Level Semantic Information Research Articles

Representation learning: a unified deep learning framework for automatic prostate MR segmentation.

Image representation plays an important role in medical image analysis. The key to the success of different medical image analysis algorithms is heavily dependent on how we represent the input data, namely features used to characterize the input image. In the literature, feature engineering remains as an active research topic, and many novel hand-crafted features are designed such as Haar wavelet, histogram of oriented gradient, and local binary patterns. However, such features are not designed with the guidance of the underlying dataset at hand. To this end, we argue that the most effective features should be designed in a learning based manner, namely representation learning, which can be adapted to different patient datasets at hand. In this paper, we introduce a deep learning framework to achieve this goal. Specifically, a stacked independent subspace analysis (ISA) network is adopted to learn the most effective features in a hierarchical and unsupervised manner. The learnt features are adapted to the dataset at hand and encode high level semantic anatomical information. The proposed method is evaluated on the application of automatic prostate MR segmentation. Experimental results show that significant segmentation accuracy improvement can be achieved by the proposed deep learning method compared to other state-of-the-art segmentation approaches.

Lexical and sublexical semantic preview benefits in Chinese reading.

Semantic processing from parafoveal words is an elusive phenomenon in alphabetic languages, but it has been demonstrated only for a restricted set of noncompound Chinese characters. Using the gaze-contingent boundary paradigm, this experiment examined whether parafoveal lexical and sublexical semantic information was extracted from compound preview characters. Results generalized parafoveal semantic processing to this representative set of Chinese characters and extended the parafoveal processing to radical (sublexical) level semantic information extraction. Implications for notions of parafoveal information extraction during Chinese reading are discussed.

LeakChaser

In large programs written in managed languages such as Java and C#, holding unnecessary references often results in memory leaks and bloat, degrading significantly their run-time performance and scalability. Despite the existence of many leak detectors for such languages, these detectors often target low-level objects; as a result, their reports contain many false warnings and lack sufficient semantic information to help diagnose problems. This paper introduces a specification-based technique called LeakChaser that can not only capture precisely the unnecessary references leading to leaks, but also explain, with high-level semantics, why these references become unnecessary. At the heart of LeakChaser is a three-tier approach that uses varying levels of abstraction to assist programmers with different skill levels and code familiarity to find leaks. At the highest tier of the approach, the programmer only needs to specify the boundaries of coarse-grained activities, referred to as transactions. The tool automatically infers liveness properties of these transactions, by monitoring the execution, in order to find unnecessary references. Diagnosis at this tier can be performed by any programmer after inspecting the APIs and basic modules of a program, without understanding of the detailed implementation of these APIs. At the middle tier, the programmer can introduce application-specific semantic information by specifying properties for the transactions. At the lowest tier of the approach is a liveness checker that does not rely on higher-level semantic information, but rather allows a programmer to assert lifetime relationships for pairs of objects. This task could only be performed by skillful programmers who have a clear understanding of data structures and algorithms in the program. We have implemented LeakChaser in Jikes RVM and used it to help us diagnose several real-world leaks. The implementation incurs a reasonable overhead for debugging and tuning. Our case studies indicate that the implementation is powerful in guiding programmers with varying code familiarity to find the root causes of several memory leaks---even someone who had not studied a leaking program can quickly find the cause after using LeakChaser's iterative process that infers and checks properties with different levels of semantic information.

A Machine Learning Based Analytical Framework for Semantic Annotation Requirements

The Semantic Web is an extension of the current web in which information is given well-defined meaning. The perspective of Semantic Web is to promote the quality and intelligence of the current web by changing its contents into machine understandable form. Therefore, semantic level information is one of the cornerstones of the Semantic Web. The process of adding semantic metadata to web resources is called Semantic Annotation. There are many obstacles against the Semantic Annotation, such as multilinguality, scalability, and issues which are related to diversity and inconsistency in content of different web pages. Due to the wide range of domains and the dynamic environments that the Semantic Annotation systems must be performed on, the problem of automating annotation process is one of the significant challenges in this domain. To overcome this problem, different machine learning approaches such as supervised learning, unsupervised learning and more recent ones like, semi-supervised learning and active learning have been utilized. In this paper we present an inclusive layered classification of Semantic Annotation challenges and discuss the most important issues in this field. Also, we review and analyze machine learning applications for solving semantic annotation problems. For this goal, the article tries to closely study and categorize related researches for better understanding and to reach a framework that can map machine learning techniques into the Semantic Annotation challenges and requirements.

The effect of semantic information on saccade trajectory deviations

In recent years, many studies have explored the conditions in which irrelevant visual distractors affect saccades trajectories. These previous studies mainly focused on the low-level stimulus characteristics and how they affect the magnitude of curvature. The present study explored the possible effect of high level semantic information on saccade curvature. Semantic saliency was manipulated by presenting irrelevant peripheral taboo versus neutral cue words in a spatial cuing paradigm that allowed for the measurement of trajectory deviations. Findings showed larger saccade trajectory deviations away from taboo (versus neutral) cue words when making a saccade towards another location. This indicates that due to their high semantic saliency, more inhibition was necessarily applied to taboo cue locations to effectively suppress their competing as saccade targets.

Audiovisual Information Fusion in Human–Computer Interfaces and Intelligent Environments: A Survey

Microphones and cameras have been extensively used to observe and detect human activity and to facilitate natural modes of interaction between humans and intelligent systems. Human brain processes the audio and video modalities, extracting complementary and robust information from them. Intelligent systems with audiovisual sensors should be capable of achieving similar goals. The audiovisual information fusion strategy is a key component in designing such systems. In this paper, we exclusively survey the fusion techniques used in various audiovisual information fusion tasks. The fusion strategy used tends to depend mainly on the model, probabilistic or otherwise, used in the particular task to process sensory information to obtain higher level semantic information. The models themselves are task oriented. In this paper, we describe the fusion strategies and the corresponding models used in audiovisual tasks such as speech recognition, tracking, biometrics, affective state recognition, and meeting scene analysis. We also review the challenges and existing solutions and also unresolved or partially resolved issues in these fields. Specifically, we discuss established and upcoming work in hierarchical fusion strategies and cross-modal learning techniques, identifying these as critical areas of research in the future development of intelligent systems.

Enabling cognitive manufacturing through automated on-machine measurement planning and feedback

With the manufacturing industry shifting from seller markets to buyer markets, cognitive capabilities have become enablers for high flexibility and changeability that facilitate efficient production for the fast changing markets. Process planning in such a manufacturing system needs to be adaptive to machine tools and manufacturing environments. Real-time acquisition of production and process data and feedback to process control is essential. However, the current information flow between manufacturing processes is segmented due to the lack of a consolidated data model to represent information. This paper presents a consolidated data model and system structure for an automatic and integrated process planning for machining, inspection, and feedback. The data model is based on the newly developed ISO standard STEP/STEP-NC, which provides semantic level information for the manufacturing system. The system structure and algorithms developed for the system are described first. A software prototype was developed and used to test the proposed data model and the integrated process planning and feedback system. The system provides important cognitive capability, such as automatic on-line monitoring, automated reasoning and adjustment to unforeseen changes during machining processes, for a manufacturing system. This facilitates the realization of cognitive manufacturing in the near future.

An exploratory study of linguistic–colour associations across languages in multilingual synaesthetes

In one of the most common forms of synaesthesia, linguistic-colour synaesthesia, colour is induced by stimuli such as numbers, letters, days of the week, and months of the year. It is not clear, however, whether linguistic-colour synaesthesia is determined more by higher level semantic information--that is, word meaning--or by lower level grapheme or phoneme structure. To explore this issue, we tested whether colour is consistently induced by grapheme or phoneme form or word meaning in bilingual and trilingual linguistic-colour synaesthetes. We reasoned that if the induced colour was related to word meaning, rather than to the acoustic or visual properties of the words, then the induced colours would remain consistent across languages. We found that colours were not consistently related to word meaning across languages. Instead, induced colours were more related to form properties of the word across languages, particularly visual structure. However, the type of inducing stimulus influenced specific colour associations. For example, colours to months of the year were more consistent across languages than were colours to numbers or days of the week. Furthermore, the effect of inducing stimuli was also associated with the age of acquisition of additional languages. Our findings are discussed with reference to a critical period in language acquisition on synaesthesia.

What do we perceive in a glance of a real-world scene?

What do we see when we glance at a natural scene and how does it change as the glance becomes longer? We asked naive subjects to report in a free-form format what they saw when looking at briefly presented real-life photographs. Our subjects received no specific information as to the content of each stimulus. Thus, our paradigm differs from previous studies where subjects were cued before a picture was presented and/or were probed with multiple-choice questions. In the first stage, 90 novel grayscale photographs were foveally shown to a group of 22 native-English-speaking subjects. The presentation time was chosen at random from a set of seven possible times (from 27 to 500 ms). A perceptual mask followed each photograph immediately. After each presentation, subjects reported what they had just seen as completely and truthfully as possible. In the second stage, another group of naive individuals was instructed to score each of the descriptions produced by the subjects in the first stage. Individual scores were assigned to more than a hundred different attributes. We show that within a single glance, much object- and scene-level information is perceived by human subjects. The richness of our perception, though, seems asymmetrical. Subjects tend to have a propensity toward perceiving natural scenes as being outdoor rather than indoor. The reporting of sensory- or feature-level information of a scene (such as shading and shape) consistently precedes the reporting of the semantic-level information. But once subjects recognize more semantic-level components of a scene, there is little evidence suggesting any bias toward either scene-level or object-level recognition.

Grammatical category disambiguation by statistical optimization

Several algorithms have been developed in the past that attempt to resolve categorial ambiguities in natural language text without recourse to syntactic or semantic level information. An innovative method (called CLAWS) was recently developed by those working with the Lancaster-Oslo/Bergen Corpus of British English. This algorithm uses a systematic calculation based upon the probabilities of co-occurrence of particular tags. Its accuracy is high, but it is very slow, and it has been manually augmented in a number of ways. The effects upon accuracy of this manual augmentation are not individually known.The current paper presents an algorithm for disambiguation that is similar to CLAWS but that operates in linear rather than in exponential time and space, and which minimizes the unsystematic augments. Tests of the algorithm using the million words of the Brown Standard Corpus of English are reported; the overall accuracy is 96%. This algorithm can provide a fast and accurate front end to any parsing or natural language processing system for English.