Search

Emotion recognition has been extensively studied in a single modality in the last decade. However, humans express their emotions usually through multiple modalities like voice, facial expressions, or text. This paper proposes a new method to learn a joint emotion representation for multimodal emotion recognition. Emotion-based feature for speech audio is learned by an unsupervised triplet-loss objective, and a text-to-text transformer network is used to extract text embedding for latent emotional meaning. Transfer learning provides a powerful and reusable technique to help fine-tune emotion recognition models trained on mega audio and text datasets respectively. The extracted emotional information from speech audio and text embedding are processed by dedicated transformer networks. The alternating co-attention mechanism is used to construct a deep transformer network. Multimodal fusion is implemented by a deep co-attention transformer network. Experimental results show the proposed method for learning a joint emotion representation achieves good performance in multimodal emotion recognition.

Emotion recognition has been extensively studied in a single modality in the last decade. However, humans express their emotions usually through multiple modalities like voice, facial expressions, or text. In this paper, we propose a new method to find a unified emotion representation for multimodal emotion recognition through speech audio, and text. Emotion-based feature representation from speech audio is learned by an unsupervised triplet-loss objective, and a text-to-text transformer network is constructed to extract latent emotional meaning. As deep neural network models trained by huge datasets exhaust a lot of unaffordable resources, transfer learning provides a powerful and reusable technique to help fine-tune emotion recognition models trained on mega audio and text datasets respectively. Automatic multimodal fusion of emotion-based features from speech audio and text is conducted by a new transformer. Both the accuracy and robustness of proposed method are evaluated, and we show that our method for multimodal fusion using transfer learning in emotion recognition achieves good results.

Organizations need to constantly learn, develop, and evaluate new strategies and policies for their effective operation. Unsupervised reinforcement learning is becoming a highly useful tool, since rewards and punishments in different forms are pervasive and present in a wide variety of decision-making scenarios. By observing the outcome of a sufficient number of repeated trials, one would gradually learn the value and usefulness of a particular policy or strategy. However, in a given environment, the outcomes resulting from different trials are subject to external chance influence and variations. In learning about the usefulness of a given policy, significant costs are involved in systematically undertaking the sequential trials; therefore, in most learning episodes, one would wish to keep the cost within bounds by adopting learning efficient stopping rules. In this Chapter, we explain the deployment of different learning strategies in given environments for reinforcement learning policy evaluation and review, and we present suggestions for their practical use and applications.

Classification situations are ubiquitous and arise in a range of practical applications. Very often classification problems do not occur in isolation but in clusters where several classification problems need to be solved. Examples of these include employee contract renewal in large organizations, patient prognosis in hospitals, and crowdsourced problem solving. With any form of classification, however, there are invariably inaccuracies occurring in different forms, especially when multiple classification tasks need to be performed. Since typical classifiers are not free from errors, classification errors tend to accumulate, and having frequent misclassifications are often unacceptable. Moreover, in unsupervised learning situations, there are typically no pre-determined ground truth classes; in such a situation the ground truth class is determined by the view of the majority of classifiers. In this paper, we study the situation of multiple classification within the Naive Bayes framework, where the ground truth is determined by the decision of most classifiers, and where there is finite resources requiring decisions to be made within limited budget. Here, we represent the classification tasks as a random walk process, and perform a stochastic analysis of the situation and predict the occurrence of erroneous classification. We obtained closed-form solutions for the probability of classification inaccuracies and derive useful error bounds as a function of the budget. We find that by raising the budget, the probability of error in classification can be controlled, and the extent of the reduction can be quantified. Experimental measurements are also performed which show good agreement with the theoretical results.

In the stylometric analysis of literary works, there are primarily specific aims, for example, at resolving the authorship, the period of the work, style, motif and purpose. These tasks, in fact, can be placed in a machine learning framework, where the quantitative features of known passages are initially learned and analyzed, and then compared with the corresponding metrics in disputable or doubtful passages. While in the past, relative laborious manual approaches relying on the human judgment of selected literary experts are mainly employed, they have shown to error-prone since human experts are unable to systematically process and methodically integrate the varied features of vast volumes of literary data. By combining a big data approach with the application of sophisticated stochastic models, more efficient, dependable and defensible decisions can be made. In this paper, we introduce two models for the quantitative analysis of the characteristics of literary work. The first is a Bernoulli model, which assumes the underlying quantitative features that are independent of each other, and the second is a Markov model, which is particularly appropriate for the analysis of poetic styles. We derive probabilistic measures of passage characteristics, and provide their illustrations in literary works in the English language. Experiments are performed the results of which corroborate the approach.

Bernoulli bandits have found to mirror many practical situations in the context of reinforcement learning, and the aim is to maximize rewards through playing the machine over a set time frame. In an actual casino setting, it is often unrealistic to fix the time when playing stops, as the termination of play may be random and dependent on the outcomes of earlier lever pulls, which in turn affects the inclination of the gambler to continue playing. It is often assumed that exploration is repeated each time the game is played, and that the game tend to go on indefinitely. In practical situations, if the casino does not change their machines often, exploration need not be carried out repeatedly as this would be inefficient. Moreover, from the gamblers' point of view, they would likely to stop at some point or when certain conditions are fulfilled. Here, the bandit problem is studied in terms of stopping rules which are dependent on earlier random outcomes and on the behavior of the players. Rewards incorporating the cost of play and the size of payouts are then calculated on the conclusion of a playing episode. Here, the rewards for Bernoulli machines are placed within the context of martingales that are commonly used in gambling situations, and the fairness of the game is expressed through the parameters of the bandit machines which can be manifested as various forms of martingales. The average rewards and regrets as well as episode durations are obtained under different martingale stopping times. Exploration costs and regrets for different bandit machines are analyzed. Experimentation has also been undertaken which corroborate the theoretical results.

Human behaviors and the emotional states that they convey have been studied by psychologist and sociologists. The tracking of behaviors and emotions is becoming more pervasive with the advent of the Internet of Things (IoT), where small and always connected sensors can continuously capture information about human gestures, movements and postures. The captured information about readable behaviors conveys significant information that can be represented as time series. Few studies in emotion recognition and affective computing have explored the connection between the time series sensors data and the emotional behavior they conveys. In this paper, an innovative approach is proposed to study the emotions and behaviors connected to the time series data. A convolutional network augmented with attention-based bidirectional LSTM is introduced to represent the correlations between behaviors and emotions. The advantage of this model is that it can well recognized emotions by exploiting the data captured by sensors. The experimental results show that the proposed deep learning method outperforms separate schemes and achieves a high degree of accuracy for modelling human behaviors and emotions.

In the search and retrieval of multimedia objects, it is impractical to either manually or automatically extract the contents for indexing since most of the multimedia contents are not machine extractable, while manual extraction tends to be highly laborious and time-consuming. However, by systematically capturing and analyzing the feedback patterns of human users, vital information concerning the multimedia contents can be harvested for effective indexing and subsequent search. By learning from the human judgment and mental evaluation of users, effective search indices can be gradually developed and built up, and subsequently be exploited to find the most relevant multimedia objects. To avoid hovering around a local maximum, we apply the epsilon-greedy method to systematically explore the search space. Through such methodic exploration, we show that the proposed approach is able to guarantee that the most relevant objects can always be discovered, even though initially it may have been overlooked or not regarded as relevant. The search behavior of the present approach is quantitatively analyzed, and closed-form expressions are obtained for the performance of two variants of the epsilon-greedy algorithm, namely EGSE-A and EGSE-B. Simulations and experiments on real data set have been performed which show good agreement with the theoretical findings. The present method is able to leverage exploration in an effective way to significantly raise the performance of multimedia information search, and enables the certain discovery of relevant objects which may be otherwise undiscoverable., 8 pages, 10 figures. IEEE ICMLA 2019

In reinforcement learning episodes, the rewards and punishments are often non-deterministic, and there are invariably stochastic elements governing the underlying situation. Such stochastic elements are often numerous and cannot be known in advance, and they have a tendency to obscure the underlying rewards and punishments patterns. Indeed, if stochastic elements were absent, the same outcome would occur every time and the learning problems involved could be greatly simplified. In addition, in most practical situations, the cost of an observation to receive either a reward or punishment can be significant, and one would wish to arrive at the correct learning conclusion by incurring minimum cost. In this paper, we present a stochastic approach to reinforcement learning which explicitly models the variability present in the learning environment and the cost of observation. Criteria and rules for learning success are quantitatively analyzed, and probabilities of exceeding the observation cost bounds are also obtained., AIKE 2018

In reinforcement learning, a decision needs to be made at some point as to whether it is worthwhile to carry on with the learning process or to terminate it. In many such situations, stochastic elements are often present which govern the occurrence of rewards, with the sequential occurrences of positive rewards randomly interleaved with negative rewards. For most practical learners, the learning is considered useful if the number of positive rewards always exceeds the negative ones. A situation that often calls for learning termination is when the number of negative rewards exceeds the number of positive rewards. However, while this seems reasonable, the error of premature termination, whereby termination is enacted along with the conclusion of learning failure despite the positive rewards eventually far outnumber the negative ones, can be significant. In this paper, using combinatorial analysis we study the error probability in wrongly terminating a reinforcement learning activity which undermines the effectiveness of an optimal policy, and we show that the resultant error can be quite high. Whilst we demonstrate mathematically that such errors can never be eliminated, we propose some practical mechanisms that can effectively reduce such errors. Simulation experiments have been carried out, the results of which are in close agreement with our theoretical findings., Short Paper in AIKE 2018

Rewards and punishments in different forms are pervasive and present in a wide variety of decision-making scenarios. By observing the outcome of a sufficient number of repeated trials, one would gradually learn the value and usefulness of a particular policy or strategy. However, in a given environment, the outcomes resulting from different trials are subject to chance influence and variations. In learning about the usefulness of a given policy, significant costs are involved in systematically undertaking the sequential trials; therefore, in most learning episodes, one would wish to keep the cost within bounds by adopting learning stopping rules. In this paper, we examine the deployment of different stopping strategies in given learning environments which vary from highly stringent for mission critical operations to highly tolerant for non-mission critical operations, and emphasis is placed on the former with particular application to aviation safety. In policy evaluation, two sequential phases of learning are identified, and we describe the outcomes variations using a probabilistic model, with closed-form expressions obtained for the key measures of performance. Decision rules that map the trial observations to policy choices are also formulated. In addition, simulation experiments are performed, which corroborate the validity of the theoretical results.

Human behaviors and the emotional states that they convey have been studied by psychologist and sociologists. The tracking of behaviors and emotions is becoming more pervasive with the advent of the Internet of Things (IoT), where small and always connected sensors can continuously capture information about human gestures, movements and postures. The captured information about readable behaviors conveys signicant information that can be represented as time series. Few studies in emotion recognition and affective computing have explored the connection between the time series sensors data and the emotional behavior they conveys. In this paper, an innovative approach is proposed to study the emotions and behaviors connected to the sensors time series data. A deep learning attention-based bidirectional LSTM is introduced to represent the correlations between data and emotions. The attention-based mechanism is then used to focus on the most significant information. The advantage of this model is that it can represent various human emotions by exploiting the data captured by sensors. The experimental results show that the proposed deep learning method outperforms shallow models and achieves a high degree of accuracy for modelling human behaviors and emotions.

Clinical Decisions Support Systems have been in practical use for a period of time. However, rate of system acceptance is still relatively low. Normally, developers and users have little communication during system development. Consequently, they fail to develop adequate understanding of the full system requirements and operations. In the Co-Development model, developers and users build and implement the system together through joint effort. In this paper, we use Acute Leukemia as an example for Preliminary Decision Tree Model. In the Co-Development model, we provide a channel for user feedback and amend the decision model for increased accuracy of diagnosis.

With the increasing number of Web images and social photograph sharing sites, effective search of real-world images becomes a formidable challenge and an important necessity. Different indexing and annotation algorithms are required for different types of Web images. To meet the challenge and provide satisfactory search results to users, we present a multi-level method with four levels differentially applied to different types of Web images. The performance of our method is evaluated by experiments on thousands of Web images and tags in different subsets, and our approach is able to yield highly promising results compared with applying a single method to all types of Web images.

Image search engines tend to return a large number of images which the engines consider to be relevant, and such pool of results generally is very large and may be regarded to be effectively inexhaustible. While the images are presented as relevant, it is normally true that many of them are actually irrelevant, and that the distribution of relevant images over the returned results is nonuniform. In this paper, we present time series models, including moving average and exponential smoothing, for predicting Image Search Engines (ISEs) results. We are able to establish models which are able to give good and robust prediction of search performance. In addition, the results of this research can have a direct bearing on search engine design to provide informative guidance to users on the retrieval of relevant images, and allow the users to optimize their strategy in the recovery and discovery of images.

As the number of web images is increasing at a rapid rate, searching them semantically presents a significant challenge. Many raw images are constantly uploaded with little meaningful direct annotations of semantic content, limiting their search and discovery. In this paper, we present a semantic annotation technique based on the use of image parametric dimensions and metadata. Using decision trees and rule induction, we develop a rule-based approach to formulate explicit annotations for images fully automatically, so that by the use of our method, semantic query such as " sunset by the sea in autumn in New York" can be answered and indexed purely by machine. Our system is evaluated quantitatively using more than 100,000 web images. Experimental results indicate that this approach is able to deliver highly competent performance, attaining good recall and precision rates of sometimes over 80%. This approach enables a new degree of semantic richness to be automatically associated with images which previously can only be performed manually.

Topological analysis of intelligent agent networks provides crucial information about the structure of agent distribution over a network. Performance analysis of agent network topologies helps multi-agent system developers to understand the impact of topology on system efficiency and effectiveness. Appropriate topology analysis enables the adoption of suitable frameworks for specific multi-agent systems. In this paper, we systematically classify agent network topologies and propose a novel hybrid topology for distributed multi-agent systems. We compare the performance of this topology with two other common agent network topologies—centralised and decentralised topologies—within a new multi-agent framework, called Agent-based Open Connectivity for DSS (AOCD). Three major aspects are studied for estimating topology performance, which include (i) transmission time for a set of requests; (ii) waiting time for processing requests; and (iii) memory consumption for storing agent information. We also conduct a set of AOCD topological experiments to compare the performance of hybrid and centralised agent network topologies and illustrate our experimental results in this paper.

We present several interrelated technical and empirical contributions to the problem of emotion-based music recommendation and show how they can be applied in a possible usage scenario. The contributions are (1) a new three-dimensional resonance-arousal-valence model for the representation of emotion expressed in music, together with methods for automatically classifying a piece of music in terms of this model, using robust regression methods applied to musical/acoustic features; (2) methods for predicting a listener’s emotional state on the assumption that the emotional state has been determined entirely by a sequence of pieces of music recently listened to, using conditional random fields and taking into account the decay of emotion intensity over time; and (3) a method for selecting a ranked list of pieces of music that match a particular emotional state, using a minimization iteration method. A series of experiments yield information about the validity of our operationalizations of these contributions. Throughout the article, we refer to an illustrative usage scenario in which all of these contributions can be exploited, where it is assumed that (1) a listener’s emotional state is being determined entirely by the music that he or she has been listening to and (2) the listener wants to hear additional music that matches his or her current emotional state. The contributions are intended to be useful in a variety of other scenarios as well.

Queries containing aggregate functions often combine multiple tables through join operations. This query is subsequently called "Groupby-Join". There is a special category of this query whereby the group-by operation can only be performed after the join operation. This is known as "Groupby-After-Join" queries--the focus of this paper. In parallel processing of such queries, it must be decided which attribute is used as a partitioning attribute, particularly join attribute or group-by attribute. Based on the partitioning attribute, two parallel processing methods, namely join partition method (JPM) and aggregate partition method (APM) are discussed. The behaviours of these parallelization methods are described in terms of cost models. Experiments are performed based on simulations. The simulation results show that the aggregate partition method performs better than the join partition method.

Skewness has been one of the major problems not only in parallel relational database systems, but also in parallel object-oriented database systems. To improve performance of object-oriented query processing, a careful and intelligent skew handling for load balancing must be established. Depending on the parallel machine environment, whether it is a shared-memory or a shared-nothing architecture, load balancing can be achieved through "physical" or "logical" data re-distribution. It is not the aim of this paper to propose or to investigate skew handling methods, but rather to analyze the impact of load balancing to query execution scheduling strategies. Our analysis shows that when load balancing is achieved, "serial" execution scheduling is preferable to "parallel" execution scheduling strategy. In other words, allocating full resources to a sub-query seems to be better than dividing resources to multiple sub-queries.

Recent findings show that online reviews, blogs, and discussion forums on chronic diseases and drugs are becoming important supporting resources for patients. Extracting information from these substantial bodies of texts is useful and challenging. We developed a generative probabilistic aspect mining model (PAMM) for identifying the aspects/topics relating to class labels or categorical meta-information of a corpus. Unlike many other unsupervised approaches or supervised approaches, PAMM has a unique feature in that it focuses on finding aspects relating to one class only rather than finding aspects for all classes simultaneously in each execution. This reduces the chance of having aspects formed from mixing concepts of different classes; hence the identified aspects are easier to be interpreted by people. The aspects found also have the property that they are class distinguishing: They can be used to distinguish a class from other classes. An efficient EM-algorithm is developed for parameter estimation. Experimental results on reviews of four different drugs show that PAMM is able to find better aspects than other common approaches, when measured with mean pointwise mutual information and classification accuracy. In addition, the derived aspects were also assessed by humans based on different specified perspectives, and PAMM was found to be rated highest.

Multimedia systems store and retrieve largeamounts of data which require extremely high disk bandwidth and their performance critically depends on the efficiency of disk storage. However, existing magnetic disksare designed for small amounts of data retrievals geared totraditional operations; with speed improvements mainly focused on how to reduce seek time and rotational latency.When the same mechanism is applied to multimedia systems,overheads in disk I/O can result in dramatic deterioration insystem performance. In this paper, we present a mathematical model to evaluate the performance of constant-density recording disks, and use this model to analyze quantitatively the performance of multimedia data request streams.We show that high disk throughput may be achieved by suitably adjusting the relevant parameters. In addition to demonstrating quantitatively that constant-density recording disksperform significantly better than traditional disks for multimedia data storage, a novel disk-partitioning scheme whichplaces data according to their bandwidths is presented.

Complex object-oriented queries generally consist of path expressions and explicit join operations. Since explicit join operations have been acknowledged as the most expensive operations, query executions normally start from the path expressions. Each path expression may form a sub-query. There are two existing strategies to sub-queries processing: ‘serial’ and ‘parallel’ execution scheduling strategies. Serial sub-queries execution corresponds to an execution of the sub-queries one-by-one, whereas parallel sub-queries execution corresponds to simultaneous execution of the sub-queries. When a sub-query is being processed, parallelization techniques may be applied. In this paper, we focus on the scheduling issues of the sub-queries, rather than the parallelization of the sub-queries themselves. Rules are formulated to guide the parallel query execution process. Our analysis shows that when there is no load skew, the serial scheduling strategy is preferred, otherwise the parallel scheduling strategy should be used.

Discovering and matching services among heterogeneous agents is the fundamental process for agent cooperation and coordination in multi-agent systems. In such a process, numerous service-providing agents are matched against requested criteria. During the matching procedure, an agent is given the chance to discover its new corresponding agents that can provide desirable services. However, in order to solve more complex problems, the number of agents in a multi-agent system is increasing rapidly; therefore making the chance for an efficient and desirable match is decreased. Thus, improving the efficiency of the agent matching process has become an important issue in the multi-agent filed. Utilising an appropriate agent-matching mechanism will enhance agent cooperation and communication efficiency in dealing with complex problems. In this chapter, we develop a new agent-matching algorithm, the Agent-Rank (AR hereafter) algorithm, which ranks service-providing agents according to their contributions to a nominated requesting agent. The AR algorithm improves agent matching process through combining the general ranking scores with the request-based ranking scores. The AR algorithm narrows down the searching targets (services), which is avail to detect some random targets in a chance discovery process.

Conventionally, oocyte injection has been conducted manually; however, long training, low throughput, and low success rates from poor reproducibility and inconsistency among technicians in manual operations call for the reduction of human involvement and automated injection systems. The past decade has witnessed significant efforts to automate oocyte injection using automation and robotics approaches. Several systems have been reported for robotically assisted oocyte injection. These systems all borrowed the architecture directly from manual operation and automated a few procedures. Commonalities include (1) a holding micropipette for immobilizing an oocyte, (2) an injection micropipette for penetrating the oocyte and depositing materials, and (3) microrobots for positioning the holding and injection micropipettes. Research emphases were placed upon precise microrobotic motion control, development of computer vision algorithms for localization and visual tracking of oocyte and micropipettes, visual servo control of the micropipette, or integration of visual and haptic interfaces. However, there exist several difficulties that prevent these systems or automation techniques from practical use for microinjection. In the state-of-the-art oocyte immobilization process, a robot mimics a human operator to control a holding micropipette to search, hold, and release individual oocyte (∼100 μm for mouse and ∼150 μm for human). The procedures pose tremendous difficulties to a robotic system to achieve a high speed and reliability. Consequently, switching from one oocyte to another has been conducted manually, making robotically assisted oocyte injection cumbersome and time-consuming.

Emotion-based music retrieval provides a natural and humanized way to help people experience music. In this paper, we utilize the three-dimensional Resonance-Arousal-Valence emotion model to represent the emotions invoked by music, and the relationship between acoustic features and their emotional impact based on this model is established. In addition, we also consider the emotional tag features for music, and then represent acoustic features and emotional tag features jointly in a low dimensional embedding space for music emotion, while the joint emotion space is optimized by minimizing the joint loss of acoustic features and emotional tag features through dimension reduction. Finally we construct a unified framework for music retrieval in joint emotion space by the means of query-by-music or query-by-tag or together, and then we utilize our proposed ranking algorithm to return an optimized ranked list that has the highest emotional similarity. The experimental results show that the joint emotion space and unified framework can produce satisfying results for emotion-based music retrieval.

In this work several semantic approaches to concept-based query expansion and reranking schemes are studied and compared with different ontology-based expansion methods in web document search and retrieval. In particular, we focus on concept-based query expansion schemes, where, in order to effectively increase the precision of web document retrieval and to decrease the users browsing time, the main goal is to quickly provide users with the most suitable query expansion. Two key tasks for query expansion in web document retrieval are to find the expansion candidates, as the closest concepts in web document domain, and to rank the expanded queries properly. The approach we propose aims at improving the expansion phase for better web document retrieval and precision. The basic idea is to measure the distance between candidate concepts using the PMING distance, a collaborative semantic proximity measure, i.e. a measure which can be computed by using statistical results from web search engine. Experiments show that the proposed technique can provide users with more satisfying expansion results and improve the quality of web document retrieval.

Sperm analysis and manipulation play a significant role in biology research and reproductive medicine (assisted reproductive technologies). This chapter reviews computer vision-based sperm tracking methods, sperm analysis techniques, and automated sperm manipulation. Based on computer vision tracking of sperm head and sperm tail, sperm motility can be quantified by calculating the sperm’s straight line velocity, curvilinear velocity, moving path linearity, and the sperm tail beating amplitude. Conventional computer-assisted sperm analysis (CASA) systems are capable of performing some of these tasks. Recent progress in this field provides additional, enhanced capabilities to biologists and clinical embryologists. This chapter also introduces recent progress in automating sperm manipulation procedures, including sperm immobilization, aspiration, and positioning inside a micropipette.

With the rapid development of multimedia technology, web images have become increasingly available and it constitutes a significant component of multimedia contents on the Internet. Since Web images can be represented in various formats and dimensions, searching such information are far more challenging than text-based search. While some basic forms of image retrieval are available on the Internet, these tend to be inflexible and have significant limitations. Currently, most of these image retrieval systems only rely on shallow image information (e.g., text annotation, metadata, color, etc.). Here, we present an approach for deep concept-based image retrieval, which focuses on high-level human perception, incorporating subtle nuances and emotional impressions on the images. We also propose a user-interactive and innovative adaptive method for image search that overcomes the current limitations.

The occurrence of transaction failures is often unavoidable in most database processing environments and such failures can have a particularly detrimental effect on long-duration transactions. In this paper, the impact of failures on the performance of long-duration database transactions is studied, and results for the transaction response time and failure overhead are presented for general failure patterns, which include Poisson failures as a special case. It is found that the failure overhead can be excessive for long-duration transactions and the transaction response time can increase sharply with increased failure rates. The benefits of transaction nesting to arrest performance deterioration is considered and results for quantifying the extent of restructuring improvement are provided. A procedure based on branch and bound techniques for determining the optimal restructuring strategy to minimize total transaction response time is also presented. It is shown that restructuring optimization can bring about significant performance advantages.

Manipulating single cells with a micropipette is the oldest, yet still a widely used technique. This paper discusses the positioning of a single cell to a target position inside the micropipette after the cell is aspirated into the micropipette. Due to the small volume of a single cell (pico-liter) and nonlinear dynamics involved, this task has high skill requirements and is labor intensive in manual operation that is solely based on trial and error and has high failure rates. We present automated techniques in this paper for achieving this task. Computer vision algorithm was developed to track a single cell inside a micropipette for automated single-cell positioning. A closed-loop robust controller integrating the dynamics of cell motion was designed to accurately and efficiently position the cell to a target position inside the micropipette. The system achieved high success rates of 97% for cell tracking (n=100) and demonstrated its capability of accurately positioning a cell inside the micropipette within 8 seconds (vs. 25 seconds by highly skilled operators).

Manipulating single cells with a micropipette is the oldest, yet still a widely used technique. This paper discusses the aspiration of a single cell into a micropipette and positioning the cell accurately to a target position inside the micropipette. Due to the small volume of a single cell (picoliter) and nonlinear dynamics involved, these tasks have high skill requirements and are labor intensive in manual operation that is solely based on trial and error and has high failure rates. We present automated techniques in this paper for achieving these tasks via computer vision microscopy and closed-loop motion control. Computer vision algorithms were developed to detect and track a single cell outside and inside a micropipette for automated single-cell aspiration. A closed-loop robust controller integrating the dynamics of cell motion was designed to accurately and efficiently position the cell to a target position inside the micropipette. The system achieved high success rates of 98% for cell detection and 97% for cell tracking (n = 100). The automated system also demonstrated its capability of aspirating a single cell into a micropipette within 2 s (versus 10 s by highly skilled operators) and accurately positioning the cell inside the micropipette within 8 s (versus 25 s by highly skilled operators).

Effective sharing of diverse social media is often inhibited by limitations in their search and discovery mechanisms, which are particularly restrictive for media that do not lend themselves to automatic processing or indexing. Here, we present the structure and mechanism of an adaptive search engine which is designed to overcome such limitations. The basic framework of the adaptive search engine is to capture human judgment in the course of normal usage from user queries in order to develop semantic indexes which link search terms to media objects semantics. This approach is particularly effective for the retrieval of multimedia objects, such as images, sounds, and videos, where a direct analysis of the object features does not allow them to be linked to search terms, for example, nontextual/icon-based search, deep semantic search, or when search terms are unknown at the time the media repository is built. An adaptive search architecture is presented to enable the index to evolve with respect to user feedback, while a randomized query-processing technique guarantees avoiding local minima and allows the meaningful indexing of new media objects and new terms. The present adaptive search engine allows for the efficient community creation and updating of social media indexes, which is able to instill and propagate deep knowledge into social media concerning the advanced search and usage of media resources. Experiments with various relevance distribution settings have shown efficient convergence of such indexes, which enable intelligent search and sharing of social media resources that are otherwise hard to discover.

Music emotion expresses inherent and high-level states of mind and spiritual quality. In this paper, a hierarchical framework is proposed, which consists of two layers: an external layer that represents preliminary and superficial emotions and an inherent layer that represents psychic and resonant emotions. Using these two layers, a Resonance-Arousal-Valence (RAV) emotion model has been constructed. Five feature sets, including intensity, timbre, rhythm, pitch and tonality, and harmony, are extracted to represent music emotions in the RAV model. In order to effectively represent emotions with extracted features, suitable weighting schemes are utilized to balance the different features. As each music clip may have rather complex emotions, a supervised multiclass label model is adopted to annotate emotions with emotion multinomial. Preliminary experimental results indicate that the proposed emotion model and retrieval approach is able to deliver good retrieval performance.

With the rapid advancement of music compression and storage technologies, digital music can be easily created, shared and distributed, not only in computers, but also in numerous portable digital devices. Music often constitutes a key component in many multimedia databases, and as they grow in size and complexity, their meaningful search and retrieval become important and necessary. Music Information Retrieval (MIR) is a relatively young and challenging research area started since the late 1990s. Although some form of music retrieval is available on the Internet, these tend to be inflexible and have significant limitations. Currently, most of these music retrieval systems only rely on low-level music information contents (e.g., metadata, album title, lyrics, etc.), and in this chapter, the authors present an adaptive indexing approach to search and discover music information. Experimental results show that through such an indexing architecture, high-level music semantics may be incorporated into search strategies.

We study several semantic concept-based query expansion and re-ranking scheme and compare different ontology-based expansion methods in image search and retrieval. In particular, we exploit the two concept similarities of different concept expansion ontology-WordNet Similarity, Wikipedia Similarity. Furthermore, we compare the keywords semantic distance with the precision of image search results with query expansion according to different concept expansion algorithms. We also compare the image retrieval precision of searching with the expanded query and original plain query. Preliminary experiments have been able to demonstrate that the two proposed retrieval mechanism has the potential to outperform unaided approaches.

This paper presents an adaptive architecture for educational games which are evolved and optimized in order to fulfill the educational goals while reflecting the specific requirements and features of the user community. The approach is based on an online genetic framework where typical genetic operators like crossover and mutation are designed to evolve a population of games, and the online fitness driving the evolution is given by a metric of user behavior/performance, evaluated on the actual community of users. The evolutionary structure adapt can also be applied for continuous game adaptation in dynamical domains where the user community and/or the educational goals are changing over the time.

An efficient service matching process is crucial for solving complex problems based on heterogeneous agents. Agent cooperation can be achieved through matching requesting agents with service-providing agents, and, through such cooperation, multi-agents can solve a variety of complex problems. Improving the efficiency of the agent-matching process has become an important issue in multi-agent research. The adoption of an appropriate agent-matching mechanism will enhance agent cooperation and communication efficiency within an agent network. In this paper, we develop a new agent-matching algorithm, the Agent-Rank algorithm, which ranks service-providing agents according to their contributions to a nominated requesting agent based on Agent Association Graphs. The Agent-Rank algorithm overcomes the problems of agent-matching in a large agent network through combining the general ranking scores with the request-based ranking scores. In our experimental evaluation, we have found that the Agent-Rank algorithm can significantly improve efficiency in the agent-matching and re-matching processes.

Optimization of web content presentation poses a key challenge for e-commerce applications. Whether considering web pages, advertising banners or any other content presentation media on the web, the choice of the appropriate structure and appearance with respect to the given audience can obtain a more effective and successful impact on users, such as gathering more readers to web sites or customers to online shops. Here, the collective optimization of web content presentation based on the online discrete Particle Swarm Optimization (PSO) model is presented. The idea behind online PSO is to evaluate the collective user feedback as the PSO objective function which drives particles’ velocities in the hybrid continuous-discrete space of web content features. The PSO coordinates the process of sampling collective user behaviour in order to optimize a given user-based metric. Experiments in the online banner optimization scenario show that the method converges faster than other methods and avoid some common drawbacks such as local optima and hybrid discrete/continuous features management. The proposed online optimization method is sufficiently general and may be applied to other web marketing or business intelligence contexts.

Tracking sperm tail movement provides important information for clinical sperm research. It is also a crucial step for sperm immobilization in Intracytoplasmic Sperm Injection (ICSI). However, the low visibility of the sperm tail under optical microscopy, coupled with the sperm fast motility, render sperm tail identification and tracking challenging tasks to execute. This paper presents two approaches for sperm tail tracking: (1) the Maximum Intensity Region (MIR) algorithm, and (2) the Optical Flow (OF) algorithm. The algorithms were evaluated by calculating the Euclidean distance error between each tail tracking algorithm's computed tail location and a user's manual input via mouse click of the tail's image location. Experimental results demonstrate that the OF algorithm and MIR algorithm are both capable of tracking the sperm tail with minimal error when viscous liquid is added to the sperm culture medium, which is the present clinical standard practice for slowing down sperm movement. The MIR algorithm outperforms the OF algorithm by 52% in tail tracking accuracy in situations where the viscous liquid is absent.

In spite of the rising demands for reusable information systems, current designs are still insufficient in providing efficient reusable mechanisms for system design. One of the major problems hindering the development of information reuse in most traditional systems, which include component-based systems and object-oriented systems, is the lack of the self-organising ability among the system components or subsystems. The emergence of intelligent agent-based technology is able to solve the problems plaguing many traditional systems. In this paper we introduce an optimised design for agent-based systems, which is able to provide an efficient process for agent capability reuse. An experimental program is developed to evaluate the performance of the proposed design.

Human users spend a vast amount of time in interacting with image contents on the Web. Their interaction entails the exercise of considerable perceptive intelligence, visual judgment and mental evaluation. For high-level semantic image features and concepts, such processes of intelligent judgment cannot be mechanized or carried out automatically by machines. In this chapter, an indexing method is described whereby the aggregate intelligence of different Web users is continuously transferred to the Web. Such intelligence is codified, reinforced, distilled and shared among users so as to enable the systematic mining and discovery of semantic image contents. This method allows the collaborative creation of image indexes, which is able to instill and propagate deep knowledge and collective wisdom into the Web concerning the advanced semantic characteristics of Web images. This method is robust and adaptive, and is able to respond dynamically to changing usage patterns caused by community trends and social networking.

Effectively finding relevant passages in a full-text database of software documentation calls for a user interface that does more than mimic a printed book. A hypertext approach, with a network of links among passages, offers great flexibility but often at the cost of high cognitive overhead and a disorienting lack of contextual cues. A tree-based approach guides users along branching paths through a hierarchy of text nodes. The “natural”, sequential implementation of such hierarchical access, however, is psychologically inept in large databases because it is order-dependent, discriminates awkwardly among key terms, clarifies each node's context incompletely, and often involves much semantic redundancy. An alternative, mixed approach, recently implemented in the on-line documentation system at the National Energy Research Supercomputer Center (NERSC), overcomes three of these four problems. It displays only local tree structure in response to “zoomin” or “zoomout” commands issued to focus a search begun with typical hypertext moves. This combination approach enjoys the benefits of cued, spatially interpreted hierarchical search while avoiding most of its known pitfalls. Usage monitoring at NERSC shows the ready acceptance of both zoom commands by documentation readers.

A new Automatic Repeat Request (ARQ) scheme, Virtual Multiple Data Links (VMDL), for satellite communications is proposed. The principle of VMDL is based on the basic go-back- N scheme and plural data links. It is not only able to achieve high performance on high speed satellite channels with large round trip delays, but its recovery procedure is also simple to implement. The performance of VMDL is analysed and it is demonstrated that VMDL is superior in terms of both throughput efficiency and error detection time compared with conventional schemes.

This tutorial will be based on the recently published book, High-Performance Parallel Database Processing and Grid Databases (John Wiley & Sons, 2008). The sizes of databases have seen exponential growth in the past and such growth is expected to accelerate in the future, with the steady drop in storage cost accompanied by a rapid increase in storage capacity. To effectively manage such volumes of data, it is necessary to allocate multiple resources to it, very often massively so. The processing of databases of such astronomical proportions requires an understanding of how high performance systems and parallelism work. Besides the massive volume of data in the database to be processed, some data has been distributed across the globe in a Grid environment. This important new book provides readers with a fundamental understanding of parallelism in data-intensive applications, and demonstrates how to develop faster capabilities to support them. It features not only the algorithms for database operations, but also quantitative analytical models, so that performance can be analyzed and evaluated more effectively.

In this paper we propose an approach to optimization of web marketing content based on online discrete particle swarm optimization (PSO) model. The idea behind online PSO is to evaluate the collective user feedback as the PSO objective function which drive particles velocities in the hybrid continuous-discrete space of web content features. The PSO coordinate the process of sampling collective user behavior in order to optimize the web marketing metric. Experiments in the online banner optimization scenario show that the method converges faster than other methods and avoid some common drawbacks such as local optimal and hybrid discrete/continuous features management. The proposed online optimization method is general and can be applied to other web marketing or business intelligent contexts.

A challenging problem facing the semantic search of multimedia data objects is the ability to index them. Here, we present an architectural paradigm for collaborative semantic indexing, which makes use of a dynamic evolutionary approach. By capturing, analyzing and interpreting user response and query behavior, the patterns of searching and finding multimedia data objects may be established. Within the present architectural paradigm, the semantic index may be dynamically constructed, validated, and built-up, where the performance of the system will increase as time progresses. Our system also incorporates a high degree of robustness and fault-tolerance whereby inappropriate index terms will be gradually eliminated from the index, while appropriate ones will be reinforced. We also incorporate genetic variations into the design to allow objects which may otherwise be hidden to be discovered. Experimental results indicate that the present approach is able to confer significant performance benefits in the semantic searching and discovery of a wide variety of multimedia data objects.