heritage Segmenter. [ABSTRACT FROM AUTHOR]

Bilingual infants must navigate the similarities and differences between their languages to achieve native proficiency in childhood. Bilinguals learning to find individual words in fluent speech face the possibility of conflicting cues to word boundaries across their languages. Despite this challenge, bilingual infants typically begin to segment and learn words in both languages around the same time as monolinguals. It is possible that early bilingual experience may support infants' abilities to track regularities relevant for word segmentation separately across their languages. In a dual speech stream statistical word segmentation task, we assessed whether 16‐month‐old infants could track syllable co‐occurrence regularities in two artificial languages despite conflicting information across the languages. We found that bilingual, but not monolingual, infants were able to segment the dual speech streams using statistical regularities. Although the two language groups did not differ on secondary measures of cognitive and linguistic development, bilingual infants' real‐world experience with bilingual speakers was predictive of their performance in the dual language statistical segmentation task. [ABSTRACT FROM AUTHOR]

This paper presents the text line and word segmentation from unconstrained handwritten documents based on horizontal projection histogram (HPH) to detect mid-points and gap trailing between lines. The midpoints are estimated from the HPH for the first 100 to 200 columns of the whole document. Then, considering the mid-points, the gap is tracked between two consecutive lines from locally computed HPH for a block having k rows and j columns. The HPH block is examined for various cases to locate optimal rows that separate adjacent lines. The proposed method segments curve, touching and skew-lines and is robust to writing variation and language independent. Word segmentation is not treated as a separate problem and goes efficiently alongside the line segmentation. As the trailing of space between neighboring lines goes on, the vertical projection Histogram (VPH) of t columns is monitored between the above and below separator of a line and find the optimal word separator. The algorithm is evaluated on two isolated datasets of different languages (Meitei Mayek and English). Text-line and word segmentation on Meitei Mayek handwritten documents achieve 91.84% and 88.96% accuracy respectively. Similarly, the handwritten English document meets 94.18% and 87.73% accuracy for line and word segmentation. [ABSTRACT FROM AUTHOR]

The present study sought to establish how a word's contextual predictability impacts the early stages of word processing when reading Chinese. Two eye-movement experiments were conducted in which the predictability of the target two-character word was manipulated; the frequency of the target's initial character was manipulated in Experiment 1, as was the target's end character frequency in Experiment 2. No reliable interaction effect of predictability with initial character frequency was observed in Experiment 1. Reliable interactions of word predictability with end character frequency were observed in Experiment 2. The end character frequency effects, in which the words with high-frequency end characters were fixated for a shorter time and re-fixated less often, were only observed when reading unpredictable words. Reliable interactions were also observed with incoming saccade length, as high-frequency end character words elicited longer saccades to themselves than low-frequency end character words when reading predictable words. The effects of pervasive predictability on measures of fixation time, probability, and saccade length were noted in both experiments. Our findings suggest that a word's contextual predictability facilitates the processing of its constituent characters. [ABSTRACT FROM AUTHOR]

Practices of word segmentation in French and Moroccan Arabic by beginning and advanced bilingual writers in two historically and linguistically divergent settings are analysed in a threefold perspective: (1) In the different sociocultural contexts of linguistically heterogeneous France in the 1870's and a town with remarkable immigration from Morocco in Germany in 2000, dictations constitute monolingual settings of language policy and normativity; (2) structurally, open and closed spellings of (clitic) function and content words indicate constraints of different orthographies, focussing either phonology or morphosyntax; (3) in the framework of contact linguistics, bilingual students write in one of their languages (French, Moroccan Arabic) with resources of other languages (like Breton, German, Classical Arabic). The results show that the students' writings are influenced by graphematic structures not directly related to the language dictated. In French Brittany, a great importance of closed spellings may be supported by the agglutinative feature of the Breton language, while the apostrophe as a striking feature of French orthography is used primarily, but often only emblematically, by the students in Gascony. Moroccan Arabic writers in Germany are influenced indirectly by their first school language, German, in the way they mark word boundaries in prepositional phrases (PP) and imperfective verb forms. Classical Arabic, however, remains of marginal influence although both varieties are historically and structurally closely related. [ABSTRACT FROM AUTHOR]

This paper proposes a new subword segmentation method for machine translation. The algorithm, which we call GenSeg, is generic in the sense that it can be applied to any language, but is designed with an emphasis on inflectional splitting, i.e. it attempts to split words on boundaries corresponding to inflectional suffixes. The main principle of the method is grouping together words that share a common middle substring, and then separating the best such substring from the rest of the word. GenSeg is a cross-language method extended with some language-specific morphological analysis rules (currently for the Latvian language). To verify its effectiveness, we performed machine translation experiments in two directions: Latvian-English and English-Latvian, obtaining minor improvements in translation quality when using our pre-processing method. [ABSTRACT FROM AUTHOR]

Infants' ability to segment words in fluent speech is affected by their language experience. In this study we investigated the conditions under which infants can segment words in a non-native language. Using the Head-turn Preference Procedure, we found that monolingual English-learning 8-month-olds can segment bisyllabic words in Spanish (trochees and iambs) but not French (iambs). Our results are incompatible with accounts that rely on distributional learning, language rhythm similarity, or target word prosodic shape alone. Instead, we show that monolingual English-learning infants are able to segment words in a non-native language as long as words have stress, as is the case in English. More specifically, we show that even in a rhythmically different non-native language, English-learning infants can find words by detecting stressed syllables and treating them as word onsets or offsets. [ABSTRACT FROM AUTHOR]

The majority of English nouns, verbs, and adjectives begin with a stressed syllable, and listeners exploit this tendency to help parse the continuous stream of speech into individual words. However, the acoustic manifestation of stress depends on the variety of English being spoken. In two visual world eye-tracking experiments, we tested if Indian English-accented speech causes Canadian English listeners to make stress-based segmentation errors. Participants heard Canadian- or Indian-accented trisyllabic sequences that could be segmented in two ways, depending on the perceived location of stress. For example, [hæ.pi.tsə] could be segmented as happy /[tsə] if it is perceived to have stress on the first syllable or as [hæ]/ pizza if it is perceived to have stress on the second syllable. Results suggest that Indian English-accented speech impairs segmentation in Canadian listeners, and that both accented pitch and other features of the Indian English accent contribute to segmentation difficulties. Findings are interpreted with respect to models of how similarity between two languages impacts the listener’s ability to segment words from the speech stream. [ABSTRACT FROM AUTHOR]

Word intuition is speakers' intuitive knowledge on wordhood. Collective word intuition is the word intuition of the whole language community. Given this definition, the optimal word segmentation result in Chinese NLP should reflect collective word intuition. It is also believed that an ideal definition of Chinese word should accord with the collective word intuition of Chinese speakers. To test the validity and feasibility of modeling collective word intuition, it is important to know to what extent Chinese speakers agree with each other on what is a word. In this study, we measured word intuition agreement using Mechanical Turk-based Chinese word segmentation experiment. Three metrics were used: proportionate agreement, Cohen's kappa, and Fleiss' kappa. The results show that Chinese speakers agree with each other almost perfectly on what is a word. And we found no evidence to support an effect of semantic transparency on word intuition agreement. Such high word intuition agreement among Chinese speakers supports the psychological reality of Chinese word and also suggests that that it is quite feasible to formulate a definition of Chinese word by modeling the collective word intuition of Chinese speakers. [ABSTRACT FROM AUTHOR]

Segmenting word units from running speech is a fundamental skill infants must develop in order to acquire language. Despite ample behavioral evidence of this skill, its neurocognitive basis remains unclear. Using behavioral testing and functional near-infrared spectroscopy, we aimed to uncover the neurocognitive substrates of word segmentation and its development. Of three age-groups of Japanese infants (5-6, 7-8, and 9-10months of age), the two older age-groups showed significantly larger temporo-parietal (particularly supramarginal gyrus) responses to target words repeatedly presented for training, than to control words. After the training, they also exhibited stronger inferior frontal responses to target words embedded in sentences. These findings suggest that word segmentation largely involves a cerebral circuit of phonological (phonetic) short-term memory. The dorsal pathway involved in encoding and decoding phonological representation may start to function stably at around 7months of age to facilitate the growth of the infant's vocabulary. [ABSTRACT FROM AUTHOR]

While American English infants typically segment words from fluent speech by 7.5-months, studies of infants from other language backgrounds have difficulty replicating this finding. One possible explanation for this cross-linguistic difference is that the input infants from different language backgrounds receive is not as infant-directed as American English infant-directed speech (Floccia et al., 2016). Against this background, the current study investigates whether German 7.5- and 9-month-old infants segment words from fluent speech when the input is prosodically similar to American English IDS. While 9-month-olds showed successful segmentation of words from exaggerated IDS, 7.5-month-olds did not. These findings highlight (a) the beneficial impact of exaggerated IDS on infant speech segmentation, (b) cross-linguistic differences in word segmentation that are based not just on the kind of input available to children and suggest (c) developmental differences in the role of IDS as an attentional spotlight in speech segmentation. [ABSTRACT FROM AUTHOR]

Over the first year, infants begin to learn the words of their language. Previous work suggests that certain statistical regularities in speech could help infants segment the speech stream into words, thereby forming a proto-lexicon that could support learning of the eventual vocabulary. However, computational models of word segmentation have typically been tested using language input that is much less variable than actual speech is. We show that using actual, transcribed pronunciations rather than dictionary pronunciations of the same speech leads to worse segmentation performance across models. We also find that phonologically variable input poses serious problems for lexicon building, because even correctly segmented word forms exhibit a complex, many-to-many relationship with speakers' intended words. Many phonologically distinct word forms were actually the same intended word, and many identical transcriptions came from different intended words. The fact that previous models appear to have substantially overestimated the utility of simple statistical heuristics suggests a need to consider the formation of the lexicon in infancy differently. [ABSTRACT FROM AUTHOR]

Before their first birthday, infants have started to identify and use information about their native language, such as frequent words, transitional probabilities, and co-occurrence of segments (phonotactics), to identify viable word boundaries. These cues can then be used to segment new words from running speech. We explored whether infants are capable of detecting a novel word form using the frequency of occurrence of the onset alone to further characterize the role of phonotactics in speech segmentation. Experiment 1 shows that English-learning 9-month-olds can successfully segment a word from natural speech if the onset is legal in English (i.e., pleet ) but not if the onset is illegal (i.e., tleet ). Experiment 2 shows that English-learning 9-month-olds are successful at word segmentation when presented with two onset clusters that vary in statistical frequency. Infants familiarized to a high-frequency onset (i.e., trom ) were successful at segmenting the target word embedded in speech, but those familiarized to the low-frequency onset (i.e., drom ) were unsuccessful. Together, these results show that infants use statistical information from the speech input and that low levels of exposure to onset phonotactics alone might not be sufficient in identifying word boundaries. [ABSTRACT FROM AUTHOR]

Human infants learn meanings for spoken words in complex interactions with other people, but the exact learning mechanisms are unknown. Among researchers, a widely studied learning mechanism is called cross-situational learning (XSL). In XSL, word meanings are learned when learners accumulate statistical information between spoken words and co-occurring objects or events, allowing the learner to overcome referential uncertainty after having sufficient experience with individually ambiguous scenarios. Existing models in this area have mainly assumed that the learner is capable of segmenting words from speech before grounding them to their referential meaning, while segmentation itself has been treated relatively independently of the meaning acquisition. In this article, we argue that XSL is not just a mechanism for word-to-meaning mapping, but that it provides strong cues for proto-lexical word segmentation. If a learner directly solves the correspondence problem between continuous speech input and the contextual referents being talked about, segmentation of the input into word-like units emerges as a by-product of the learning. We present a theoretical model for joint acquisition of proto-lexical segments and their meanings without assuming a priori knowledge of the language. We also investigate the behavior of the model using a computational implementation, making use of transition probability-based statistical learning. Results from simulations show that the model is not only capable of replicating behavioral data on word learning in artificial languages, but also shows effective learning of word segments and thenmeanings from continuous speech. Moreover, when augmented with a simple familiarity preference during learning, the model shows a good fit to human behavioral data in XSL tasks. These results support the idea of simultaneous segmentation and meaning acquisition and show that comprehensive models of early word segmentation should take referential word meanings into account. [ABSTRACT FROM AUTHOR]

Abstract: reported that when adult subjects are exposed to an unsegmented artificial language composed from trisyllabic words such as ABX, YBC, and AZC, they are unable to distinguish between these words and what they coined as the “phantom-word” ABC in a subsequent test. This suggests that statistical learning generates knowledge about the transitional probabilities (TPs) within each pair of syllables (AB, BC, and A⋯C), which are common to words and phantom-words, but, crucially, does not lead to the extraction of genuine word-like units. This conclusion is definitely inconsistent with chunk-based models of word segmentation, as confirmed by simulations run with the MDLChunker () and PARSER (), which successfully discover the words without computing TPs. Null results, however, can be due to multiple causes, and notably, in the case of Endress and Mehler, to the reduced level of intelligibility of their synthesized speech flow. In three experiments, we observed positive results in conditions similar to Endress and Mehler after only 5min of exposure to the language, hence providing strong evidence that statistical information is sufficient to extract word-like units. [Copyright &y& Elsevier]

Abstract: Infants have been described as ‘statistical learners’ capable of extracting structure (such as words) from patterned input (such as language). Here, we investigated whether prior knowledge influences how infants track transitional probabilities in word segmentation tasks. Are infants biased by prior experience when engaging in sequential statistical learning? In a laboratory simulation of learning across time, we exposed 9- and 10-month-old infants to a list of either disyllabic or trisyllabic nonsense words, followed by a pause-free speech stream composed of a different set of disyllabic or trisyllabic nonsense words. Listening times revealed successful segmentation of words from fluent speech only when words were uniformly disyllabic or trisyllabic throughout both phases of the experiment. Hearing trisyllabic words during the pre-exposure phase derailed infants’ abilities to segment speech into disyllabic words, and vice versa. We conclude that prior knowledge about word length equips infants with perceptual expectations that facilitate efficient processing of subsequent language input. [Copyright &y& Elsevier]

Statistical segmentation of continuous speech, i.e., the ability to utilise transitional probabilities between syllables in order to detect word boundaries, is reflected in the brain's auditory event-related potentials (ERPs). The N1 and N400 ERP components are typically enhanced for word onsets compared to random syllables during active listening. We used magnetoencephalography (MEG) to record event-related fields (ERFs) simultaneously with ERPs to syllables in a continuous sequence consisting of ten repeating tri-syllabic pseudowords and unexpected syllables presented between these pseudowords. We found the responses to differ between the syllables within the pseudowords and between the expected and unexpected syllables, reflecting an implicit process extracting the statistical characteristics of the sequence and monitoring for unexpected syllables. [ABSTRACT FROM AUTHOR]

This article is intended to investigate the question of sensitivity to language and its sound aspects, as related to writing acquisition processes in preschool children (4 to 6 years old). The aim is to study certain tasks that imply extracting and segmenting words and to relate this to the process of learning to write. Specifically, these tasks relate to rhyme (Task 1) and segmentation (Task 2) in the context of word games presented to preschool children. The results show that there is a connection between the extraction and segmentation process and the level of conceptualization of what is written. The better at extracting rhyme and segmenting syllables the children were, the higher their level of writing. Only a few responses with phonemic segmentation were found, even in literate children, since understanding of the system did not automatically give rise to exhaustive segmentations. [ABSTRACT FROM AUTHOR]

Abstract: To identify Song Ci style automatically, we put forward a novel stylistic text categorization approach based on words and their semantic in this paper. And a modified special word segmentation method, a new semantic relativity computing method based on HowNet along with the corresponding word sense disambiguation method are proposed to extract words and semantic features from Song Ci. Experiments are carried out and the results show that these methods are effective. [ABSTRACT FROM AUTHOR]

Traditional language models rely on lexical units that are defined as entities separated from each other by word boundary markers. Since there are no such boundaries in Thai, alternative definitions of lexical units have to be pursued. The problem is to find the optimal set of lexical units that constitutes the vocabulary of the language model and yields the best final result. The word is a traditional lexical unit recognized by Thai people and is used by most of the natural language processing systems, including an automatic speech recognition system. This paper discusses problems with using words as a lexical unit and investigates other lexical units for the Thai large vocabulary continuous speech recognition (LVCSR) system. The pseudo-morpheme is introduced in the paper and shown to be unsuitable for use as a lexical unit directly. A technique using pseudo-morphemes to improve the system based on the traditional word model is introduced and some improvements can be gained by this technique. Then, a new lexical unit for Thai, the compound pseudo-morpheme, and an algorithm to build compound pseudo-morphemes are presented. The experimental results show that the system using compound pseudo-morphemes outperforms other systems. Thus, the compound pseudo-morpheme is the most suitable lexical unit for Thai LVCSR system. [Copyright &y& Elsevier]

How can infants detect where words or morphemes start and end in the continuous stream of speech? Previous computational studies have investigated this question mainly for English, where morpheme and word boundaries are often isomorphic. Yet in many languages, words are often multimorphemic, such that word and morpheme boundaries do not align. Our study employed corpora of two languages that differ in the complexity of inflectional morphology, Chintang (Sino-Tibetan) and Japanese (in Experiment 1), as well as corpora of artificial languages ranging in morphological complexity, as measured by the ratio and distribution of morphemes per word (in Experiments 2 and 3). We used two baselines and three conceptually diverse word segmentation algorithms, two of which rely purely on sublexical information using distributional cues, and one that builds a lexicon. The algorithms' performance was evaluated on both word- and morpheme-level representations of the corpora. Segmentation results were better for the morphologically simpler languages than for the morphologically more complex languages, in line with the hypothesis that languages with greater inflectional complexity could be more difficult to segment into words. We further show that the effect of morphological complexity is relatively small, compared to that of algorithm and evaluation level. We therefore recommend that infant researchers look for signatures of the different segmentation algorithms and strategies, before looking for differences in infant segmentation landmarks across languages varying in complexity. [ABSTRACT FROM AUTHOR]

Abstract: The study examined intonational cues to word beginnings in French. French has an optional “early rise” in fundamental frequency (F 0) starting at the beginning of a content word. The role of this rise in segmentation by human listeners had previously been suggested, but not empirically tested. In Experiment 1, participants listened to noise-masked items like le ballon de mémentos/le ballon de mes manteaux, differing in segmentation and presence of an early rise. They interpreted early rises as markers of content word beginnings. In Experiment 2, the alignment of the early rise was manipulated in nonword sequences like [me.la.m.din]. Listeners were more likely to perceive two words (mes lamondines) when the early rise started at the second syllable and a single content (non)word (mélamondine) when it started at the first. Experiment 3 showed that a simple F 0 elbow at a function word–content word boundary also cued a content word beginning. This pattern and its potential use in word segmentation had not been previously reported in the literature. These intonational cues, like other cues to word segmentation, influenced rather than determined segmentation decisions. The influence of other cues (e.g., duration, word frequency) is also discussed. These results are the first evidence that French listeners use intonational information as cues to content word beginning. These cues are particularly important because the beginning of a word is a privileged position in word recognition and because, unlike many other cues (e.g., stress in English), they identify actual rather than potential word boundaries. The results provide support for an autosegmental-metrical account of the intonational phonology of French in which the early rise is a bitonal (LH) phrase accent that serves as a cue to content word beginnings. The cue is strongest when both tones (LH) are realized, which leads to an early rise, but can still be used if only the L is realized, which leads to a simple elbow. These results illustrate the importance of expanding studies of the range of cues to word segmentation to include intonational cues. [Copyright &y& Elsevier]

Infants come to learn several hundreds of word forms by two years of age, and it is possible this involves carving these forms out from continuous speech. It has been proposed that the task is facilitated by the presence of prosodic boundaries. We revisit this claim by running computational models of word segmentation, with and without prosodic information, on a corpus of infant-directed speech. We use five cognitively-based algorithms, which vary in whether they employ a sub-lexical or a lexical segmentation strategy and whether they are simple heuristics or embody an ideal learner. Results show that providing expert-annotated prosodic breaks does not uniformly help all segmentation models. The sub-lexical algorithms, which perform more poorly, benefit most, while the lexical ones show a very small gain. Moreover, when prosodic information is derived automatically from the acoustic cues infants are known to be sensitive to, errors in the detection of the boundaries lead to smaller positive effects, and even negative ones for some algorithms. This shows that even though infants could potentially use prosodic breaks, it does not necessarily follow that they should incorporate prosody into their segmentation strategies, when confronted with realistic signals. [ABSTRACT FROM AUTHOR]

Without inter-word spaces, Chinese readers need to use high-level information to segment words. The current study investigated how prior context influences word segmentation of overlapping ambiguous strings during Chinese reading. In two eye-tracking experiments, Chinese readers read sentences containing a 3-character overlapping ambiguous string (ABC) where both AB and BC were 2-character words. In the informative condition, prior contexts supported either the first word segmentation (AB-C) or the second word segmentation (A-BC), while didn't provide word-segmentation information in the neutral condition. The posttarget contexts were also manipulated to favor either the first word segmentation (AB-C) or the second word segmentation (A-BC). Results showed that when prior context was informative, first-pass reading times were shorter in the AB-C condition than that in the A-BC condition, while no significant difference was found in the neutral condition. However, the effect in the informative condition was eliminated when word frequency of the words BC was higher than that of the words AB. In the informative condition, the difference between AB-C and A-BC conditions was also significant for regression-in probability. These results imply that during word segmentation process, prior context exerts its effect at an early stage but inefficiently. Readers still utilize word frequency information in situation of informative context. And the first word segmentation (AB-C) has processing advantages, suggesting a left segmentation priority. [ABSTRACT FROM AUTHOR]

In Chinese, there are some incremental words, in which multi-character words (e.g., M, means landlady) contain a subset of characters that constitute another word (referred to as the embedded word, e.g., II, means boss). In the current talk, we will introduce some recent progress on experimental and modelling studies on how Chinese readers segment words during natural sentence reading. In one of these studies, Chinese readers read sentences with some incremental words while their eye movements were monitored. We manipulated the plausibility of incremental words and their embedded words. Results showed an interaction between the plausibility of the whole words and the plausibility of the embedded words on the reading times on the target word region. When the whole incremental word was plausible, the plausibility of the embedded words did not affect reading time on the target region. However, when the whole incremental word was implausible, reading times on the implausible embedded word was longer than plausible embedded word. These results suggest that Chinese readers tend to segment incremental word as a whole, but they can adjust their segmentation strategy very quickly once they found the segmentation is incongruent with sentence context. [ABSTRACT FROM AUTHOR]