The success of deep learning in natural language processing raises intriguing questions about the nature of linguistic meaning and ways in which it can be processed by natural and artificial systems. One such question has to do with subword segmentation algorithms widely employed in language modeling, machine translation, and other tasks since 2016. These algorithms often cut words into semantically opaque pieces, such as ‘period’, ‘on’, ‘t’, and ‘ist’ in ‘period|on|t|ist’. The system then represents the resulting segments in (...) a dense vector space, which is expected to model grammatical relations among them. This representation may in turn be used to map ‘period|on|t|ist’ (English) to ‘par|od|ont|iste’ (French). Thus, instead of being modeled at the lexical level, translation is reformulated more generally as the task of learning the best bilingual mapping between the sequences of subword segments of two languages; and sometimes even between pure character sequences: ‘p|e|r|i|o|d|o|n|t|i|s|t’ → ‘p|a|r|o|d|o|n|t|i|s|t|e’. Such subword segmentations and alignments are at work in highly efficient end-to-end machine translation systems, despite their allegedly opaque nature. The computational value of such processes is unquestionable. But do they have any linguistic or philosophical plausibility? I attempt to cast light on this question by reviewing the relevant details of the subword segmentation algorithms and by relating them to important philosophical and linguistic debates, in the spirit of making artificial intelligence more transparent and explainable. (shrink)

In general, stimuli that are familiar and recognizable have an advantage of predominance during binocular rivalry. Recent research has demonstrated that familiar and recognizable stimuli such as upright faces and words in a native language could break interocular suppression faster than their matched controls. In this study, a visible word prime was presented binocularly then replaced by a high-contrast dynamic noise pattern presented to one eye and either a semantically related or unrelated word was introduced to the other eye. We (...) measured how long it took for target words to break from suppression. To investigate word-parts priming, a second experiment also included word pairs that had overlapping subword fragments. Results from both experiments consistently show that semantically related words and words that shared subword fragments were faster to gain dominance compared to unrelated words, suggesting that words, even when interocularly suppressed and invisible, can benefit from semantic and subword priming. (shrink)

The proposed work describes an approach for segmentation of activated sludge flocs from the microscopic images for monitoring wastewater treatment. The morphological features of flocs and filaments are related to the state of an activated sludge wastewater treatment plant and must be monitored for proper functioning. Hence, image processing and analysis could be a time-saving monitoring tool. To address this challenge, we propose a novel framework involving a multiphase edge detection algorithm based on information theory. The proposed framework is (...) evaluated and scrutinized critically considering the artifacts found in the photographs tested. To evaluate the segmentations, gold approximation of estimated truth images is created. In addition, the performance was subjectively evaluated for its potential for segmenting activated sludge images. Experimental results show that the proposed framework exhibits the good results and demonstrates its effectiveness. (shrink)

We provide a new interpretation of Zeno’s Paradox of Measure that begins by giving a substantive account, drawn from Aristotle’s text, of the fact that points lack magnitude. The main elements of this account are (1) the Axiom of Archimedes which states that there are no infinitesimal magnitudes, and (2) the principle that all assignments of magnitude, or lack thereof, must be grounded in the magnitude of line segments, the primary objects to which the notion of linear magnitude applies. Armed (...) with this account, we are ineluctably driven to introduce a highly constructive notion of (outer) measure based exclusively on the total magnitude of potentially infinite collections of line segments. The Paradox of Measure then consists in the proof that every finite or potentially infinite collection of points lacks magnitude with respect to this notion of measure. We observe that the Paradox of Measure, thus understood, troubled analysts into the 1880’s, despite their knowledge that the linear continuum is uncountable. The Paradox was ultimately resolved by Borel in his thesis of 1893, as a corollary to his celebrated result that every countable open cover of a closed line segment has a finite sub-cover, a result he later called the “First Fundamental Theorem of Measure Theory.” This achievement of Borel has not been sufficiently appreciated. We conclude with a metamathematical analysis of the resolution of the paradox made possible by recent results in reverse mathematics. (shrink)

We discuss here a series of testable hypotheses concerning the role of chromosome folding into topologically associating domains (TADs). Several lines of evidence suggest that segmental packaging of chromosomal neighborhoods may underlie features of chromatin that span large domains, such as heterochromatin blocks, association with the nuclear lamina and replication timing. By defining which DNA elements preferentially contact each other, the segmentation of chromosomes into TADs may also underlie many properties of long‐range transcriptional regulation. Several observations suggest that TADs (...) can indeed provide a structural basis to regulatory landscapes, by controlling enhancer sharing and allocation. We also discuss how TADs may shape the evolution of chromosomes, by causing maintenance of synteny over large chromosomal segments. Finally we suggest a series of experiments to challenge these ideas and provide concrete examples illustrating how they could be practically applied. (shrink)

The formation of the segmentation pattern in Drosophila embryos provides an excellent model for investigating the process of pattern formation in multicellular organisms. Several genes required in an embryo for normal segmentation have been analyzed by classical and molecular genetic and morphological techniques. A detailed consideration of these results suggests that these segmentation genes are combinatorially involved in translating the positional identities of individual cells at an early stage in Drosophila development.

This paper reconsiders the diphone-based word segmentation model of Cairns, Shillcock, Chater, and Levy (1997) and Hockema (2006), previously thought to be unlearnable. A statistically principled learning model is developed using Bayes’ theorem and reasonable assumptions about infants’ implicit knowledge. The ability to recover phrase-medial word boundaries is tested using phonetic corpora derived from spontaneous interactions with children and adults. The (unsupervised and semi-supervised) learning models are shown to exhibit several crucial properties. First, only a small amount of language (...) exposure is required to achieve the model’s ceiling performance, equivalent to between 1 day and 1 month of caregiver input. Second, the models are robust to variation, both in the free parameter and the input representation. Finally, both the learning and baseline models exhibit undersegmentation, argued to have significant ramifications for speech processing as a whole. (shrink)

The formation of segments in the Drosophila early embryo is understood in greater detail than any other complex developmental process. Now, by studying other types of insect embryo, we can hope to deduce something of the ancestral mechanism of segmentation and the ways in which it has been modified in evolution. The parasitic wasp, Copidosoma floridanum, is spectacularly atypical of insects in that the small egg cell divides extensively, with no initial syncytial phase, and forms eventually some 2000 embryos(1). (...) This process raises intriguing questions about the control of embryonic polarity and segmentation. (shrink)

When infants and toddlers are confronted with sequences of sounds, they are required to segment the sounds into meaningful units to achieve sufficient understanding. Rhythm has been regarded as a crucial cue for segmentation of speech sounds. Although previous intermodal methods indicated that infants and toddlers could detect differences in speech sounds based on stress-timed and syllable-timed units, these methods could not clearly indicate how infants and toddlers perform sound segmentation. Thus, the present study examined whether Japanese infants (...) and toddlers could segment word speech sounds comprising basic morae, on the basis of concurrent basic mora units within syllable units, using the new intermodal matching procedure. The results indicated that, regardless of their ages and linguistic abilities, Japanese infants and toddlers aged 6–25 months tended to segment Japanese words comprising basic morae sounds on the basis of concurrent basic mora units within syllable units. This implies that infants' and toddlers' use of syllable units for segmentation of speech sounds at an early age could be evident among many infants and toddlers learning various languages. Although this finding should be interpreted carefully, the present study demonstrated the utility of the new intermodal matching procedure for examining segmentation of speech sounds and word sounds by infants and toddlers, on the basis of specific rhythm units. (shrink)

chical reinforcement learning that does not rely on a pri ori hierarchical structures Thus the approach deals with a more di cult problem compared with existing work It in volves learning to segment sequences to create hierarchical structures based on reinforcement received during task ex ecution with di erent levels of control communicating with each other through sharing reinforcement estimates obtained by each others The algorithm segments sequences to re duce non Markovian temporal dependencies to facilitate the learning of the (...) overall task Initial experiments demon strated the basic promise of the approach.. (shrink)

Despite decades of research, morphogenesis along the various body axes remains one of the major mysteries in developmental biology. A milestone in the field was the realisation that a set of closely related regulators, called Hox genes, specifies the identity of body segments along the anterior–posterior (AP) axis in most animals. Hox genes have been highly conserved throughout metazoan evolution and code for homeodomain‐containing transcription factors. Thus, they exert their function mainly through activation or repression of downstream genes. However, while (...) much is known about Hox gene structure and molecular function, only a few target genes have been identified and studied in detail. Our knowledge of Hox downstream genes is therefore far from complete and consequently Hox‐controlled morphogenesis is still poorly understood. Genome‐wide approaches have facilitated the identification of large numbers of Hox downstream genes both in Drosophila and vertebrates, and represent a crucial step towards a comprehensive understanding of how Hox proteins drive morphological diversification. In this review, we focus on the role of Hox genes in shaping segmental morphologies along the AP axis in Drosophila, discuss some of the conclusions drawn from analyses of large target gene sets and highlight methods that could be used to gain a more thorough understanding of Hox molecular function. In addition, the mechanisms of Hox target gene regulation are considered with special emphasis on recent findings and their implications for Hox protein specificity in the context of the whole organism. BioEssays 30:965–979, 2008. © 2008 Wiley Periodicals, Inc. (shrink)

We show that in the lattice E Π of Π 0 1 classes there are initial segments [ $\emptyset$ , P] = L(P) which are not Boolean algebras, but which have a decidable theory. In fact, we will construct for any finite distributive lattice L which satisfies the dual of the usual reduction property a Π 0 1 class P such that L is isomorphic to the lattice L(P)*, which is L(P), modulo finite differences. For the 2-element lattice, we obtain (...) a minimal class, first constructed by Cenzer, Downey, Jockusch and Shore in 1993. For the simplest new Π 0 1 class P constructed, P has a single, non-computable limit point and L(P)* has three elements, corresponding to $\emptyset$ , P and a minimal class P $_0 \subset$ P. The element corresponding to P 0 has no complement in the lattice. On the other hand, the theory of L(P) is shown to be decidable. A Π 0 1 class P is said to be decidable if it is the set of paths through a computable tree with no dead ends. We show that if P is decidable and has only finitely many limit points, then L(P)* is always a Boolean algebra. We show that if P is a decidable Π 0 1 class and L(P) is not a Boolean algebra, then the theory of L(P)interprets the theory of arithmetic and is therefore undecidable. (shrink)

I review how both the parasegmental and segmental frames are used in constructing the body plan of the arthropods. The parasegment is the primary genetic unit, as shown by Hox gene expression, and the parasegmental design is maintained in the nerve cord. It is, however, not maintained in the epidermis, where the cuticle grooves are segmental, and in the musculature, which is segmental in organisation. This frame shift is reflected in the sensory and motor nerve connections between the ganglia and (...) the periphery. I suggest that the need for movement in an organism equipped with a hard exoskeleton was the functional constraint that shaped this apparently complex mode of development. BioEssays 26:1117–1125, 2004. © 2004 Wiley Periodicals, Inc. (shrink)

In heterogeneous markets, one of the many consumer groups is that of green product buyers. With rising ethical values, the green market is assuming its place in a general growth trend. Given this, it is important to determine the profile of green product buyers. This study aims to find out whether there are sub-markets for green product buyers, based on their personal values and consumption values, and to determine a detailed profile for these buyers. Both personal values and consumption values (...) are basic factors guiding consumer behaviour and affecting consumption preferences. The data was collected, through surveys in Turkey, from green consumers who were members of the TEMA (the Turkish Foundation for Combating Soil Erosion, for Reforestation and the Protection of Natural Habitats) council. The result of our clustering analysis indicates that green product buyers could be segmented into sub-groups according to their personal values and consumption values. (shrink)

According to mainstream linguistic phonetics, speech can be modeled as a string of discrete sound segments or “phones” drawn from a universal phonetic inventory. Recent work has argued that a mature phonetics should refrain from theorizing about speech and speech processing using sound segments, and that the phone concept should be eliminated from linguistic theory. The paper lays out the tenets of the phone methodology and evaluates its prospects in light of the eliminativist arguments. I claim that the eliminativist arguments (...) fail to show that the phone concept should be eliminated from linguistic theory. (shrink)

Using properties of${\cal K}$-pairs of sets, we show that every nonzero enumeration degreeabounds a nontrivial initial segment of enumeration degrees whose nonzero elements have all the same jump asa. Some consequences of this fact are derived, that hold in the local structure of the enumeration degrees, including: There is an initial segment of enumeration degrees, whose nonzero elements are all high; there is a nonsplitting high enumeration degree; every noncappable enumeration degree is high; every nonzero low enumeration degree can be (...) capped by degrees of any possible local jump ; every enumeration degree that bounds a nonzero element of strictly smaller jump, is bounding; every low enumeration degree below a non low enumeration degreeacan be capped belowa. (shrink)

MOTIVATION The constructivization of ; => D(_£^') poses several problems. For some of these the tools of MD can be modified; for others new methods will need to be established. What must we do to make a full approximation to ; * D? We ...

Cet article se propose d’appréhender le segment du point de vue de la production, en analysant les séquences identifiées lors du processus d’écriture, c’est-à-dire les unités langagières produites de manière spontanée et ininterrompue (entre deux pauses). Ces unités, que nous appelons jets textuels, ont été enregistrées grâce à l’outil de suivi de rédaction Inputlog. Nous avons tenté de caractériser les jets textuels en les mettant en regard avec divers types d’unités segmentales de langage, qu’elles soient issues de la segmentation (...) des textes finalisés, ou reconnues comme des segments préfabriqués propres à un système lexical ou syntaxique, c’est-à-dire des figements au sens large du terme. Nous nous sommes plus particulièrement penchés sur un parcours de segmentation parallèle, en production (jets textuels) et en réception (segments répétés), pour interroger quelques concepts théoriques et méthodologiques qui sous-tendent les démarches de segmentation : pré-construit, automatisme, mémorisation, figement, enracinement cognitif, fréquence, répétition. (shrink)

It is known that every non-universal self-full degree in the structure of the degrees of computably enumerable equivalence relations (ceers) under computable reducibility has exactly one strong minimal cover. This leaves little room for embedding wide partial orders as initial segments using self-full degrees. We show that considerably more can be done by staying entirely inside the collection of non-self-full degrees. We show that the poset can be embedded as an initial segment of the degrees of ceers with infinitely many (...) classes. A further refinement of the proof shows that one can also embed the free distributive lattice generated by the lower semilattice as an initial segment of the degrees of ceers with infinitely many classes. (shrink)

Dans cet article, nous appuyant sur un corpus de 3 444 075 tweets correspondant à 44 107 210 tokens (mots, signes de ponctuation, émojis, émoticônes, etc.) recueillis en décembre 2016, nous nous intéressons aux procédés de segmentation à l’œuvre dans les tweets. Après avoir évoqué certaines caractéristiques de ces écrits particuliers, nous rappelons les procédés généraux de segmentation à l’écrit : les signes de ponctuation et les connecteurs. Nous nous penchons ensuite sur la segmentation opérée dans les (...) tweets par ces deux procédés généraux. Enfin, nous montrons que les émoticônes et les émojis constituent des procédés spécifiques permettant de diversifier les stratégies de segmentation des utilisateurs de tweets (et d’autres écrits numériques, tels les SMS et les courriels). (shrink)

Cette livraison de Corpus propose de s’intéresser aux faits phraséologiques, c’est-à-dire à cet ensemble de « prêts-à-dire », de « déjà-là », de formules « clef-en-main » aux fonctions communicationnelles variées : zones de confort pour les uns, les faits phraséologiques peuvent faciliter la prise de parole ou les tâches rédactionnelles (Née, Sitri, Veniard 2014 ; Cislaru, Sitri, Pugnière-Saavedra 2013), contribuer à marquer une identité sociale ou professionnelle (Römer 2010 ; Née, Sitri, Ve...

Studies on Neurospora chromosome segment duplications (Dps) performed since the publication of Perkins's comprehensive review in 1997 form the focus of this article. We present a brief summary of Perkins's seminal work on chromosome rearrangements, specifically, the identification of insertional and quasiterminal translocations that can segregate Dp progeny when crossed with normal sequence strains (i.e., T × N). We describe the genome defense process called meiotic silencing by unpaired DNA that renders Dp‐heterozygous crosses (i.e., Dp × N) barren, which provides (...) a basis for identifying Dps, and discuss whether other processes also might contribute to the barren phenotype of Dp × N and Dp × Dp crosses. We then turn to studies suggesting that large Dps (i.e., >300 kbp) can allow smaller gene‐sized duplications to escape another genome defense process called repeat‐induced point mutation (RIP), possibly by titration of the RIP machinery. Finally, we assess whether in natural populations dominant RIP suppressor Dps provide an “RIP‐free” niche for evolution of new genes following the duplication of existing genes. (shrink)

Special Care: Medical Decisions at the Beginning of Life, by Fred FronhockPlaying God in the Nursery, by Jeff LyonsWithholding Treatment From Defective Newborn Children, Joseph E. Magnet and Eike‐Henner W. KlugeCenter for Ethics, Medicine and Public Issues Baylor College of MedicineLAW AND MORALS by Simon Lee. OxfordThe Foetus as Organ donor: Scientific, Social and Ethical Perspectives. By Peter McCullaghTransplantation Unit The Walter & Eliza Hall Institute of Medical Research.

This study analyzes the most important predictors of acceptance of social network sites in a sample of Chilean elder people. We employ a novelty procedure to explore this phenomenon. This procedure performs apriori segmentation based on gender and generation. It then applies the deep learning technique to identify the predictors by segments. The predictor variables were taken from the literature on the use of social network sites, and an empirical study was carried out by quota sampling with a sample (...) size of 395 older people. The results show different predictors of social network sites considering all the samples, baby boomer males and females, silent males and females. The high heterogeneity among older people is confirmed; this means that dealing with older adults as a uniform set of users of social network sites is a mistake. This study demonstrates that the four segments behave differently, and many diverse variables influence the acceptance of social network sites. (shrink)

River segmentation of remote sensing images is of important research significance and application value for environmental monitoring, disaster warning, and agricultural planning in an area. In this study, we propose a river segmentation model in remote sensing images based on composite attention network to solve the problems of abundant river details in images and the interference of non-river information including bridges, shadows, and roads. To improve the segmentation efficiency, a composite attention mechanism is firstly introduced in the (...) central region of the network to obtain the global feature dependence of river information. Next, in this study, we dynamically combine binary cross-entropy loss that is designed for pixel-wise segmentation and the Dice coefficient loss that measures the similarity of two segmentation objects into a weighted one to optimize the training process of the proposed segmentation network. The experimental results show that compared with other semantic segmentation networks, the evaluation indexes of the proposed method are higher than those of others, and the river segmentation effect of CoANet model is significantly improved. This method can segment rivers in remote sensing images more accurately and coherently, which can meet the needs of subsequent research. (shrink)

A standard result about hyperdegrees is proved (under PD) for all $\triangle^1_{2n + 1}$-degrees.

A standard result about hyperdegrees is proved for allΔ2n+11-degrees.

The relatively simple central nervous system (CNS) of the Drosophila embryo provides a useful model system for investigating the mechanisms that generate and pattern complex nervous systems. Central to the generation of different types of neurons by precursor neuroblasts is the initial specification of neuroblast identity and the Drosophila segment polarity genes, genes that specify regions within a segment or repeating unit of the Drosophila embryo, have emerged recently as significant players in this process. During neurogenesis the segment polarity genes (...) are expressed in the neuroectodermal cells from which neuroblasts delaminate and they continue to be expressed in neuroblasts and their progeny. Loss-of-function mutations in these genes lead to a failure in the formation of neuroblasts and/or specification of neuroblast identity. Results from several recent studies suggest that regulatory interactions between segment polarity genes during neurogenesis lead to an increase in the number of neuroblasts and specification of different identities to neuroblasts within a population of cells. BioEssays 21:472–485, 1999. © 1999 John Wiley & Sons, Inc. (shrink)