scholarly journals On the Difficulty of Translating Free-Order Case-Marking Languages

2021 ◽  
Vol 9 ◽  
pp. 1233-1248
Author(s):  
Arianna Bisazza ◽  
Ahmet Üstün ◽  
Stephan Sportel
Keyword(s):  
Language Processing ◽  
Processing Technologies ◽  
Neural Machine Translation ◽  
Semantic Cues ◽  
Verb Agreement ◽  
Case Marking ◽  
Order Case ◽  
Fixed Order ◽  
Free Order

Abstract Identifying factors that make certain languages harder to model than others is essential to reach language equality in future Natural Language Processing technologies. Free-order case-marking languages, such as Russian, Latin, or Tamil, have proved more challenging than fixed-order languages for the tasks of syntactic parsing and subject-verb agreement prediction. In this work, we investigate whether this class of languages is also more difficult to translate by state-of-the-art Neural Machine Translation (NMT) models. Using a variety of synthetic languages and a newly introduced translation challenge set, we find that word order flexibility in the source language only leads to a very small loss of NMT quality, even though the core verb arguments become impossible to disambiguate in sentences without semantic cues. The latter issue is indeed solved by the addition of case marking. However, in medium- and low-resource settings, the overall NMT quality of fixed-order languages remains unmatched.

scholarly journals Predictive processes during simultaneous interpreting from German into English

Interpreting ◽  
2017 ◽  
Vol 19 (1) ◽  
pp. 1-20 ◽  
Author(s):  
Ena Hodzik ◽  
John N. Williams
Keyword(s):  
Language Processing ◽  
English Language ◽  
Native Speakers ◽  
Transitional Probability ◽  
Semantic Cues ◽  
Simultaneous Interpreting ◽  
Spoken Language Processing ◽  
Advanced Students ◽  
Native Speakers Of English ◽  
Predictive Processes

We report a study on prediction in shadowing and simultaneous interpreting (SI), both considered as forms of real-time, ‘online’ spoken language processing. The study comprised two experiments, focusing on: (i) shadowing of German head-final sentences by 20 advanced students of German, all native speakers of English; (ii) SI of the same sentences into English head-initial sentences by 22 advanced students of German, again native English speakers, and also by 11 trainee and practising interpreters. Latency times for input and production of the target verbs were measured. Drawing on studies of prediction in English-language reading production, we examined two cues to prediction in both experiments: contextual constraints (semantic cues in the context) and transitional probability (the statistical likelihood of words occurring together in the language concerned). While context affected prediction during both shadowing and SI, transitional probability appeared to favour prediction during shadowing but not during SI. This suggests that the two cues operate on different levels of language processing in SI.

What's my App?

2021 ◽  
Vol 48 (4) ◽  
pp. 41-44
Author(s):  
Dena Markudova ◽  
Martino Trevisan ◽  
Paolo Garza ◽  
Michela Meo ◽  
Maurizio M. Munafo ◽  
...  
Keyword(s):  
Machine Learning ◽  
Natural Language Processing ◽  
Real Time ◽  
Language Processing ◽  
Traffic Management ◽  
Quality Of Experience ◽  
Broadband Internet ◽  
Management Policies ◽  
Control Traffic

With the spread of broadband Internet, Real-Time Communication (RTC) platforms have become increasingly popular and have transformed the way people communicate. Thus, it is fundamental that the network adopts traffic management policies that ensure appropriate Quality of Experience to users of RTC applications. A key step for this is the identification of the applications behind RTC traffic, which in turn allows to allocate adequate resources and make decisions based on the specific application's requirements. In this paper, we introduce a machine learning-based system for identifying the traffic of RTC applications. It builds on the domains contacted before starting a call and leverages techniques from Natural Language Processing (NLP) to build meaningful features. Our system works in real-time and is robust to the peculiarities of the RTP implementations of different applications, since it uses only control traffic. Experimental results show that our approach classifies 5 well-known meeting applications with an F1 score of 0.89.

scholarly journals Joint Multimodal Embedding and Backtracking Search in Vision-and-Language Navigation

Sensors ◽  
2021 ◽  
Vol 21 (3) ◽  
pp. 1012
Author(s):  
Jisu Hwang ◽  
Incheol Kim
Keyword(s):  
Natural Language ◽  
Language Processing ◽  
Input Data ◽  
Language Instruction ◽  
Scoring Method ◽  
Processing Technologies ◽  
Backtracking Search ◽  
Panoramic Images ◽  
Benchmark Datasets ◽  
Vision And Language

Due to the development of computer vision and natural language processing technologies in recent years, there has been a growing interest in multimodal intelligent tasks that require the ability to concurrently understand various forms of input data such as images and text. Vision-and-language navigation (VLN) require the alignment and grounding of multimodal input data to enable real-time perception of the task status on panoramic images and natural language instruction. This study proposes a novel deep neural network model (JMEBS), with joint multimodal embedding and backtracking search for VLN tasks. The proposed JMEBS model uses a transformer-based joint multimodal embedding module. JMEBS uses both multimodal context and temporal context. It also employs backtracking-enabled greedy local search (BGLS), a novel algorithm with a backtracking feature designed to improve the task success rate and optimize the navigation path, based on the local and global scores related to candidate actions. A novel global scoring method is also used for performance improvement by comparing the partial trajectories searched thus far with a plurality of natural language instructions. The performance of the proposed model on various operations was then experimentally demonstrated and compared with other models using the Matterport3D Simulator and room-to-room (R2R) benchmark datasets.

scholarly journals Machine Learning for Dissimulating Reality

Proceedings ◽  
2021 ◽  
Vol 77 (1) ◽  
pp. 17
Author(s):  
Andrea Giussani
Keyword(s):  
Machine Learning ◽  
Language Processing ◽  
Huge Amount ◽  
Generative Adversarial Network ◽  
Adversarial Network ◽  
Technological Advances ◽  
Textual Data ◽  
Musical Scores ◽  
Mathematical Formulas

In the last decade, advances in statistical modeling and computer science have boosted the production of machine-produced contents in different fields: from language to image generation, the quality of the generated outputs is remarkably high, sometimes better than those produced by a human being. Modern technological advances such as OpenAI’s GPT-2 (and recently GPT-3) permit automated systems to dramatically alter reality with synthetic outputs so that humans are not able to distinguish the real copy from its counteracts. An example is given by an article entirely written by GPT-2, but many other examples exist. In the field of computer vision, Nvidia’s Generative Adversarial Network, commonly known as StyleGAN (Karras et al. 2018), has become the de facto reference point for the production of a huge amount of fake human face portraits; additionally, recent algorithms were developed to create both musical scores and mathematical formulas. This presentation aims to stimulate participants on the state-of-the-art results in this field: we will cover both GANs and language modeling with recent applications. The novelty here is that we apply a transformer-based machine learning technique, namely RoBerta (Liu et al. 2019), to the detection of human-produced versus machine-produced text concerning fake news detection. RoBerta is a recent algorithm that is based on the well-known Bidirectional Encoder Representations from Transformers algorithm, known as BERT (Devlin et al. 2018); this is a bi-directional transformer used for natural language processing developed by Google and pre-trained over a huge amount of unlabeled textual data to learn embeddings. We will then use these representations as an input of our classifier to detect real vs. machine-produced text. The application is demonstrated in the presentation.

scholarly journals Honorifics: A sociocultural verb agreement cue in Japanese sentence processing

2010 ◽  
Vol 31 (3) ◽  
pp. 551-569 ◽  
Author(s):  
YUKI YOSHIMURA ◽  
BRIAN MACWHINNEY
Keyword(s):  
Second Language ◽  
Language Learners ◽  
Sentence Processing ◽  
Word Order ◽  
Second Language Learners ◽  
Native Speakers ◽  
Role Assignment ◽  
Verb Agreement ◽  
Case Marking ◽  
New Type

ABSTRACTCase marking is the major cue to sentence interpretation in Japanese, whereas animacy and word order are much weaker. However, when subjects and their cases markers are omitted, Japanese honorific and humble verbs can provide information that compensates for the missing case role markers. This study examined the usage of honorific and humble verbs as cues to case role assignment by Japanese native speakers and second-language learners of Japanese. The results for native speakers replicated earlier findings regarding the predominant strength of case marking. However, when case marking was missing, native speakers relied more on honorific marking than word order. In these sentences, the processing that relied on the honorific cue was delayed by about 100 ms in comparison to processing that relied on the case-marking cue. Learners made extensive use of the honorific agreement cue, but their use of the cue was much less accurate than that of native speakers. In particular, they failed to systematically invoke the agreement cue when case marking was missing. Overall, the findings support the predictions of the model and extend its coverage to a new type of culturally determined cue.

Children With Cochlear Implants Use Semantic Prediction to Facilitate Spoken Word Recognition

2021 ◽  
pp. 1-14
Author(s):  
Christina Blomquist ◽  
Rochelle S. Newman ◽  
Yi Ting Huang ◽  
Jan Edwards
Keyword(s):  
Word Recognition ◽  
Target Word ◽  
Cochlear Implants ◽  
Language Processing ◽  
Spoken Word Recognition ◽  
Spoken Language ◽  
Spoken Word ◽  
Semantic Cues ◽  
Matched Group ◽  
Spoken Language Processing

Purpose Children with cochlear implants (CIs) are more likely to struggle with spoken language than their age-matched peers with normal hearing (NH), and new language processing literature suggests that these challenges may be linked to delays in spoken word recognition. The purpose of this study was to investigate whether children with CIs use language knowledge via semantic prediction to facilitate recognition of upcoming words and help compensate for uncertainties in the acoustic signal. Method Five- to 10-year-old children with CIs heard sentences with an informative verb ( draws ) or a neutral verb ( gets ) preceding a target word ( picture ). The target referent was presented on a screen, along with a phonologically similar competitor ( pickle ). Children's eye gaze was recorded to quantify efficiency of access of the target word and suppression of phonological competition. Performance was compared to both an age-matched group and vocabulary-matched group of children with NH. Results Children with CIs, like their peers with NH, demonstrated use of informative verbs to look more quickly to the target word and look less to the phonological competitor. However, children with CIs demonstrated less efficient use of semantic cues relative to their peers with NH, even when matched for vocabulary ability. Conclusions Children with CIs use semantic prediction to facilitate spoken word recognition but do so to a lesser extent than children with NH. Children with CIs experience challenges in predictive spoken language processing above and beyond limitations from delayed vocabulary development. Children with CIs with better vocabulary ability demonstrate more efficient use of lexical-semantic cues. Clinical interventions focusing on building knowledge of words and their associations may support efficiency of spoken language processing for children with CIs. Supplemental Material https://doi.org/10.23641/asha.14417627

Compound feed from different producers for piglets and their comparative assessment

2021 ◽  
pp. 3-12
Author(s):  
N. I. Kulmakova ◽  
T. A. Magomadov ◽  
N. M. Kostomakhin ◽  
M. N. Dmitrieva ◽  
H. Saleh
Keyword(s):  
Raw Materials ◽  
Comparative Assessment ◽  
Feed Additives ◽  
The Body ◽  
Animal Origin ◽  
Quality And Safety ◽  
Processing Technologies ◽  
Starting Compound ◽  
Federal Standard

The quality of raw materials and goods of animal origin depends first of all on the safety of feed, their balance in terms of the main nutrients and digestibility by the body. As a rule, the lower the quality and safety of feed, the lower the productivity of all types of animals and poultry. The quality of feed is influenced by all stages of their production: growing conditions, harvesting and storage, processing technologies, canning and preparation for feeding. Where high-quality feed is used in diets, maximum productivity and high realization of the genetic potential of animals are obtained. During the storage and processing of plant and animal raw materials its physic and mechanical, biochemical, sanitary and microbiological properties change. This can be avoided only by observing the sanitary and hygienic requirements for bagging, harvesting, and storing feed. The quality of feed is confirmed by its physical, chemical, organoleptic, microbiological and other indicators, which determines the variety of feed control methods at all stages of their turnover. In this connection, the development and strengthening of control over the quality and safety of feed and feed additives is one of the important tasks of modern animal feeding science. The purpose of the work was to carried out a comparative assessment of four samples of the starting compound feed SK-3 for piglets of different producers. The experimental part of the work has been carried out in the laboratory of veterinary expertise at the compound feed plant LLC “Athens-Volga”. For the study of compound feed an average sample was taken from each sample, separating from the combined sample using a hand scoop according to Federal standard 13496.0-2016. Methods of sampling. Quality and safety control was carried out according to organoleptic, physic and chemical, microbiological indicators and the content of mycotoxins in compound feed. Monitoring of compound feed of the compared samples of the starting compound feed for piglets SK-3 according to the studied indicators has shown that all samples meet the sanitary and hygienic requirements of Federal standard of our country.

scholarly journals German children’s processing of morphosyntactic cues in wh-questions

2018 ◽  
Vol 39 (6) ◽  
pp. 1279-1318 ◽  
Author(s):  
ATTY SCHOUWENAARS ◽  
PETRA HENDRIKS ◽  
ESTHER RUIGENDIJK
Keyword(s):  
Working Memory ◽  
Eye Tracking ◽  
Selection Task ◽  
Typically Developing ◽  
Garden Path ◽  
Verb Agreement ◽  
Case Marking ◽  
Wh Questions ◽  
German Children

ABSTRACTTwo experiments investigated the effects of case and verb agreement cues on the comprehension and production of which-questions in typically developing German children (aged 7–10) and adults. Our aims were to determine (a) whether they make use of morphosyntactic cues (case marking and verb agreement) for the comprehension of which-questions, (b) how these questions are processed, and (c) whether the presence and position of morphosyntactic cues available for the listener influence the speaker’s production of which-questions. Performance on a picture selection task with eye tracking shows that children with low working memory make less use of morphosyntactic cues than children with high working memory and adults when interpreting object questions. Gaze data of both groups reveal garden-path effects and revisions for object and passive questions, which can be explained by a constraint-based account. Furthermore, children’s difficulties with object questions are related to the type of disambiguation cue. In a question elicitation task with patient-initial items, children overall prefer production of passives, whereas adults’ productions depend on the availability of disambiguation cues for the listener.

scholarly journals Recurrent Stacking of Layers for Compact Neural Machine Translation Models

2019 ◽  
Vol 33 ◽  
pp. 6292-6299 ◽  
Author(s):  
Raj Dabre ◽  
Atsushi Fujita
Keyword(s):  
Machine Translation ◽  
Single Layer ◽  
Training Data ◽  
Neural Machine Translation ◽  
Parallel Corpora ◽  
Translation Quality ◽  
Sequence Generation ◽  
Sequence Modeling ◽  
Back Translation

In encoder-decoder based sequence-to-sequence modeling, the most common practice is to stack a number of recurrent, convolutional, or feed-forward layers in the encoder and decoder. While the addition of each new layer improves the sequence generation quality, this also leads to a significant increase in the number of parameters. In this paper, we propose to share parameters across all layers thereby leading to a recurrently stacked sequence-to-sequence model. We report on an extensive case study on neural machine translation (NMT) using our proposed method, experimenting with a variety of datasets. We empirically show that the translation quality of a model that recurrently stacks a single-layer 6 times, despite its significantly fewer parameters, approaches that of a model that stacks 6 different layers. We also show how our method can benefit from a prevalent way for improving NMT, i.e., extending training data with pseudo-parallel corpora generated by back-translation. We then analyze the effects of recurrently stacked layers by visualizing the attentions of models that use recurrently stacked layers and models that do not. Finally, we explore the limits of parameter sharing where we share even the parameters between the encoder and decoder in addition to recurrent stacking of layers.

scholarly journals Association Between Patient Survival and Clinician Variability in Treatment Rates for Aortic Valve Stenosis

2021 ◽  
Author(s):  
J. Matthew Brennan ◽  
Angela Lowenstern ◽  
Paige Sheridan ◽  
Isabel J. Boero ◽  
Vinod H. Thourani ◽  
...  
Keyword(s):  
Aortic Valve ◽  
Language Processing ◽  
Severe Aortic Stenosis ◽  
The United States ◽  
Risk Models ◽  
Transcatheter Aortic Valve ◽  
Competing Risk Models ◽  
The Impact ◽  
Potential Challenge

Background Patients with symptomatic severe aortic stenosis (ssAS) have a high mortality risk and compromised quality of life. Surgical/transcatheter aortic valve replacement (AVR) is a Class I recommendation, but it is unclear if this recommendation is uniformly applied. We determined the impact of managing cardiologists on the likelihood of ssAS treatment. Methods and Results Using natural language processing of Optum electronic health records, we identified 26 438 patients with newly diagnosed ssAS (2011–2016). Multilevel, multivariable Fine‐Gray competing risk models clustered by cardiologists were used to determine the impact of cardiologists on the likelihood of 1‐year AVR treatment. Within 1 year of diagnosis, 35.6% of patients with ssAS received an AVR; however, rates varied widely among managing cardiologists (0%, lowest quartile; 100%, highest quartile [median, 29.6%; 25th–75th percentiles, 13.3%–47.0%]). The odds of receiving AVR varied >2‐fold depending on the cardiologist (median odds ratio for AVR, 2.25; 95% CI, 2.14–2.36). Compared with patients with ssAS of cardiologists with the highest treatment rates, those treated by cardiologists with the lowest AVR rates experienced significantly higher 1‐year mortality (lowest quartile, adjusted hazard ratio, 1.22, 95% CI, 1.13–1.33). Conclusions Overall AVR rates for ssAS were low, highlighting a potential challenge for ssAS management in the United States. Cardiologist AVR use varied substantially; patients treated by cardiologists with lower AVR rates had higher mortality rates than those treated by cardiologists with higher AVR rates.

Sign in / Sign up

Export Citation Format

Share Document