Publications

Abstract

When we communicate with others, we often repeat aspects of each other's communicative behavior such as sentence structures and words. Such behavioral alignment has been mostly studied for speech or text. Yet, language use is mostly multimodal, flexibly using speech and gestures to convey messages. Here, we explore the use of alignment in speech (words) and co-speech gestures (iconic gestures) in a referential communication task aimed at finding labels for novel objects in interaction. In particular, we investigate how people flexibly use lexical and gestural alignment to create shared labels for novel objects and whether alignment in speech and gesture are related over time. The present study shows that interlocutors establish shared labels multimodally, and alignment in words and iconic gestures are used throughout the interaction. We also show that the amount of lexical alignment positively associates with the amount of gestural alignment over time, suggesting a close relationship between alignment in the vocal and manual modalities.

Abstract

Most work on part-of-speech (POS) tagging is focused on high resource languages, or examines low-resource and active learning settings through simulated studies. We evaluate POS tagging techniques on an actual endangered language, Griko. We present a resource that contains 114 narratives in Griko, along with sentence-level translations in Italian, and provides gold annotations for the test set. Based on a previously collected small corpus, we investigate several traditional methods, as well as methods that take advantage of monolingual data or project cross-lingual POS tags. We show that the combination of a semi-supervised method with cross-lingual transfer is more appropriate for this extremely challenging setting, with the best tagger achieving an accuracy of 72.9%. With an applied active learning scheme, which we use to collect sentence-level annotations over the test set, we achieve improvements of more than 21 percentage points

Abstract

Recognizing the limitations of standard vision assessments in capturing the real-world capabilities of individuals with low vision, we investigated the potential of the Seguin Form Board Test (SFBT), a widely-used intelligence assessment employing a visuo-haptic shape-fitting task, as an estimator of vision's practical utility. We present findings from 23 children from India, who underwent treatment for congenital bilateral dense cataracts, and 21 control participants. To assess the development of functional visual ability, we conducted the SFBT and the standard measure of visual acuity, before and longitudinally after treatment. We observed a dissociation in the development of shape-fitting and visual acuity. Improvements of patients' shape-fitting preceded enhancements in their visual acuity after surgery and emerged even with acuity worse than that of control participants. Our findings highlight the importance of incorporating multi-modal and cognitive aspects into evaluations of visual proficiency in low-vision conditions, to better reflect vision's impact on daily activities.

Abstract

Learning a language involves learning how to map specific forms onto their associated meanings. Such mappings can utilise arbitrariness and non-arbitrariness, yet, our understanding of how these two systems operate at different stages of vocabulary development is still not fully understood. The Sound-Symbolism Bootstrapping Hypothesis (SSBH) proposes that sound-symbolism is essential for word learning to commence, but empirical evidence of exactly how sound-symbolism influences language learning is still sparse. It may be the case that sound-symbolism supports acquisition of categories of meaning, or that it enables acquisition of individualized word meanings. In two Experiments where participants learned form-meaning mappings from either sound-symbolic or arbitrary languages, we demonstrate the changing roles of sound-symbolism and arbitrariness for different vocabulary sizes, showing that sound-symbolism provides an advantage for learning of broad categories, which may then transfer to support learning individual words, whereas an arbitrary language impedes acquisition of categories of sound to meaning.

Abstract

In natural communication, we typically both see and hear our conversation partner. Speech comprehension thus requires the integration of auditory and visual information from the speech signal. This is for instance evidenced by the Manual McGurk effect, where the perception of lexical stress is biased towards the syllable that has a beat gesture aligned to it. However, there is considerable individual variation in how heavily gestural timing is weighed as a cue to stress. To assess within-individualconsistency, this study investigated the test-retest reliability of the Manual McGurk effect. We reran an earlier Manual McGurk experiment with the same participants, over 1.5 years later. At the group level, we successfully replicated the Manual McGurk effect with a similar effect size. However, a correlation of the by-participant effect sizes in the two identical experiments indicated that there was only a weak correlation between both tests, suggesting that the weighing of gestural information in the perception of lexical stress is stable at the group level, but less so in individuals. Findings are discussed in comparison to other measures of audiovisual integration in speech perception. Index Terms: Audiovisual integration, beat gestures, lexical stress, test-retest reliability

Abstract

Speaker diarization (answering 'who spoke when') is a widely researched subject within speech technology. Numerous experiments have been run on datasets built from broadcast news, meeting data, and call centers—the task sometimes appears close to being solved. Much less work has begun to tackle the hardest diarization task of all: spontaneous conversations in real-world settings. Such diarization would be particularly useful for studies of language acquisition, where researchers investigate the speech children produce and hear in their daily lives. In this paper, we study audio gathered with a recorder worn by small children as they went about their normal days. As a result, each child was exposed to different acoustic environments with a multitude of background noises and a varying number of adults and peers. The inconsistency of speech and noise within and across samples poses a challenging task for speaker diarization systems, which we tackled via retraining and data augmentation techniques. We further studied sources of structured variation across raw audio files, including the impact of speaker type distribution, proportion of speech from children, and child age on diarization performance. We discuss the extent to which these findings might generalize to other samples of speech in the wild.

Abstract

In two experiments, Mandarin listeners resolved potential syntactic ambiguities in spoken utterances in (a) their native language (L1) and (b) English which they had learned as a second language (L2). A new disambiguation task was used, requiring speeded responses to select the correct meaning for structurally ambiguous sentences. Importantly, the ambiguities used in the study are identical in Mandarin and in English, and production data show that prosodic disambiguation of this type of ambiguity is also realised very similarly in the two languages. The perceptual results here showed however that listeners’ response patterns differed for L1 and L2, although there was a significant increase in similarity between the two response patterns with increasing exposure to the L2. Thus identical ambiguity and comparable disambiguation patterns in L1 and L2 do not lead to immediate application of the appropriate L1 listening strategy to L2; instead, it appears that such a strategy may have to be learned anew for the L2.

Abstract

A distinction between strong and weak vowels can be drawn on the basis of vowel quality, of stress, or of both factors. An experiment was conducted in which sets of contextually matched word-intial vowels ranging from clearly strong to clearly weak were cross-spliced, and the naturalness of the resulting words was rated by listeners. The ratings showed that in general cross-spliced words were only significantly less acceptable than unspliced words when schwa was not involved; this supports a categorical distinction based on vowel quality.

Abstract

Using a phoneme detection task, the present series of
experiments examines whether listeners can entrain to
different combinations of prosodic cues to predict where focus
will fall in an utterance. The stimuli were recorded by four
female native speakers of Australian English who happened to
have used different prosodic cues to produce sentences with
prosodic focus: a combination of duration cues, mean and
maximum F0, F0 range, and longer pre-target interval before
the focused word onset, only mean F0 cues, only pre-target
interval, and only duration cues. Results revealed that listeners
can entrain in almost every condition except for where
duration was the only reliable cue. Our findings suggest that
listeners are flexible in the cues they use for focus processing.

Abstract

Listeners adapt rapidly to previously unheard talkers by
adjusting phoneme categories using lexical knowledge, in a
process termed lexically-guided perceptual learning. Although
this is firmly established for listening in the native language
(L1), perceptual flexibility in second languages (L2) is as yet
less well understood. We report two experiments examining L1
and L2 perceptual learning, the first in Mandarin-English late
bilinguals, the second in Australian learners of Mandarin. Both
studies showed stronger learning in L1; in L2, however,
learning appeared for the English-L1 group but not for the
Mandarin-L1 group. Phonological mapping differences from
the L1 to the L2 are suggested as the reason for this result.

Abstract

Speakers and listeners have a shared goal: to communicate. The processes of speech perception and of speech production interact in many ways under the constraints of this communicative goal; such interaction is as characteristic of prosodic processing as of the processing of other aspects of linguistic structure. Two of the major uses of prosodic information in situations of communication are to encode salience and segmentation, and these themes unite the contributions to the symposium introduced by the present review.

Abstract

Biologically inspired spiking networks are an important tool to study the nature of computation and cognition in neural systems. In this work, we investigate the representational capacity of spiking networks engaged in an identity mapping task. We compare two schemes for encoding symbolic input, one in which input is injected as a direct current and one where input is delivered as a spatio-temporal spike pattern. We test the ability of networks to discriminate their input as a function of the number of distinct input symbols. We also compare performance using either membrane potentials or filtered spike trains as state variable. Furthermore, we investigate how the circuit behavior depends on the balance between excitation and inhibition, and the degree of synchrony and regularity in its internal dynamics. Finally, we compare different linear methods of decoding population activity onto desired target labels. Overall, our results suggest that even this simple mapping task is strongly influenced by design choices on input encoding, state-variables, circuit characteristics and decoding methods, and these factors can interact in complex ways. This work highlights the importance of constraining computational network models of behavior by available neurobiological evidence.

Abstract

We analyze the problem of response suggestion in a closed domain along a real-world scenario of a digital library. We present a text-processing pipeline to generate question-answer pairs from chat transcripts. On this limited amount of training data, we compare retrieval-based, conditioned-generation, and dedicated representation learning approaches for response suggestion. Our results show that retrieval-based methods that strive to find similar, known contexts are preferable over parametric approaches from the conditioned-generation family, when the training data is limited. We, however, identify a specific representation learning approach that is competitive to the retrieval-based approaches despite the training data limitation.

Abstract

We present multi-modal adversarial autoencoders for recommendation and evaluate them on two different tasks: citation recommendation and subject label recommendation. We analyze the effects of adversarial regularization, sparsity, and different input modalities. By conducting 408 experiments, we show that adversarial regularization consistently improves the performance of autoencoders for recommendation. We demonstrate, however, that the two tasks differ in the semantics of item co-occurrence in the sense that item co-occurrence resembles relatedness in case of citations, yet implies diversity in case of subject labels. Our results reveal that supplying the partial item set as input is only helpful, when item co-occurrence resembles relatedness. When facing a new recommendation task it is therefore crucial to consider the semantics of item co-occurrence for the choice of an appropriate model.

Abstract

Conversation requires a substantial amount of coordination between dialogue participants, from managing turn taking to negotiating mutual understanding. Part of this coordination effort surfaces as the reuse of linguistic behaviour across speakers, a process often referred to as alignment. While the presence of linguistic alignment is well documented in the literature, several questions remain open, including the extent to which patterns of reuse across speakers have an impact on the emergence of labelling conventions for novel referents. In this study, we put forward a methodology for automatically detecting shared lemmatised constructions---expressions with a common lexical core used by both speakers within a dialogue---and apply it to a referential communication corpus where participants aim to identify novel objects for which no established labels exist. Our analyses uncover the usage patterns of shared constructions in interaction and reveal that features such as their frequency and the amount of different constructions used for a referent are associated with the degree of object labelling convergence the participants exhibit after social interaction. More generally, the present study shows that automatically detected shared constructions offer a useful level of analysis to investigate the dynamics of reference negotiation in dialogue.

Abstract

Gestures are integral components of face-to-face communication. They unfold over time, often following predictable movement phases of preparation, stroke, and re-
traction. Yet, the prevalent approach to automatic gesture detection treats the problem as binary classification, classifying a segment as either containing a gesture or not, thus failing to capture its inherently sequential and contextual nature. To address this, we introduce a novel framework that reframes the task as a multi-phase sequence labeling problem rather than binary classification. Our model processes sequences of skeletal movements over time windows, uses Transformer encoders to learn contextual embeddings, and leverages Conditional Random Fields to perform sequence labeling. We evaluate our proposal on a large dataset of diverse co-speech gestures in task-oriented face-to-face dialogues. The results consistently demonstrate that our method significantly outperforms strong baseline models in detecting gesture strokes. Furthermore, applying Transformer encoders to learn contextual embeddings from movement sequences substantially improves gesture unit detection. These results highlight our framework’s capacity to capture the fine-grained dynamics of co-speech gesture phases, paving the way for more nuanced and accurate gesture detection and analysis.

Abstract

What makes some words harder to learn than others in a second language? Although some robust factors have been identified based on small scale experimental studies, many relevant factors are difficult to study in such experiments due to the amount of data necessary to test them. Here, we investigate what factors affect the ease of learning of a word in a second language using a large data set of users learning English as a second language through the Duolingo mobile app. In a regression analysis, we test and confirm the well-studied effect of cognate status on word learning accuracy. Furthermore, we find significant effects for both cross-linguistic semantic alignment and English semantic density, two novel predictors derived from large scale distributional models of lexical semantics. Finally, we provide data on several other psycholinguistically plausible word level predictors. We conclude with a discussion of the limits, benefits and future research potential of using big data for investigating second language learning.

Abstract

A common assumption in the cognitive sciences is that artificial and natural language learning rely on shared mechanisms. However, attempts to bridge the two have yielded ambiguous results. We suggest that an empirical disconnect between the computations employed during learning and the methods employed at test may explain these mixed results. Further, we propose statistically-based chunking as a potential computational link between artificial and natural language learning. We compare the acquisition of non-adjacent dependencies to that of natural language structure using two types of tasks: reflection-based 2AFC measures, and processing-based recall measures, the latter being more computationally analogous to the processes used during language acquisition. Our results demonstrate that task-type significantly influences the correlations observed between artificial and natural language acquisition, with reflection-based and processing-based measures correlating within – but not across – task-type. These findings have fundamental implications for artificial-to-natural language comparisons, both methodologically and theoretically.

Abstract

Indian Sign Language has limited resources for developing machine learning and data-driven approaches for automated language processing. Though text/audio-based language processing techniques have shown colossal research interest and tremendous improvements in the last few years, Sign Languages still need to catch up due to the need for more resources. To bridge this gap, in this work, we propose iSign: a benchmark for Indian Sign Language (ISL) Processing. We make three primary contributions to this work. First, we release one of the largest ISL-English datasets with more than video-sentence/phrase pairs. To the best of our knowledge, it is the largest sign language dataset available for ISL. Second, we propose multiple NLP-specific tasks (including SignVideo2Text, SignPose2Text, Text2Pose, Word Prediction, and Sign Semantics) and benchmark them with the baseline models for easier access to the research community. Third, we provide detailed insights into the proposed benchmarks with a few linguistic insights into the working of ISL. We streamline the evaluation of Sign Language processing, addressing the gaps in the NLP research community for Sign Languages. We release the dataset, tasks and models via the following website: https://exploration-lab.github.io/iSign/

Abstract

It has been suggested that some individuals may benefit more from social robots than do others. Using second
language (L2) as an example, the present study examined how individual differences in attitudes toward robots and personality
traits may be related to learning outcomes. Preliminary results with 24 Turkish-speaking adults suggest that negative attitudes
toward robots, more specifically thoughts and anxiety about the negative social impact that robots may have on the society,
predicted how well adults learned L2 words from a social robot. The possible implications of the findings as well as future directions are also discussed

Abstract

Human speakers often produce sentences incrementally. They can start speaking having in mind only a fragmentary idea of what they want to say, and while saying this they refine the contents underlying subsequent parts of the utterance. This capability imposes a number of constraints on the design of a syntactic processor. This paper explores these constraints and evaluates some recent computational sentence generators from the perspective of incremental production.

Abstract

Citations play a crucial role in the scientific discourse, in information retrieval, and in bibliometrics. Many initiatives are currently promoting the idea of having free and open citation data. Creation of citation data, however, is not part of the cataloging workflow in libraries nowadays.
In this paper, we present our project Linked Open Citation Database, in which we design distributed processes and a system infrastructure based on linked data technology. The goal is to show that efficiently cataloging citations in libraries using a semi-automatic approach is possible. We specifically describe the current state of the workflow and its implementation. We show that we could significantly improve the automatic reference extraction that is crucial for the subsequent data curation. We further give insights on the curation and linking process and provide evaluation results that not only direct the further development of the project, but also allow us to discuss its overall feasibility.

Abstract

It is widely held that smells and flavors are impossible to put into words. In this paper we test this claim by seeking predictive patterns in wine reviews, which ostensibly aim to provide guides to perceptual content. Wine reviews have previously been critiqued as random and meaningless. We collected an English corpus of wine reviews with their structured metadata, and applied machine learning techniques to automatically predict the wine's color, grape variety, and country of origin. To train the three supervised classifiers, three different information sources were incorporated: lexical bag-of-words features, domain-specific terminology features, and semantic word embedding features. In addition, using regression analysis we investigated basic review properties, i.e., review length, average word length, and their relationship to the scalar values of price and review score. Our results show that wine experts do share a common vocabulary to describe wines and they use this in a consistent way, which makes it possible to automatically predict wine characteristics based on the review text alone. This means that odors and flavors may be more expressible in language than typically acknowledged.

Abstract

The past year has seen a steep rise in generative AI systems that claim to be open. But how open are they really? The question of what counts as open source in generative AI is poised to take on particular importance in light of the upcoming EU AI Act that regulates open source systems differently, creating an urgent need for practical openness assessment. Here we use an evidence-based framework that distinguishes 14 dimensions of openness, from training datasets to scientific and technical documentation and from licensing to access methods. Surveying over 45 generative AI systems (both text and text-to-image), we find that while the term open source is widely used, many models are `open weight' at best and many providers seek to evade scientific, legal and regulatory scrutiny by withholding information on training and fine-tuning data. We argue that openness in generative AI is necessarily composite (consisting of multiple elements) and gradient (coming in degrees), and point out the risk of relying on single features like access or licensing to declare models open or not. Evidence-based openness assessment can help foster a generative AI landscape in which models can be effectively regulated, model providers can be held accountable, scientists can scrutinise generative AI, and end users can make informed decisions.

Abstract

Speakers and listeners show an informativity bias in the use and interpretation of color modifiers. For example, speakers use color more often when referring to objects that vary in color than to objects with a prototypical color. Likewise, listeners look away from objects with prototypical colors upon hearing that color mentioned. Here we test whether speakers and listeners account for another factor related to informativity: the strength of the association between lexical categories and color. Our results demonstrate that speakers and listeners' choices are indeed influenced by this factor; as such, it should be integrated into current pragmatic theories of informativity and computational models of color reference.

Abstract

We present the Narrative Brain Dataset, an fMRI dataset that was collected during spoken presentation of short excerpts of three
stories in Dutch. Together with the brain imaging data, the dataset contains the written versions of the stimulation texts. The texts are
accompanied with stochastic (perplexity and entropy) and semantic computational linguistic measures. The richness and unconstrained
nature of the data allows the study of language processing in the brain in a more naturalistic setting than is common for fMRI studies.
We hope that by making NBD available we serve the double purpose of providing useful neural data to researchers interested in natural
language processing in the brain and to further stimulate data sharing in the field of neuroscience of language.

Abstract

For (semi-)automated subject indexing systems in digital libraries, it is often more practical to use metadata such as the title of a publication instead of the full-text or the abstract. Therefore, it is desirable to have good text mining and text classification algorithms that operate well already on the title of a publication. So far, the classification performance on titles is not competitive with the performance on the full-texts if the same number of training samples is used for training. However, it is much easier to obtain title data in large quantities and to use it for training than full-text data. In this paper, we investigate the question how models obtained from training on increasing amounts of title training data compare to models from training on a constant number of full-texts. We evaluate this question on a large-scale dataset from the medical domain (PubMed) and from economics (EconBiz). In these datasets, the titles and annotations of millions of publications are available, and they outnumber the available full-texts by a factor of 20 and 15, respectively. To exploit these large amounts of data to their full potential, we develop three strong deep learning classifiers and evaluate their performance on the two datasets. The results are promising. On the EconBiz dataset, all three classifiers outperform their full-text counterparts by a large margin. The best title-based classifier outperforms the best full-text method by 9.4%. On the PubMed dataset, the best title-based method almost reaches the performance of the best full-text classifier, with a difference of only 2.9%.

Abstract

Human conversations are inherently multimodal, including auditory speech, visual articulatory cues, and hand gestures. Recent studies demonstrated that the timing of a simple up-and-down hand movement, known as a beat gesture, can affect speech perception. A beat gesture falling on the first syllable of a disyllabic word induces a bias to perceive a strong-weak stress pattern (i.e., “CONtent”), while a beat gesture falling on the second syllable combined with the same acoustics biases towards a weak-strong stress pattern (“conTENT”). This effect, termed the “manual McGurk effect”, has been studied in both in-lab and online studies, employing standard experimental sessions lasting approximately forty minutes. The present work tests whether the manual McGurk effect can be observed in an online short version (“mini-test”) of the original paradigm, lasting only ten minutes. Additionally, we employ two different response modalities, namely a two-alternative forced choice and a visual analog scale. A significant manual McGurk effect was observed with both response modalities. Overall, the present study demonstrates the feasibility of employing a ten-minute manual McGurk mini-test to obtain a measure of gesture-speech integration. As such, it may lend itself for inclusion in large-scale test batteries that aim to quantify individual variation in language processing.

Abstract

The ultimate goal of our research is to improve an existing speech-based computational model of human speech recognition on the task of simulating the role of fine-grained phonetic information in human speech processing. As part of this work we are investigating articulatory feature classifiers that are able to create reliable and accurate transcriptions of the articulatory behaviour encoded in the acoustic speech signal. Articulatory feature (AF) modelling of speech has received a considerable amount of attention in automatic speech recognition research. Different approaches have been used to build AF classifiers, most notably multi-layer perceptrons. Recently, deep neural networks have been applied to the task of AF classification. This paper aims to improve AF classification by investigating two different approaches: 1) investigating the usefulness of a deep Convolutional neural network (CNN) for AF classification; 2) integrating the Mel filtering operation into the CNN architecture. The results showed a remarkable improvement in classification accuracy of the CNNs over state-of-the-art AF classification results for Dutch, most notably in the minority classes. Integrating the Mel filtering operation into the CNN architecture did not further improve classification performance.

Abstract

With the recent advancements in the field of robotics and the increased focus on having general-purpose robots widely available to the general public, it has become increasingly necessary to pursue research into Human-robot interaction (HRI). While there have been a lot of works discussing frameworks for teaching HRI in educational institutions with a few institutions already offering courses to students, a consensus on the course content still eludes the field. In this work, we highlight a few challenges and opportunities while designing an HRI course from an Indian perspective. These topics warrant further deliberations as they have a direct impact on the design of HRI courses and wider implications for the entire field.

Abstract

Analyzing EEG signals recorded while participants are listening to continuous speech with the purpose of testing linguistic hypotheses is complicated by the fact that the signals simultaneously reflect exogenous acoustic excitation and endogenous linguistic processing. This makes it difficult to trace subtle differences that occur in mid-sentence position. We apply an analysis based on multivariate temporal response functions to uncover subtle mid-sentence effects. This approach is based on a per-stimulus estimate of the response of the neural system to speech input. Analyzing EEG signals predicted on the basis of the response functions might then bring to light conditionspecific differences in the filtered signals. We validate this approach by means of an analysis of EEG signals recorded with isolated word stimuli. Then, we apply the validated method to the analysis of the responses to the same words in the middle of meaningful sentences.

Abstract

Natural speech involves repair. These repairs are often highlighted through prosodic marking (Levelt & Cutler, 1983). Prosodic marking usually entails an increase in pitch, loudness, and/or duration that draws attention to the corrected word. While it is established that natural self-repairs typically elicit prosodic marking, the exact cause of this is unclear. This study investigates whether producing a prosodic marking emerges from an automatic correction process or has a communicative purpose. In the current study, we elicit corrections to test whether all self-corrections elicit prosodic marking. Participants carried out a picture-naming task in which they described two images presented on-screen. To prompt self-correction, the second image was altered in some cases, requiring participants to abandon their initial utterance and correct their description to match the new image. This manipulation was compared to a control condition in which only the orientation of the object would change, eliciting no self-correction while still presenting a visual change. We found that the replacement of the item did not elicit a prosodic marking, regardless of the type of change. Theoretical implications and research directions are discussed, in particular theories of prosodic planning.

Abstract

Word count estimation (WCE) from audio recordings has a number of applications, including quantifying the amount of speech that language-learning infants hear in their natural environments, as captured by daylong recordings made with devices worn by infants. To be applicable in a wide range of scenarios and also low-resource domains, WCE tools should be extremely robust against varying signal conditions and require minimal access to labeled training data in the target domain. For this purpose, earlier work has used automatic syllabification of speech, followed by a least-squares-mapping of syllables to word counts. This paper compares a number of previously proposed syllabifiers in the WCE task, including a supervised bi-directional long short-term memory (BLSTM) network that is trained on a language for which high quality syllable annotations are available (a “high resource language”), and reports how the alternative methods compare on different languages and signal conditions. We also explore additive noise and varying-channel data augmentation strategies for BLSTM training, and show how they improve performance in both matching and mismatching languages. Intriguingly, we also find that even though the BLSTM works on languages beyond its training data, the unsupervised algorithms can still outperform it in challenging signal conditions on novel languages.

Abstract

It has been shown that when speakers produce hand gestures, addressees are attentive towards these gestures, using them to facilitate speech processing. Even relatively simple “beat” gestures are taken into account to help process aspects of speech such as prosodic prominence. In fact, recent evidence suggests that the timing of a beat gesture can influence spoken word recognition. Termed the manual McGurk Effect, Dutch participants, when presented with lexical stress minimal pair continua in Dutch, were biased to hear lexical stress on the syllable that coincided with a beat gesture. However, little is known about how this manual McGurk effect would surface in languages other than Dutch, with different acoustic cues to prominence, and variable gestures. Therefore, this study tests the effect in Spanish where lexical stress is arguably even more important, being a contrastive cue in the regular verb conjugation system. Results from 24 participants corroborate the effect in Spanish, namely that when given the same auditory stimulus, participants were biased to perceive lexical stress on the syllable that visually co-occurred with a beat gesture. These findings extend the manual McGurk effect to a different language, emphasizing the impact of gestures' timing on prosody perception and spoken word recognition.

Abstract

Recent research on multimodal language production has revealed that prominence in speech and gesture go hand-in-hand. Specifically, peaks in gesture (i.e., the apex) seem to closely coordinate with peaks in fundamental frequency (F0). The nature of this relationship may also be bi-directional, as it has also been shown that the production of gesture directly affects speech acoustics. However, most studies on the topic have largely focused on stress-based languages, where fundamental frequency has a prominence-lending function. Less work has been carried out on lexical tone languages such as Mandarin, where F0 is lexically distinctive. In this study, four native Mandarin speakers were asked to produce single monosyllabic CV words, taken from minimal lexical tone triplets (e.g., /pi1/, /pi2/, /pi3/), either with or without a beat gesture. Our analyses of the timing of the gestures showed that the gesture apex most stably occurred near vowel onset, with consonantal duration being the strongest predictor of apex placement. Acoustic analyses revealed that words produced with gesture showed raised F0 contours, greater intensity, and shorter durations. These findings further our understanding of gesture-speech alignment in typologically diverse languages, and add to the discussion about multimodal prominence.

Abstract

For almost two decades, the poor performance observed with the so-called Director task has been interpreted as evidence of limited use of Theory of Mind in communication. Here we propose a probabilistic model of common ground in referential communication that derives three inferences from an utterance: what the speaker is talking about in a visual context, what she knows about the context, and what referential expressions she prefers. We tested our model by comparing its inferences with those made by human participants and found that it closely mirrors their judgments, whereas an alternative model compromising the hearer’s expectations of cooperativeness and efficiency reveals a worse fit to the human data. Rather than assuming that common ground is fixed in a given exchange and may or may not constrain reference resolution, we show how common ground can be inferred as part of the process of reference assignment.

Abstract

We investigated the relationship between visual experience and pragmatic development by testing the socio-communicative skills of a unique population: the Prakash children of India, who received treatment for congenital cataracts after years of visual deprivation. Using two different referential communication tasks, our study investigated Prakash' children ability to produce sufficiently informative referential expressions (e.g., ‘the green pear' or ‘the small plate') and pay attention to their interlocutor's face during the task (Experiment 1), as well as their ability to recognize a speaker's referential intent through non-verbal cues such as head turning and pointing (Experiment 2). Our results show that Prakash children have strong pragmatic skills, but do not look at their interlocutor's face as often as neurotypical children do. However, longitudinal analyses revealed an increase in face fixations, suggesting that over time, Prakash children come to utilize their improved visual skills for efficient referential communication.

Abstract

While there are many studies on information retrieval models using full-text, there are presently no comparison studies of full-text retrieval vs. retrieval only over the titles of documents. On the one hand, the full-text of documents like scientific papers is not always available due to, e.g., copyright policies of academic publishers. On the other hand, conducting a search based on titles alone has strong limitations. Titles are short and therefore may not contain enough information to yield satisfactory search results. In this paper, we compare different retrieval models regarding their search performance on the full-text vs. only titles of documents. We use different datasets, including the three digital library datasets: EconBiz, IREON, and PubMed. The results show that it is possible to build effective title-based retrieval models that provide competitive results comparable to full-text retrieval. The difference between the average evaluation results of the best title-based retrieval models is only 3% less than those of the best full-text-based retrieval models.

Abstract

Despite decades of study, we still know less than we would like about the association between joint attention (JA) and language acquisition. This is partly because of disagreements on how to operationalise JA. In this study, we examine the impact of applying two different, influential JA operationalisation schemes to the same dataset of child-caregiver interactions, to determine which yields a better fit to children's later vocabulary size. Two coding schemes— one defining JA in terms of gaze overlap and one in terms of social aspects of shared attention—were applied to video-recordings of dyadic naturalistic toy-play interactions (N=45). We found that JA was predictive of later production vocabulary when operationalised as shared focus (study 1), but also that its operationalisation as shared social awareness increased its predictive power (study 2). Our results emphasise the critical role of methodological choices in understanding how and why JA is associated with vocabulary size.

Abstract

Fine-Tracker is a speech-based model of human speech
recognition. While previous work has shown that Fine-Tracker
is successful at modelling aspects of human spoken-word
recognition, its speech recognition performance is not
comparable to that of human performance, possibly due to
suboptimal intermediate articulatory feature (AF)
representations. This study investigates the effect of improved
AF representations, obtained using a state-of-the-art deep
convolutional network, on Fine-Tracker’s simulation and
recognition performance: Although the improved AF quality
resulted in improved speech recognition; it, surprisingly, did
not lead to an improvement in Fine-Tracker’s simulation power.

Abstract

We report an individual with music-odor synaesthesia who experiences automatic and vivid odor sensations when she hears music. S’s odor associations were recorded on two days, and compared with those of two control participants. Overall, S produced longer descriptions, and her associations were of multiple odors at once, in comparison to controls who typically reported a single odor. Although odor associations were qualitatively different between S and controls, ratings of the consistency of their descriptions did not differ. This demonstrates that crossmodal associations between music and odor exist in non-synaesthetes too. We also found that S is better at discriminating between odors than control participants, and is more likely to experience emotion, memories and evaluations triggered by odors, demonstrating the broader impact of her synaesthesia.

Abstract

Sequences of reaction times (RT) produced by participants in an experiment are not only influenced by the stimuli, but by many other factors as well, including fatigue, attention, experience, IQ, handedness, etc. These confounding factors result in longterm effects (such as a participant’s overall reaction capability) and in short- and medium-time fluctuations in RTs (often referred to as ‘local speed effects’). Because stimuli are usually presented in a random sequence different for each participant, local speed effects affect the underlying ‘true’ RTs of specific trials in different ways across participants. To be able to focus statistical analysis on the effects of the cognitive process under study, it is necessary to reduce the effect of confounding factors as much as possible. In this paper we propose and compare techniques and criteria for doing so, with focus on reducing (‘filtering’) the local speed effects. We show that filtering matters substantially for the significance analyses of predictors in linear mixed effect regression models. The performance of filtering is assessed by the average between-participant correlation between filtered RT sequences and by Akaike’s Information Criterion, an important measure of the goodness-of-fit of linear mixed effect regression models.

Abstract

The recent advent of deep learning techniques in speech tech-nology and in particular in automatic speech recognition hasyielded substantial performance improvements. This suggeststhat deep neural networks (DNNs) are able to capture structurein speech data that older methods for acoustic modeling, suchas Gaussian Mixture Models and shallow neural networks failto uncover. In image recognition it is possible to link repre-sentations on the first couple of layers in DNNs to structuralproperties of images, and to representations on early layers inthe visual cortex. This raises the question whether it is possi-ble to accomplish a similar feat with representations on DNNlayers when processing speech input. In this paper we presentthree different experiments in which we attempt to untanglehow DNNs encode speech signals, and to relate these repre-sentations to phonetic knowledge, with the aim to advance con-ventional phonetic concepts and to choose the topology of aDNNs more efficiently. Two experiments investigate represen-tations formed by auto-encoders. A third experiment investi-gates representations on convolutional layers that treat speechspectrograms as if they were images. The results lay the basisfor future experiments with recursive networks.

Abstract

We estimate lexical Concreteness for millions of words across 77 languages. Using a simple regression framework, we combine vector-based models of lexical semantics with experimental norms of Concreteness in English and Dutch. By applying techniques to align vector-based semantics across distinct languages, we compute and release Concreteness estimates at scale in numerous languages for which experimental norms are not currently available. This paper lays out the technique and its efficacy. Although this is a difficult dataset to evaluate immediately, Concreteness estimates computed from English correlate with Dutch experimental norms at $\rho$ = .75 in the vocabulary at large, increasing to $\rho$ = .8 among Nouns. Our predictions also recapitulate attested relationships with word frequency. The approach we describe can be readily applied to numerous lexical measures beyond Concreteness

Abstract

Do all languages convey semantic knowledge in the same way? If language simply mirrors the structure of the world, the answer should be a qualified “yes”. If, however, languages impose structure as much as reflecting it, then even ostensibly the “same” word in different languages may mean quite different things. We provide a first pass at a large-scale quantification of cross-linguistic semantic alignment of approximately 1000 meanings in 55 languages. We find that the translation equivalents in some domains (e.g., Time, Quantity, and Kinship) exhibit high alignment across languages while the structure of other domains (e.g., Politics, Food, Emotions, and Animals) exhibits substantial cross-linguistic variability. Our measure of semantic alignment correlates with known phylogenetic distances between languages: more phylogenetically distant languages have less semantic alignment. We also find semantic alignment to correlate with cultural distances between societies speaking the languages, suggesting a rich co-adaptation of language and culture even in domains of experience that appear most constrained by the natural world

Abstract

In situated communication, reference to an entity in the shared visual context can be established using eitheranexpression that conveys precise (minimally specified) or redundant (over-specified) information. There is, however, along-lasting debate in psycholinguistics concerningwhether the latter hinders referential processing. We present evidence from an eyetrackingexperiment recordingfixations as well asthe Index of Cognitive Activity –a novel measure of cognitive workload –supporting the view that over-specifications facilitate processing. We further present originalevidence that, above and beyond the effect of specificity,referring expressions thatuniformly reduce referential entropyalso benefitprocessing

Abstract

The human brain deals with the infinite variability of speech through multiple mechanisms. Some of them rely solely on information in the speech input (i.e., signal-driven) whereas some rely on linguistic or real-world knowledge (i.e., knowledge-driven). Many signal-driven perceptual processes rely on the enhancement of acoustic differences between incoming speech sounds, producing contrastive adjustments. For instance, when an ambiguous voiceless fricative is preceded by a high fundamental frequency (f0) sentence, the fricative is perceived as having lower a spectral center of gravity (CoG). However, it is not clear whether knowledge of a talker’s typical f0 can lead to similar contrastive effects. This study investigated a possible talker f0 effect on fricative CoG perception. In the exposure phase, two groups of participants (N=16 each) heard the same talker at high or low f0 for 20 minutes. Later, in the test phase, participants rated fixed-f0 /?ɔk/ tokens as being /sɔk/ (i.e., high CoG) or /ʃɔk/ (i.e., low CoG), where /?/ represents a fricative from a 5-step /s/-/ʃ/ continuum. Surprisingly, the data revealed the opposite of our contrastive hypothesis, whereby hearing high f0 instead biased perception towards high CoG. Thus, we demonstrated that talker f0 information affects fricative CoG perception.

Abstract

The task of automatic playlist continuation is generating a list of recommended tracks that can be added to an existing playlist. By suggesting appropriate tracks, i. e., songs to add to a playlist, a recommender system can increase the user engagement by making playlist creation easier, as well as extending listening beyond the end of current playlist. The ACM Recommender Systems Challenge 2018 focuses on such task. Spotify released a dataset of playlists, which includes a large number of playlists and associated track listings. Given a set of playlists from which a number of tracks have been withheld, the goal is predicting the missing tracks in those playlists. We participated in the challenge as the team Unconscious Bias and, in this paper, we present our approach. We extend adversarial autoencoders to the problem of automatic playlist continuation. We show how multiple input modalities, such as the playlist titles as well as track titles, artists and albums, can be incorporated in the playlist continuation task.

Abstract

This paper reports two experiments with vowels and consonants as phoneme detection targets in real words. In the first experiment, two relatively distinct vowels were compared with two confusible stop consonants. Response times to the vowels were longer than to the consonants. Response times correlated negatively with target phoneme length. In the second, two relatively distinct vowels were compared with their corresponding semivowels. This time, the vowels were detected faster than the semivowels. We conclude that response time differences between vowels and stop consonants in this task may reflect differences between phoneme categories in the variability of tokens, both in the acoustic realisation of targets and in the' representation of targets by subjects.

Abstract

Contrastive pitch accents can mark sentence elements occupying parallel roles. In “Mary kissed John, not Peter”, a pitch accent on Mary or John cues the implied syntactic role of Peter. Van der Burght, Friederici, Goucha, and Hartwigsen (2021) showed that listeners can build expectations concerning syntactic and semantic properties of upcoming words, derived from pitch accent information they heard previously. To further explore these expectations, we attempted a partial replication of the original German study in Dutch. In the experimental sentences “Yesterday, the police officer arrested the thief, not the inspector/murderer”, a pitch accent on subject or object cued the subject/object role of the ellipsis clause. Contrasting elements were additionally cued by the thematic role typicality of the nouns. Participants listened to sentences in which the ellipsis clause was omitted and selected the most plausible sentence-final noun (presented visually) via button press. Replicating the original study results, listeners based their sentence-final preference on the pitch accent information available in the sentence. However, as in the original study, individual differences between listeners were found, with some following prosodic information and others relying on a structural bias. The results complement the literature on ellipsis resolution and on interindividual variability in cue weighting.

Abstract

Before infants become mature speakers of their native language, they must acquire a robust word-recognition system which allows them to strike the balance between allowing some variation (mood, voice, accent) and recognizing variability that potentially changes meaning (e.g. cat vs hat). The current meta-analysis quantifies how the latter, termed mispronunciation sensitivity, changes over infants’ first three years, testing competing predictions of mainstream language acquisition theories. Our results show that infants were sensitive to mispronunciations, but accepted them as labels for target objects. Interestingly, and in contrast to predictions of mainstream theories, mispronunciation sensitivity was not modulated by infant age, suggesting that a sufficiently flexible understanding of native language phonology is in place at a young age.

Abstract

Early word learning involves mapping individual words to their meanings and building organized semantic representations among words. Previous corpus-based studies (e.g., using text from websites, newspapers, child-directed speech corpora) demonstrated that linguistic information such as word co-occurrence alone is sufficient to build semantically organized word knowledge. The present study explored two new research directions to advance understanding of how infants acquire semantically organized word knowledge. First, infants in the real world hear words surrounded by contextual information. Going beyond inferring semantic knowledge merely from language input, we examined the role of extra-linguistic contextual information in learning semantic knowledge. Second, previous research relies on large amounts of linguistic data to demonstrate in-principle learning, which is unrealistic compared with the input children receive. Here, we showed that incorporating extra-linguistic information provides an efficient mechanism through which semantic knowledge can be acquired with a small amount of data infants perceive in everyday learning contexts, such as toy play.

Abstract

Prosodic prominence can enhance memory for the prominent words. This mnemonic benefit has been linked to listeners’ allocation of attention and deeper processing, which leads to more robust semantic representations. We investigated whether, in addition to the well-established effect at the semantic level, there was a memory benefit for prominent words at the phonological level. To do so, participants (48 native speakers of Dutch), first performed an accent judgement task, where they had to discriminate accented from unaccented words, and accented from unaccented pseudowords. All stimuli were presented in lists. They then performed an old/new recognition task for the stimuli. Accuracy in the accent judgement task was equally high for words and pseudowords. In the recognition task, performance was, as expected, better for words than pseudowords. More importantly, there was an interaction of accent with word type, with a significant advantage for accented compared to unaccented words, but not for pseudowords. The results confirm the memory benefit for accented compared to unaccented words seen in earlier studies, and they are consistent with the view that prominence primarily affects the semantic encoding of words. There was no evidence for an additional memory benefit arising at the phonological level.

Abstract

In Swedish, words are associated with either of two pitch contours known as Accent 1 and Accent 2. Using a psychometric test, we investigated how listeners judge pitch accent violations while interpreting discourse. Forty native speakers of Central Swedish were presented with auditory dialogues, where test words were appropriately or inappropriately accented in a given context, and asked to judge the correctness of sentences containing the test words. Data indicated a statistically significant effect of wrong accent pattern on the correctness judgment. Both Accent 1 and Accent 2 violations interfered with the coherent interpretation of discourse and were judged as incorrect by the listeners. Moreover, there was a statistically significant difference in the perceived correctness between the accent patterns. Accent 2 violations led to a lower correctness score compared to Accent 1 violations, indicating that the listeners were more sensitive to pitch accent violations in Accent 2 words than in Accent 1 words. This result is in line with the notion that Accent 2 is marked and lexically represented in Central Swedish. Taken together, these findings indicate that listeners use both Accent 1 and Accent 2 to arrive at the correct interpretation of the linguistic input, while assigning varying degrees of relevance to them depending on their markedness.