This work reimplements a recent semantic bootstrapping child language acquisition (CLA) model, which was originally designed for English, and trains it to learn a new language: Hebrew. The model learns from pairs of utterances and logical forms as meaning representations, and acquires both syntax and word meanings simultaneously. The results show that the model mostly transfers to Hebrew, but that a number of factors, including the richer morphology in Hebrew, makes the learning slower and less robust. This suggests that a clear direction for future work is to enable the model to leverage the similarities between different word forms.

Large language models (LLMs) have shown impressive performance in code understanding and generation, making coding tasks a key focus for researchers due to their practical applications and value as a testbed for LLM evaluation. Data synthesis and filtering techniques have been widely adopted and shown to be highly effective in this context. In this paper, we present a focused survey and taxonomy of these techniques, emphasizing recent advancements. We highlight key challenges, explore future research directions, and offer practical guidance for new researchers entering the field.

This work develops a probabilistic child language acquisition model to learn a range of linguistic phenonmena, most notably long-range syntactic dependencies of the sort found in object wh-questions, among other constructions. The model is trained on a corpus of real child-directed speech, where each transcribed utterance is paired with a logical form as a meaning representation. It then learns both word meanings and language-specific syntax simultaneously. After training, the model can deduce the correct parse tree and word meanings for a given string-meaning pair, and can infer the meaning given only the string.

Narrative Question Answering is an important task for evaluating and improving reading comprehension abilities in both humans and machines. However, there is a lack of consensus on the skill taxonomy that would enable systematic and comprehensive assessment and learning of the various aspects of Narrative Question Answering. Existing task-level skill views oversimplify the multidimensional nature of tasks, while question-level taxonomies face issues in evaluation and methodology. To address these challenges, we introduce a more inclusive skill taxonomy that synthesizes and redefines narrative understanding skills from previous taxonomies and includes a generation skill dimension from the answering perspective.

Large Language Models (LLMs) are claimed to be capable of Natural Language Inference (NLI), necessary for applied tasks like question answering and summarization. We present a series of behavioral studies on several LLM families (LLaMA, GPT-3.5, and PaLM) which probe their behavior using controlled experiments. We establish two biases originating from pretraining which predict much of their behavior, and show that these are major sources of hallucination in generative LLMs. First, memorization at the level of sentences: we show that, regardless of the premise, models falsely label NLI test samples as entailing when the hypothesis is attested in training data, and that entities are used as “indices” to access the memorized data. Second, statistical patterns of usage learned at the level of corpora: we further show a similar effect when the premise predicate is less frequent than that of the hypothesis in the training data, a bias following from previous studies. We demonstrate that LLMs perform significantly worse on NLI test samples which do not conform to these biases than those which do, and we offer these as valuable controls for future LLM evaluation.

The diversity and Zipfian frequency distribution of natural language predicates in corpora leads to sparsity in Entailment Graphs (EGs) built by Open Relation Extraction (ORE). EGs are computationally efficient and explainable models of natural language inference, but as symbolic models, they fail if a novel premise or hypothesis vertex is missing at test-time. We present theory and methodology for overcoming such sparsity in symbolic models. First, we introduce a theory of optimal smoothing of EGs by constructing transitive chains. We then demonstrate an efficient, open-domain, and unsupervised smoothing method using an off-the-shelf Language Model to find approximations of missing premise predicates. This improves recall by 25.1 and 16.3 percentage points on two difficult directional entailment datasets, while raising average precision and maintaining model explainability. Further, in a QA task we show that EG smoothing is most useful for answering questions with lesser supporting text, where missing premise predicates are more costly. Finally, controlled experiments with WordNet confirm our theory and show that hypothesis smoothing is difficult, but possible in principle.

While it is well established that humans develop stronger relationship bonds when they share stressful experiences, there is little known on how shared stressful experiences may influence relationship bonding in animals. Here, we present a study looking at social proximity between individuals in small groups of Merino ewes following a shared stressful experience compared with control sheep that were not exposed to stress. Some sheep were familiar to each other. Analyses of social proximity using real-time-kinematic Global Navigation Satellite System (GNSS) on-animal devices showed sheep preferred to be closest to familiar individuals, but across the study duration they also developed a preference for the individuals they shared the stressful experience with, relative to their proximity to control individuals. These results contribute to limited research on what factors may instigate the development of bonds between unfamiliar sheep. Between-individual bonds may develop as a means of socially mediated stress buffering. Social bonding following a shared stressful experience aligns with human social relationships and increases our understanding of how animals perceive their conspecifics in relation to stressful environmental change.

There has recently been an increased interest in studying the language development of non-western languages. This is not new - it began in 1960's and continued into the 1980's and 1990's. The current renewed interest is much welcomed, and will benefit from many of the experimental methods and theoretical insights developed over the past decades.

Sensory data is only meaningful if correctly associated with originating locations, e.g., in three-dimensional wireless sensor systems inside an aircraft cabin. Received signal strength is a cost-effective option to locate low-cost sensors due to its universal availability, but suffers from coarse ranging accuracy with multiplicative errors. This paper presents a new received signal strength-based approach to locating a large number of low-cost sensors given their possible three-dimensional installation points. Our approach achieves the accuracy which has not been achieved in the literature. The approach first cooperatively locates the sensors in the continuous three-dimensional spaces, then associates the continuous location estimates to the installation points, and finally refines the association with likelihood ascent search. A new convex relaxation-based optimization is designed for cooperative localization in continuous three-dimensional spaces. The Kuhn-Munkres algorithm is generalized for the association. Cramér-Rao Lower Bounds are derived to specify the local three-dimensional regions in which refinement is carried out to improve the final accuracy with little complexity overhead. The proposed approach is validated experimentally with signal strength measurements collected in a Fokker 100 passenger plane, achieves 100% accuracy in all experiments conducted, and outperforms state-of-the-art metaheuristics significantly with much shorter execution time. Note to Practitioners—This paper is motivated by a goal to realize various aircraft automation applications through three-dimensional localization and association of wireless sensors. For example, automatic identification of seat locations can be implemented efficiently with wireless sensors, which enables effortless re-association between control buttons in the seats and the corresponding functions after cabin refurbishment. Meanwhile, automatic detection of missing/misplaced safety equipment can be carried out to improve the efficiency of flight preparation. There currently lacks an received signal strength-based three-dimensional sensor association method in indoor environments. The theoretical framework proposed in this paper aims to utilize the readily available received signal strength measurements to associate Commercial-Off-The-Shelf sensors to their possible installation points. This approach avoids the requirement of labor-intensive fingerprinting effort and high-cost equipment. The proposed approach has been tested and evaluated using both synthetic data and real data collected in a passenger plane as a proof of concept.

Sheep are gregarious animals that can form close inter-individual relationships. Social bonds are affected by environmental and within-animal factors, but there is limited understanding of individual factors that influence social relationships, in part due to the challenges of accurately measuring individual social proximity within a group. This study used geospatially precise RTK-GNSS localization devices (Real-Time-Kinematic Global Navigation Satellite System) to evaluate social proximity between individuals with different personality traits in small sheep flocks. First, unsupervised clustering algorithms were used to categorize 78 individual Merino sheep (from a research flock of 108 grazed together for 11 months prior) based on their behavior measured in four behavioral tests: arena, novel object, flight speed, and isolation box. Sixty-five sheep were then categorized into five cohorts of 13 individuals each (clustered into three bold, six bold/fearful, and four fearful animals per cohort). Bold/fearful animals were bold toward a novel object but fearful of a human. Sheep within a single cohort were released into a 0.70 Ha test paddock with RTK-GNSS devices attached to their backs for 3 days to monitor relative positions. All cohorts were tested sequentially over a 5-week period. Social network analyses of GNSS locational data with a distance threshold of up to 30 m for all cohorts together showed no significant effect of personality traits (P = 0.11). There were, however, some significant but inconsistent differences in distances based on personality traits within separate cohorts [i.e., cohorts 1, 3, 4 (P ≤ 0.02), cohorts 2, 5 (P ≥ 0.44)]. When GNSS locational data were analyzed with a distance threshold of up to 2 m (day 2 excluded from all cohorts due to missing values), there was a significant effect of personality traits (P = 0.01) across all cohorts combined. The bold animals showed greater distances from each other relative to their distances from fearful or bold/fearful individuals. Additionally, regardless of personality traits, the average distance between animals decreased 2.2 m over the 3 test days. Social interactions in sheep flocks may be influenced by personality traits of fearfulness and boldness, but further research with larger numbers of sheep across varying environments is needed.

Mark A. I. Johnson,1, 2 Mateusz T. Mądzik,1, 2 Fay E. Hudson,1 Kohei M. Itoh,3 Alexander M. Jakob,4, 2 David N. Jamieson,4, 2 Andrew Dzurak,1 and Andrea Morello1, 2, ∗ 1School of Electrical Engineering and Telecommunications, UNSW Sydney, Sydney, NSW 2052, Australia 2Centre of Excellence for Quantum Computation & Communication Technology 3School of Fundamental Science and Technology, Keio University, Kohoku-ku, Yokohama, Japan 4School of Physics, University of Melbourne, Melbourne, VIC 3010, Australia (Dated: October 6, 2021)

Treatments of atherosclerosis depend on the severity of the disease at the diagnosis time. Non-invasive diagnosis techniques, capable of detecting stenosis at early stages, are essential to reduce associated costs and mortality rates. We used computational fluid dynamics and acoustics analysis to extensively investigate the sound sources arising from high-turbulent fluctuating flow through stenosis. The frequency spectral analysis and proper orthogonal decomposition unveiled the frequency contents of the fluctuations for different severities and decomposed the flow into several frequency bandwidths. Results showed that high-intensity turbulent pressure fluctuations appeared inside the stenosis for severities above 70%, concentrated at plaque surface, and immediately in the post-stenotic region. Analysis of these fluctuations with the progression of the stenosis indicated that (a) there was a distinct break frequency for each severity level, ranging from 40 to 230 Hz, (b) acoustic spatial-frequency maps demonstrated the variation of the frequency content with respect to the distance from the stenosis, and (c) high-energy, high-frequency fluctuations existed inside the stenosis only for severe cases. This information can be essential for predicting the severity level of progressive stenosis, comprehending the nature of the sound sources, and determining the location of the stenosis with respect to the point of measurements.

Understanding linguistic modality is widely seen as important for downstream tasks such as Question Answering and Knowledge Graph Population. Entailment Graph learning might also be expected to benefit from attention to modality. We build Entailment Graphs using a news corpus filtered with a modality parser, and show that stripping modal modifiers from predicates in fact increases performance. This suggests that for some tasks, the pragmatics of modal modification of predicates allows them to contribute as evidence of entailment.

Computational Linguistics and Natural Language Processing have changed considerably in the past few decades. Early research focused on representing and using linguistic knowledge in computational processes such as parsers, while these days the field focuses on practically-useful tasks such as information retrieval and chatbots. Currently our Deep Learning models have little to do with linguistic theory For example, the Oracle Digital Assistant is built on top of generic "Foundation" Deep Learning models. An intermediate Focusing step adapts these models to specific enterprise domains. Transfer Learning is used to refocus these models onto specific customer-oriented tasks such as Intent Classification, Named Entity Recognition, as well as more advanced models such as text-to-SQL sequence-to-sequence models. These technologies have revolutionised the application of NLP to practical problems with commercial relevance, enabling us to build better systems faster and cheaper than ever before. Linguistic insights aren't gone from the field, however; they play a critical role in data manufacturing and evaluation. This talk explain how we use hundreds of different evaluations to understand the strengths and weaknesses of our models in the Oracle Digital Assistant, and how we automatically use this in hyper-parameter tuning. It also describes areas where additional research is still required before we can claim that NLP has become an engineering field.

A data-oriented parsing or DOP model for statistical parsing associates fragments of linguistic representations with numerical weights, where these weights are estimated by normalizing the empirical frequency of each fragment in a training corpus (see Bod [1998] and references cited therein). This note observes that this estimation method is biased and inconsistent that is, the estimated distribution does not in general converge on the true distribution as the size of the training corpus increases.

For the early detection of atherosclerosis, it is imperative to explore the capabilities of new, effective noninvasive diagnosis techniques to significantly reduce the associated treatment costs and mortality rates. In this study, a multifaceted comprehensive approach involving advanced computational fluid dynamics combined with signal processing techniques was exploited to investigate the highly turbulent fluctuating flow through arterial stenosis. The focus was on localizing high-energy mechano-acoustic source potential to transmit to the epidermal surface. The flow analysis results showed the existence of turbulent pressure fluctuations inside the stenosis and in the post-stenotic region. After analyzing the turbulent kinetic energy and pressure fluctuations on the flow centerline and the vessel wall, the point of maximum excitation in the flow was observed around two diameters downstream of the stenosis within the fluctuating zone. It was also found that the concentration of pressure fluctuation closer to the wall was higher inside the stenosis compared to the post-stenotic region. Additionally, the visualization of the most energetic proper orthogonal decomposition (POD) mode and spectral decomposition of the flow indicated that the break frequencies ranged from 80 to 220 Hz and were correlated to the eddies generated within these regions.

Relation prediction informed from a combination of text corpora and curated knowledge bases, combining knowledge graph completion with relation extraction, is a relatively little studied task. A system that can perform this task has the ability to extend an arbitrary set of relational database tables with information extracted from a document corpus. OpenKi[1] addresses this task through extraction of named entities and predicates via OpenIE tools then learning relation embeddings from the resulting entity-relation graph for relation prediction, outperforming previous approaches. We present an extension of OpenKi that incorporates embeddings of text-based representations of the entities and the relations. We demonstrate that this results in a substantial performance increase over a system without this information.

Sequence-to-Sequence (S2S) neural text generation models, especially the pre-trained ones (e.g., BART and T5), have exhibited compelling performance on various natural language generation tasks. However, the black-box nature of these models limits their application in tasks where specific rules (e.g., controllable constraints, prior knowledge) need to be executed. Previous works either design specific model structure (e.g., Copy Mechanism corresponding to the rule "the generated output should include certain words in the source input") or implement specialized inference algorithm (e.g., Constrained Beam Search) to execute particular rules through the text generation. These methods require careful design case-by-case and are difficult to support multiple rules concurrently. In this paper, we propose a novel module named Neural Rule-Execution Tracking Machine that can be equipped into various transformer-based generators to leverage multiple rules simultaneously to guide the neural generation model for superior generation performance in a unified and scalable way. Extensive experimental results on several benchmarks verify the effectiveness of our proposed model in both controllable and general text generation.

This paper focuses on Seq2Seq (S2S) constrained text generation where the text generator is constrained to mention specific words, which are inputs to the encoder, in the generated outputs. Pre-trained S2S models such as T5 or a Copy Mechanism can be trained to copy the surface tokens from encoders to decoders, but they cannot guarantee constraint satisfaction. Constrained decoding algorithms always produce hypotheses satisfying all constraints. However, they are computationally expensive and can lower the generated text quality. In this paper, we propose Mention Flags (MF), which trace whether lexical constraints are satisfied in the generated outputs of an S2S decoder. The MF models are trained to generate tokens until all constraints are satisfied, guaranteeing high constraint satisfaction. Our experiments on the Common Sense Generation task (CommonGen) (Lin et al., 2020), End2end Data-to-Text task (E2ENLG) (Dusek et al., 2020) and Novel Object Captioning task (nocaps) (Agrawal et al., 2019) show that the MF models maintain higher constraint satisfaction and text quality than the baseline models and other constrained text generation algorithms, achieving state-of-the-art performance on all three tasks. These results are achieved with a much lower run-time than constrained decoding algorithms. We also show that the MF models work well in the low-resource setting.(1)