Search

Synthetic aperture imaging, which has been proved to be an effective approach for occluded object imaging, is one of the challenging problems in the field of computational imaging. Currently most of the related researches focus on fixed synthetic aperture which usually accompanies with mixed observation angle and foreground de-focus blur. But the existence of them is frequently a source of perspective effect decrease and occluded object imaging quality degradation. In order to solve this problem, we propose a novel data-driven variable synthetic aperture imaging based on semantic feedback. The semantic content we concerned for better de-occluded imaging is the foreground occlusions rather than the whole scene. Therefore, unlike other methods worked on pixel-level, we start from semantic layer and present a semantic labeling method based on feedback. Semantic labeling map deeply mines visual data in synthetic image and preserves the semantic information of foreground occluder. On the basis of semantic feedback strategy, semantic labeling map will conversely pass to synthetic imaging process. The proposed data-driven variable synthetic aperture imaging contains two levels: one is adaptive changeable imaging aperture driven by synthetic depth and perspective angle, the other is light ray screening driven by visual information in semantic labeling map. On this basis, the hybrid camera view and superimposition of foreground occlusion can be removed. Evaluations on several complex indoor scenes and real outdoor environments demonstrate the superiority and robustness performance of our proposed approach.

Studies on teaching of reading strategies have found that summarizing is of tremendous help to reading comprehension. However grading students’ summary writings is laborious, but given the importance of summarizing, an effective summarizing learning module is important. This study developed an automatic summary assessment and feedback system based on Latent Semantic Analysis (LSA) to provide score, concept and semantic feedback, and then investigated the effects of concept and semantic feedback on the writing of summaries by students in the sixth grade. The design involved two between-subject factors: semantic feedback (with, without) and concept feedback (with, without). 120 sixth-grade students from an elementary school were recruited for the study, and then were randomly assigned to each group. The overall results demonstrated the effectiveness of the proposed system in improving the summary writing skills of students. The effects of semantic feedback and concept feedback were also discussed.

Automated essay evaluation (AEE) is a widely used practical solution for replacing time-consuming manual grading of student essays. Automated systems are used in combination with human graders in different high-stake assessments, as well as in classrooms. During the last 50 years, since the beginning of the development of the field, many challenges have arisen in the field, including seeking ways to evaluate the semantic content, providing automated feedback, determining reliability of grades, making the field more "exposed", and others. In this dissertation we address several of these challenges and propose novel solutions for semantic based essay evaluation. Most of the AEE research has been conducted by commercial organizations that protect their investments by releasing proprietary systems where details are not publicly available. We provide comparison (as detailed as possible) of 20 state-of-the-art approaches for automated essay evaluation and we propose a new automated essay evaluation system named SAGE (Semantic Automated Grader for Essays) with all the technological details revealed to the scientific community. Lack of consideration of text semantics is one of the main weaknesses of the existing state-of-the-art systems. We address the evaluation of essay semantics from perspectives of essay coherence and semantic error detection. Coherence describes the flow of information in an essay and allows us to evaluate the connections between the discourse. We propose two groups of coherence attributes: coherence attributes obtained in a highly dimensional semantic space and coherence attributes obtained from a sentence-similarity networks. Furthermore, we propose the Automated Error Detection (AED) system and evaluate the essay semantics from the perspective of essay consistency. The system detects semantic errors using information extraction and logic reasoning and is able to provide semantic feedback for the writer. The proposed system SAGE achieves significantly higher grading accuracy compared with other state-of-the-art automated essay evaluation systems. In the last part of the dissertation we address the question of reliability of grades. Despite the unified grading rules, human graders introduce bias into scores. Consequently, a grading model has to implement a grading logic that may be a mixture of grading logics from various graders. We propose an approach for separating a set of essays into subsets that represent different graders, which uses an explanation methodology and clustering. The results show that learning from the ensemble of separated models significantly improves the average prediction accuracy on artificial and real-world datasets. Avtomatsko ocenjevanje esejev predstavlja praktično rešitev za številne težave, povezane s časovno zahtevnim ročnim ocenjevanjem. Avtomatizirani sistemi se uporabljajo v kombinaciji s človeškimi ocenjevalci pri številnih standardiziranih testih, vse več pa tudi v učilnicah. V zadnjih 50 letih, od začetka razvoja področja, so se pojavili številni izzivi, vključno z iskanjem pristopov za ocenjevanje semantične vsebine, zagotavljanjem avtomatskih povratnih informacij, določanjem zanesljivosti ocen, težnjo po dostopnosti podrobnosti delovanja sistemov in s tem odprtosti področja, in drugi. V pričujoči disertaciji obravnavamo te izzive in predlagamo nove rešitve za semantično usmerjeno avtomatsko ocenjevanje esejev. Eden od glavnih problemov sistemov za avtomatsko ocenjevanje esejev je problem ocenjevanja semantične pravilnosti besedila. V disertaciji obravnavamo ocenjevanje semantike besedila z različnimi pristopi: ocenjevanje koherence esejev in zaznavanje semantičnih napak. Koherenca opisuje pretok informacij v eseju in nam omogoča, da ocenimo povezanost besedila. Predlagamo dve skupini atributov za ocenjevanje koherence: atributi, pridobljeni v visoko dimenzionalnem semantičnem prostoru, in atributi, pridobljeni iz omrežij stavčne podobnosti. Poleg tega predlagamo sistem za avtomatsko odkrivanje napak, ki nam pomaga oceniti semantiko eseja z vidika doslednosti. Sistem zaznava semantične napake z uporabo ekstrakcije informacij in logičnega sklepanja ter zagotavlja povratno semantično informacijo. Predlagani sistem SAGE (Semantic Automated Grader for Essays) dosega višjo napovedno točnost v primerjavi z drugimi sodobnimi sistemi za avtomatsko ocenjevanje esejev. V zadnjem delu disertacije se posvečamo vprašanju zanesljivosti ocen. Kljub poenotenim kriterijem za človeške ocenjevalce, ocenjevalci vnašajo pristranskost v rezultate. Zato mora napovedni model uporabiti napovedno logiko, ki je lahko mešanica ocenjevalne logike različnih ocenjevalcev. Predlagamo pristop za ločevanje množice esejev v podmnožice, ki predstavljajo različne ocejevalce, kjer uporabimo metodologijo razlage napovedi in gručenje. Rezultati kažejo, da učenje na ansamblu ločenih modelov bistveno izboljša povprečno točnost napovedi na umetnih in realnih podatkovnih množicah.

Automation of security inspections is crucial for improving the efficiency and reducing security risks. In this paper, we focus on automatically recognizing and localizing prohibited items in airport X-ray security images. A top-down attention mechanism is applied to enhance a CNN classifier to additionally locate the prohibited items. We introduce a high-level semantic feedback loop to map the targets semantic signal to the input X-ray image space for generating task-specic attention maps. And the attention maps indicate the location and general outline of prohibited items in the input images. Furthermore, to obtain more accurate location information, we combine the lateral inhibition and contrastive attention to suppress noise and non-target interference in attention maps. The experiments on the GDX-ray image dataset have demonstrated the efficiency and stability of the proposed scheme in both single target detection and multi-target detection.

Previous research has reported an imageability effect on visual word recognition. Words that are high in imageability are recognised more rapidly than are those lower in imageability. However, later researchers argued that imageability was confounded with age of acquisition. In the current research, these two factors were manipulated in a factorial design to assess their effect in a lexical decision task and a progressive demasking task. Across both tasks, there was a clear and robust effect of age of acquisition. In contrast, the imageability effect was only evident in the progressive demasking task. Both effects are explained within the connectionist framework in terms of network plasticity and semantic feedback activation.

We design and study ExamParser, an innovative intelligent semantic automatic indexing method, for orchestrating today's programming classes. ExamParser automatically processes paper-based exams by associating sets of concepts to the exam questions, which provide graders semantic grading guidelines and leave personalized semantic feedback. Results showed that the ExamPraser significantly extract more and diverse concepts from exams. It also achieves high coherence within exam, indicating the automatic concept extraction from exams is promising and could be a potential technological solution to provide personalized feedback for large-size programming classes.

Background: Few studies have addressed rehabilitation in semantic dementia. A potentially promising method is formal‐semantic therapy, which consists of tasks in which the names of concepts and their semantic characteristics are presented. It could also be enhanced by spaced retrieval, a learning method improving retention through recalling information after increasing recall intervals. Aims: This study explores the efficacy of both a formal‐semantic therapy and the spaced retrieval method to restore lost concepts in TBo, a woman with semantic dementia. Methods & Procedures: The formal‐semantic therapy consisted of giving TBo semantic feedback followed by a cueing technique to facilitate naming. Formal‐semantic therapy with simple repetition was compared to formal‐semantic therapy with spaced retrieval. TBo's performance was measured throughout the study with picture naming and generation of verbal attributes. Two untrained lists were also measured for generalisation effects. Outcomes & Results: Results indicate that, after therapy, TBo could name 3/8 of the trained items, compared to no items on the untrained lists. She also showed an increase in performance for the evocation of specific semantic attributes of concepts, reaching 6/8 of correct responses. Moreover, she maintained her performance up to 5 weeks after the end of the study. Finally, when compared to simple repeated practice, spaced retrieval did not enhance learning and no generalisation was observed between trained and non‐trained categories. Conclusions: Along with recent results reported in the literature, TBo's results confirm that people with semantic dementia can improve their naming performance with training but that this is limited. However, formal‐semantic therapy seems very promising for retraining specific semantic attributes. Instead of focusing on naming, we suggest that therapies used in semantic dementia should aim at restoring specific and functionally relevant concepts to enable the individuals to be more autonomous in daily living.

In this paper, we report a study of an innovative mobile application to support grading paper-based programming exams. We call the app – Programming Grading Assistant (PGA). It scans pre-generated QR-codes of paper-based question-and-concepts associations and uses OCR to recognize handwritten answers. PGA provides interfaces for teachers to calibrate recognition results, as well as to adjust partial credit assignment according to conceptual incorrectness of the answers. We evaluate the mobile grading process and the quality of grading results based on the assessed semantic information. The results demonstrate that the mobile grading approach keeps persistent traces of students’ performance, including semantic feedback and ultimately enhances grading consistency.

Pedagogical algorithm visualization (AV) systems produce graphical representations that aim to assist learners in understanding the dynamic behavior of computer algorithms. In order to foster active learning, computer science educators have developed AV systems that empower learners to construct their own visualizations of algorithms under study. Notably, these systems support a similar development model in which coding an algorithm is temporally distinct from viewing and interacting with the resulting visualization. Given that they are known to have problems both with formulating syntactically correct code, and with understanding how code executes, novice learners would appear likely to benefit from a more ''live'' development model that narrows the gap between coding an algorithm and viewing its visualization. In order to explore this possibility, we have implemented ''What You See Is What You Code,'' an algorithm development and visualization model geared toward novices first learning to program under the imperative paradigm. In the model, the line of algorithm code currently being edited is reevaluated on every edit, leading to immediate syntactic feedback, along with immediate semantic feedback in the form of an AV. Analysis of usability and field studies involving introductory computer science students suggests that the immediacy of the model's feedback can help novices to quickly identify and correct programming errors, and ultimately to develop semantically correct code.

The present study used a mediated priming paradigm to examine whether developmental differences exist in the integration of semantic information with orthographic and phonological information during visual word recognition. In Experiment 1, we found that the integration of semantics with phonology and orthography differed among third-grade, sixth-grade and college students: orthographically based mediated inhibition effects were found in third-grade children, whereas phonologically based mediated inhibition effects were found in sixth-grade children and college students. A second experiment was performed with adults to test the hypothesis that the orthographically based mediated inhibition effect observed with young children was due to deficits in orthographic processing. When stimulus quality was manipulated within the mediated priming paradigm, orthographically based mediated inhibition effects were found when targets were dim, whereas phonologically based mediated inhibition effects were found when targets were bright. Taken together, these results suggest that the allocation of activation during reading may depend on the processing demands of the word recognition system.

Information can only be defined as an equivalence class of sequences with respect to alphabet changes and encodings, hence it is an abstract entity. However, a sequence is necessarily borne by a physical support. Information thus appears as a bridge between the abstract and the concrete. In the living world, the information borne by genomes acts on matter by instructing the assembly of objects. Physical perturbations degrade the supports, hence the sequences they bear, but the information they represent can be almost indefinitely conserved provided they are endowed with error-correcting codes as powerful as to enable their regeneration, and regenerated frequently enough. Codes made of nested components which successively originated over the ages account for the conservation of the oldest parts of the genomes. The component codes are ‘soft’, meaning that they result from any kind of non-mathematical constraint, e.g., physical-chemical or linguistic. The genomic information instructs the assembly of living structures by the agency of interwoven semantic feedback loops which lock themselves because the synthetized enzymes catalyse their own assembly.

In this paper, we propose a novel way to perform top-down semantic feedback in convolutional deep networks for efficient and accurate image parsing. We also show how to add global appearance/semantic features, which have shown to improve image parsing performance in state-of-the-art methods, and was not present in previous convolutional approaches. The proposed method is characterised by an efficient training and a sufficiently fast testing. We use the well known SIFTflow dataset to numerically show the advantages provided by our contributions, and to compare with state-of-the-art image parsing convolutional based approaches.

It is argued in Chap. 9 that information is specific to life. The concept of ‘semantic feedback loops’ is first introduced in order to explain the onset of living structures and their conservation. The basic fact is that certain proteins act as enzymes which catalyze the operations which are needed for their own synthesis: transcription of DNA into messenger or premessenger RNA, splicing out the introns (in the case of eukaryotic cells) and translation as operated by the ribosome, i.e., synthesizing proteins under the control of messenger RNA. This can be interpreted as a set of interwoven feedback loops which all involve the genetic ‘code’ (or mapping). This set of loops acts as a trap since, once closed, it keeps its structure. If any of its parts fails, however, the whole system ceases operating. We refer to it as ‘semantic’ because the genetic mapping consists of a set of semantic rules. Hence, the set of feedback loops implements the semantics which enables the assembly of proteins (and more generally of phenotypes) in terms of the information that DNA bears. This shows how information, as an abstract fundamental entity, controls the assembly of physical structures. ‘Abstract’ should be understood here as opposed to physical, since assuming that information is physical leads to results which contradict its very definition. The set of semantic feedback loops, although each loop is closed, does not prevent the lengthening of genomes by horizontal genetic transfer, hence is compatible with evolution. The last section of this chapter examines the achievements of Nature as an engineer and pleads for a collaboration of engineers and biologists.

In models of visual word recognition that incorporate an interactive activation framework, ac- tivation spreads from orthographic and phonological units to semantic units and from seman- tic units back to phonological and orthographic units. The present research examined whether semantic feedback changes over the time course of lexical processing and as a function of stimulus quality. Using a mediated priming paradigm, prime-target word pairs were associa- tively related (frog-toad), homophonically mediated (frog-towed), or orthographically medi- ated (frog-told). Evidence of semantic feedback to both orthography and phonology was found when the prime duration was 146 ms (Experiment 1) and only to phonology when the prime duration was 253 ms (Experiment 2a). However, when the prime duration was 253 ms and target words were degraded (Experiment 2b), feedback spread only to orthography. The results suggest that the dynamics of semantic feedback change as the function of processing demands in the visual word recognition system. Both localist and parallel distributed processing models of visual word recognition typically contain distinct groups or levels of representations (e.g., orthographic, phonological, semantic). As a result, multiple models of visual word recognition (e.g., Col

In this paper we discuss the Recommendz recommender system. This domain-independent system combines the advantages of collaborative and content-based filtering in a novel way. By allowing users to provide feedback not only about an item as a whole, but also properties of an item that motivated their opinion, increased performance seems to be achieved. The features used to describe items are specified by the users of the system rather than predetermined using manual knowledge-engineering. We describe a method for combining descriptive features and simple ratings, and provide a performance analysis.

International audience; Based on previous works on geographical information (GI) modeling and information retrieval, we report on the development of a human-system interface for the AntéServeur Géographique project. The AntéServeur project is a middleware system that implements a framework for retrieving geographical data of heterogeneous nature. After a bref overview of various existing frameworks for GI retrieval and a short descrption of the principles of the bimodal interface, this paper focuses on the spatial model underlying our approach and \\backslashll\semnatic feedback\\backslashgg\ carried out.

This paper presents an open-domain textual Question-Answering system that uses several feedback loops to enhance its performance. These feedback loops combine in a new way statistical results with syntactic, semantic or pragmatic information derived from texts and lexical databases. The paper presents the contribution of each feedback loop to the overall performance of 76% human-assessed precise answers.

THESEUS++ is an object-oriented high level user interface toolkit especially designed to ease development of application-specific interaction with 2 1 2 D graphics. THESEUS++ enforces the separation between the application and user interface in a flexible way. These parts of the dialogue, that don't require application-specific knowledge, can be described on a high level of abstraction, while semantic feedback on a low level is possible for those dialogue cycles needing application knowledge. This is accomplished by the following features: • • A description model for the separate specification of interaction components like prompt, feedback, and semantics on each level of the interaction hierarchy • • Dynamic composition mechanisms both for the definition of complex interactions (dialogue hierarchy) and for complex graphical objects (picture hierarchy) • • Support for constraint-based direct manipulation of application-defined graphics including continuous interactions like dragging and stretching • • Mechanisms to support extensibility by new classes of domain-specific interaction techniques and graphical objects. THESEUS++ has been implemented in object-oriented C++ on top of X Window System and OSF/Motif.

This paper describes the implementation of semantic snapping — an interaction technique that provides semantic feedback at the lexical level while dragging a graphical object on the screen. Like conventional snapping, or gravity fields, semantic snapping includes a geometric component where objects in close proximity are drawn together or “snap” into position. However, semantic snapping goes further by allowing non-geometric (semantic) properties of objects to place additional constraints on snapping. Semantic snapping also provides more complex lexical feedback which reflects potential semantic consequences of a snap. This paper motivates the use of semantic snapping and describes how this technique has been implemented in a window-based toolkit. This implementation works in an adaptive manner to provide the best interactive response in situations where semantic tests are very time consuming and strain the limits of acceptable performance.

Feedback is an important component of assessment in learning environments, because it allows students to know their learning flaws. Feedback information is indeed useful for teachers to design learning contents adapted to the needs of the students. Therefore, the availability of feedback constitutes a new learning opportunity. In this paper we describe an approach based on Semantic Web technologies for generating useful semantic feedback for both teachers and students. © 2010 Elsevier Ltd. All rights reserved.