Image Chunking: Defining Spatial Building Blocks for Scene Analysis

Rapid judgments about the properties and spatial relations of objects are the crux of visually guided interaction with the world. Vision begins, however, with essentially pointwise representations of the scene, such as arrays of pixels or small edge fragments. For adequate time-performance in recognition, manipulation, navigation, and reasoning, the processes that extract meaningful entities from the pointwise representations must exploit parallelism. This report develops a framework for the fast extraction of scene entities, based on a simple, local model of parallel computation.sAn image chunk is a subset of an image that can act as a unit in the course of spatial analysis. A parallel preprocessing stage constructs a variety of simple chunks uniformly over the visual array. On the basis of these chunks, subsequent serial processes locate relevant scene components and assemble detailed descriptions of them rapidly. This thesis defines image chunks that facilitate the most potentially time-consuming operations of spatial analysis---boundary tracing, area coloring, and the selection of locations at which to apply detailed analysis. Fast parallel processes for computing these chunks from images, and chunk-based formulations of indexing, tracing, and coloring, are presented. These processes have been simulated and evaluated on the lisp machine and the connection machine.

Le chunking perceptif de la parole : sur la nature du groupement temporel et son effet sur la mémoire immédiate

Dans de nombreux comportements qui reposent sur le rappel et la production de séquences, des groupements temporels émergent spontanément, créés par des délais ou des allongements. Ce « chunking » a été observé tant chez les humains que chez certains animaux et plusieurs auteurs l’attribuent à un processus général de chunking perceptif qui est conforme à la capacité de la mémoire à court terme. Cependant, aucune étude n’a établi comment ce chunking perceptif s’applique à la parole. Nous présentons une recension de la littérature qui fait ressortir certains problèmes critiques qui ont nui à la recherche sur cette question. C’est en revoyant ces problèmes qu’on propose une démonstration spécifique du chunking perceptif de la parole et de l’effet de ce processus sur la mémoire immédiate (ou mémoire de travail). Ces deux thèmes de notre thèse sont présentés séparément dans deux articles. Article 1 : The perceptual chunking of speech: a demonstration using ERPs Afin d’observer le chunking de la parole en temps réel, nous avons utilisé un paradigme de potentiels évoqués (PÉ) propice à susciter la Closure Positive Shift (CPS), une composante associée, entre autres, au traitement de marques de groupes prosodiques. Nos stimuli consistaient en des énoncés et des séries de syllabes sans sens comprenant des groupes intonatifs et des marques de groupements temporels qui pouvaient concorder, ou non, avec les marques de groupes intonatifs. Les analyses démontrent que la CPS est suscitée spécifiquement par les allongements marquant la fin des groupes temporels, indépendamment des autres variables. Notons que ces marques d’allongement, qui apparaissent universellement dans la langue parlée, créent le même type de chunking que celui qui émerge lors de l’apprentissage de séquences par des humains et des animaux. Nos résultats appuient donc l’idée que l’auditeur chunk la parole en groupes temporels et que ce chunking perceptif opère de façon similaire avec des comportements verbaux et non verbaux. Par ailleurs, les observations de l’Article 1 remettent en question des études où on associe la CPS au traitement de syntagmes intonatifs sans considérer les effets de marques temporels. Article 2 : Perceptual chunking and its effect on memory in speech processing:ERP and behavioral evidence Nous avons aussi observé comment le chunking perceptif d’énoncés en groupes temporels de différentes tailles influence la mémoire immédiate d’éléments entendus. Afin d’observer ces effets, nous avons utilisé des mesures comportementales et des PÉ, dont la composante N400 qui permettait d’évaluer la qualité de la trace mnésique d’éléments cibles étendus dans des groupes temporels. La modulation de l’amplitude relative de la N400 montre que les cibles présentées dans des groupes de 3 syllabes ont bénéficié d’une meilleure mise en mémoire immédiate que celles présentées dans des groupes plus longs. D’autres mesures comportementales et une analyse de la composante P300 ont aussi permis d’isoler l’effet de la position du groupe temporel (dans l’énoncé) sur les processus de mise en mémoire. Les études ci-dessus sont les premières à démontrer le chunking perceptif de la parole en temps réel et ses effets sur la mémoire immédiate d’éléments entendus. Dans l’ensemble, nos résultats suggèrent qu’un processus général de chunking perceptif favorise la mise en mémoire d’information séquentielle et une interprétation de la parole « chunk par chunk ».

Expressões formulaicas e collocations em inglês: uma proposta de descrição por meio de esquemas de imagem e chunking

Pós-graduação em Linguística e Língua Portuguesa - FCLAR

Usage-based chunking of software architecture information to assist information finding

One of the key problems with Software Architecture Documents (ADs) 2 is the difficulty of finding in- formation required from them. Most existing studies focus on the production of ADs or Architectural Knowledge (AK) 3 , to allow them to support information finding. However, there has been little focus placed on the consumption of ADs. To address this, we postulate the existence of a concept of “usage- based chunks”of architectural information discoverable from consumers’ usage of ADs when they engage in information-seeking tasks. In a set of user studies, we have found evidence that such usage-based chunks exist and that useful chunks can be identified from one type of usage data, namely, consumer’s ratings of sections of ADs. This has implications for tool design to support the effective reuse of AK.

Early relational reasoning and the novice programmer : swapping as the “hello world” of relational reasoning

We report on a longitudinal research study of the development of novice programmers in their first semester of programming. In the third week, almost half of our sample of students could not answer an explain-in-plain-English question, for code consisting of just three assignment statements, which swapped the values in two variables. We regard code that swaps the values of two variables as the simplest case of where a programming student can manifest a SOLO relational response. Our results demonstrate that the problems many students face with understanding code can begin very early, on relatively trivial code. However, using traditional programming exercises, these problems often go undetected until late in the semester. New approaches are required to detect and fix these problems earlier.

Spontaanin puheen prosodinen jaksottelu

The common focus of the studies brought together in this work is the prosodic segmentation of spontaneous speech. The theoretically most central aspect is the introduction and further development of the IJ-model of intonational chunking. The study consists of a general introduction and five detailed studies that approach prosodic chunking from different perspectives. The data consist of recordings of face-to-face interaction in several spoken varieties of Finnish and Finland Swedish; the methodology is usage-based and qualitative. The term “speech prosody” refers primarily to the melodic and rhythmic characteristics of speech. Both speaking and understanding speech require the ability to segment the flow of speech into suitably sized prosodic chunks. In order to be usage-based, a study of spontaneous speech consequently needs to be based on material that is segmented into prosodic chunks of various sizes. The segmentation is seen to form a hierarchy of chunking. The prosodic models that have so far been developed and employed in Finland have been based on sentences read aloud, which has made it difficult to apply these models in the analysis of spontaneous speech. The prosodic segmentation of spontaneous speech has not previously been studied in detail in Finland. This research focuses mainly on the following three questions: (1) What are the factors that need to be considered when developing a model of prosodic segmentation of speech, so that the model can be employed regardless of the language or dialect under analysis? (2) What are the characteristics of a prosodic chunk, and what are the similarities in the ways chunks of different languages and varieties manifest themselves that will make it possible to analyze different data according to the same criteria? (3) How does the IJ-model of intonational chunking introduced as a solution to question (1) function in practice in the study of different varieties of Finnish and Finland Swedish? The boundaries of the prosodic chunks were manually marked in the material according to context-specific acoustic and auditory criteria. On the basis of the data analyzed, the IJ-model was further elaborated and implemented, thus allowing comparisons between different language varieties. On the basis of the empirical comparisons, a prosodic typology is presented for the dialects of Swedish in Finland. The general contention is that the principles of the IJ-model can readily be used as a methodological tool for prosodic analysis irrespective of language varieties.

Constructing skilled images

When experts construct mental images, they do not rely only on perceptual features; they also access domain-specific knowledge and skills in long-term memory, which enables them to exceed the capacity limitations of the short-term working memory system. The central question of the present dissertation was whether the facilitating effect of long-term memory knowledge on working memory imagery tasks is primarily based on perceptual chunking or whether it relies on higher-level conceptual knowledge. Three domains of expertise were studied: chess, music, and taxi driving. The effects of skill level, stimulus surface features, and the stimulus structure on incremental construction of mental images were investigated. A method was developed to capture the chunking mechanisms that experts use in constructing images: chess pieces, street names, and visual notes were presented in a piecemeal fashion for later recall. Over 150 experts and non-experts participated in a total of 13 experiments, as reported in five publications. The results showed skill effects in all of the studied domains when experts performed memory and problem solving tasks that required mental imagery. Furthermore, only experts' construction of mental images benefited from meaningful stimuli. Manipulation of the stimulus surface features, such as replacing chess pieces with dots, did not significantly affect experts' performance in the imagery tasks. In contrast, the structure of the stimuli had a significant effect on experts' performance in every task domain. For example, taxi drivers recalled more street names from lists that formed a spatially continuous route than from alphabetically organised lists. The results suggest that the mechanisms of conceptual chunking rather than automatic perceptual pattern matching underlie expert performance, even though the tasks of the present studies required perception-like mental representations. The results show that experts are able to construct skilled images that surpass working memory capacity, and that their images are conceptually organised and interpreted rather than merely depictive.

汉语韵律短语切分及语调降阶和下倾研究

Firstly, prosodic boundaries of 1991 common sentences were labeled based on speech perception experiment, relation between prosodic structure and syntactic structure was examined after immediate constituent analysis, an example of prosodic phrasing from text sentences was provided using CART. Then, using designed sentences, phenomena of downstep and declination in pitch downtrend of Chinese declarative sentences were examined, commonness and speciality of Chinese intonation were discussed. The main results of the study are: 1 The distribution patterns of prosodic phrase boundaries for different syntactic structures are different, and there is great freedom in prosodic chunking. The relation between syntactic structure and prosodic structure can only be discussed in statistical sense. 2 Besides of syntactic relation, the second most important factor which influences prosodic phrase boundaries is length. The distances to the front boundary and the back boundary are more important than the lengths of the left syntactic contituent and the right one. In our corpus, the length distributions of prosodic phrases are 5±3 syllables. 3 Automatic downstep can lower intonation linearly, but is affected by stress easily. Non-automatic downstep lowers the higher part of pitch contours and has no effect on the lower one of the intonation. 4 The downtrend reason of low point is declination. The extent of declination relates to not only tones of low points, but also their positions in prosodic words, the baselines decline much faster when low point are in the initial position of a prosodic word. In long sentences, the baselines of prosodic phrases are the basic declination units, and the whole declination pattern of a sentence is related to syntactic relations between two neighboring prosodic phrases.

顿悟中的蔡格尼克效应：左右脑在解题失败与提示信息加工时的活动差异

Zeigarnik effect refers to the enhanced memory performance for unfinished tasks and studies on insight using hemi-visual field presentation technology also find that after failing to solve an problem, hints to the problem are more effective received and lead to insight experience when presented to the left-visual field (Right hemisphere) than presented to the right-visual field, especial when the hints appeared with a delay. Thus, it seems that right hemisphere may play an important role in preserving information of unsolved problems and processing related cues. To further examine the finding above, we introduce an Chinese character chunking task to investigate the brain activities during the stage of failure to resolve problems and of hint presentation using Event-Related Potentials (ERP) and functional MRI technology. Our FMRI results found that bilateral BA10 showed more activation when seeing hints for unsolved problems and we proposed that it may reflect the processes of information to failure problems, howerver, there was no hemispheric difference. The ERP results after the effort to the problems showed that unsolved problems elicited a more positive P150 over the right frontal cortex while solved problems demonstrated a left hemispheric advantage of P150. When hints present, P2 amplitudes of hints were modulated by the status of problem only in the right hemisphere but not in the left hemisphere. Our results confirmed the hypothesis that failure to solve problems would trigger the perseverance processes in right hemisphere, which would make right hemisphere more sensitive to related information of failure problems.

From Parallel Sequence Representations to Calligraphic Control: A Conspiracy of Neural Circuits

Calligraphic writing presents a rich set of challenges to the human movement control system. These challenges include: initial learning, and recall from memory, of prescribed stroke sequences; critical timing of stroke onsets and durations; fine control of grip and contact forces; and letter-form invariance under voluntary size scaling, which entails fine control of stroke direction and amplitude during recruitment and derecruitment of musculoskeletal degrees of freedom. Experimental and computational studies in behavioral neuroscience have made rapid progress toward explaining the learning, planning and contTOl exercised in tasks that share features with calligraphic writing and drawing. This article summarizes computational neuroscience models and related neurobiological data that reveal critical operations spanning from parallel sequence representations to fine force control. Part one addresses stroke sequencing. It treats competitive queuing (CQ) models of sequence representation, performance, learning, and recall. Part two addresses letter size scaling and motor equivalence. It treats cursive handwriting models together with models in which sensory-motor tmnsformations are performed by circuits that learn inverse differential kinematic mappings. Part three addresses fine-grained control of timing and transient forces, by treating circuit models that learn to solve inverse dynamics problems.

Laminar Cortical Dynamics of Cognitive and Motor Working Memory, Sequence Learning and Performance: Toward a Unified Theory of How the Cerebral Cortex Works

How do the layered circuits of prefrontal and motor cortex carry out working memory storage, sequence learning, and voluntary sequential item selection and performance? A neural model called LIST PARSE is presented to explain and quantitatively simulate cognitive data about both immediate serial recall and free recall, including bowing of the serial position performance curves, error-type distributions, temporal limitations upon recall, and list length effects. The model also qualitatively explains cognitive effects related to attentional modulation, temporal grouping, variable presentation rates, phonemic similarity, presentation of non-words, word frequency/item familiarity and list strength, distracters and modality effects. In addition, the model quantitatively simulates neurophysiological data from the macaque prefrontal cortex obtained during sequential sensory-motor imitation and planned performance. The article further develops a theory concerning how the cerebral cortex works by showing how variations of the laminar circuits that have previously clarified how the visual cortex sees can also support cognitive processing of sequentially organized behaviors.

Understanding and production of prosody in children with Williams syndrome: A developmental trajectory approach

This study investigated the development of three aspects of linguistic prosody in a group of children with Williams syndrome compared to typically developing children. The prosodic abilities investigated were: (1) the ability to understand and use prosody to make specific words or syllables stand out in an utterance (focus); (2) the ability to understand and use prosody to disambiguate complex noun phrases (chunking); (3) the ability to understand and use prosody to regulate conversational behaviour (turn-end). The data were analysed using a cross-sectional developmental trajectory approach. The results showed that, relative to chronological age, there was a delayed onset in the development of the ability of children with WS to use prosody to signal the most important word in an utterance (the focus function). Delayed rate of development was found for all the other aspects of expressive and receptive prosody under investigation. However, when non-verbal mental age was taken into consideration, there were no differences between the children with WS and the controls neither with the onset nor with the rate of development for any of the prosodic skills under investigation apart from the ability to use prosody in order to regulate conversational behaviour. We conclude that prosody is not a ‘preserved’ cognitive skill in WS. The genetic factors, development in other cognitive domains and environmental influences affect developmental pathways and as a result, development proceeds along an atypical trajectory.

Hierarchical semi-markov conditional random fields for recursive sequential data

Inspired by the hierarchical hidden Markov models (HHMM), we present the hierarchical semi-Markov conditional random field (HSCRF), a generalisation of embedded undirected Markov chains to model complex hierarchical, nested Markov processes. It is parameterised in a discriminative framework and has polynomial time algorithms for learning and inference. Importantly, we develop efficient algorithms for learning and constrained inference in a partially-supervised setting, which is important issue in practice where labels can only be obtained sparsely. We demonstrate the HSCRF in two applications: (i) recognising human activities of daily living (ADLs) from indoor surveillance cameras, and (ii) noun-phrase chunking. We show that the HSCRF is capable of learning rich hierarchical models with reasonable accuracy in both fully and partially observed data cases.

La memoria di lavoro nell'interpretazione simultanea

Riassunto Tesi Nella presente dissertazione viene esposto un esperimento condotto con la partecipazione di 13 studenti del Corso Magistrale di Interpretazione di Conferenza della Scuola di Lingue e Letterature, Traduzione e Interpretazione di Forlì al fine di comprendere se la memoria di lavoro possa essere potenziata in seguito ad un periodo di 7 mesi nel quale gli studenti si sono costantemente confrontati con lezioni ed esercizi di memorizzazione, interpretazione simultanea e interpretazione consecutiva. Nel primo capitolo, dopo una breve introduzione agli studi riguardanti la memoria, vengono descritti i diversi tipi di quest'ultima che congiuntamente alla memoria di lavoro, vengono attivati durante l'interpretazione simultanea. Nel secondo capitolo la memoria di lavoro viene messa in relazione con l'interpretazione simultanea. Oltre alle teorie più influenti circa i meccanismi cognitivi e le singole attività che sottendono il processo dell’IS, vengono esposti i principali modelli di elaborazione dell'informazione durante l'attività in questione. In chiusura del secondo capitolo, vengono esposti brevemente altri aspetti cruciali per l'interpretazione simultanea, utili al fine di comprendere le teorie sviluppate e gli studi condotti in merito: attenzione, chunking, soppressione articolatoria e ripetizione subvocalica, accennando inoltre ad un'interessante ricerca condotta per monitorare l'attivazione e il mutamento delle diverse aree cerebrali durante l'utilizzo di una lingua straniera. Nel terzo capitolo vengono passati in rassegna i principali studi empirici condotti sull'interpretazione simultanea per comprendere il comportamento della memoria di lavoro. Infine, nel quarto capitolo viene presentato l'esperimento condotto allo scopo di verificare l'ipotesi alla base della presente tesi: un aumento dell'esperienza in IS può condurre al potenziamento ML. Dopo aver esposto lo scopo, il metodo e il materiale vengono presentati i risultati che, sebbene siano stati raccolti su un campione abbastanza esiguo, vanno a confermare l'ipotesi all'origine della ricerca.

Beyond the aspect hypothesis:tense-aspect development in advanced L2 French

The Aspect Hypothesis (AH) claims that the association of any verb category (lexical aspect) with any grammatical aspect (perfective or imperfective) constitutes the endpoint of acquisition. The present book evaluates the explanatory power of the Aspect Hypothesis for the acquisition of French past tenses, which constitutes a serious stumbling block for foreign learners, even at the highest levels of proficiency. The present research applies the Aspect Hypothesis to the production of 61 Anglophone 'advanced learners' in a tutored environment. In so doing, it tests concurrent explanations, including the influence of the input, the influence of chunking, and the hypothesis of cyclic development. It discusses the cotextual and contextual factors that still provoke «non-native glitches» at the final stage of the Aspect Hypothesis. The book shows that the AH fails to account for the complex phenomenon of past tense development, as it adopts a local and linear approach.