Point clouds have attracted attention as a representation of an object’s surface. Deep generative models have typically used a continuous map from a dense set in a latent space to express their variations. However, a continuous map cannot adequately express the varying numbers of holes. That is, previous approaches disregarded the topological structure of point clouds. Furthermore, a point cloud comprises several subparts, making it difficult to express it using a continuous map. This paper proposes ChartPointFlow, a flow-based deep generative model that forms a map conditioned on a label. Similar to a manifold chart, a map conditioned on a label is assigned to a continuous subset of a point cloud. Thus, ChartPointFlow is able to maintain the topological structure with clear boundaries and holes, whereas previous approaches generated blurry point clouds with fuzzy holes. The experimental results show that ChartPointFlow achieves state-of-the-art performance in various tasks, including generation, reconstruction, upsampling, and segmentation.

Accurate and automated detection of anomalous samples in an image dataset can be accomplished with a probabilistic model. Such images have heterogeneous complexity, however, and a probabilistic model tends to overlook simply shaped objects with small anomalies. The reason is that a probabilistic model assigns undesirable lower likelihoods to complexly shaped objects, which are nevertheless consistent with the current set standards. This difficulty is critical, especially for a defect detection task, where the anomaly can be a small scratch or grime. To overcome this difficulty, we propose an unregularized score for deep generative models (DGMs). We found that the regularization terms of the DGMs considerably influence the anomaly score depending on the complexity of the samples. By removing these terms, we obtain an unregularized score, which we evaluated on toy datasets, two in-house manufacturing datasets, and on open manufacturing and medical datasets. The empirical results demonstrate that the unregularized score is robust to the apparent complexity of given samples and detects anomalies selectively.

A point cloud serves as a representation of the surface of a three-dimensional (3D) shape. Deep generative models have been adapted to model their variations typically using a map from a ball-like set of latent variables. However, previous approaches did not pay much attention to the topological structure of a point cloud, despite that a continuous map cannot express the varying numbers of holes and intersections. Moreover, a point cloud is often composed of multiple subparts, and it is also difficult to express. In this study, we propose ChartPointFlow, a flow-based generative model with multiple latent labels for 3D point clouds. Each label is assigned to points in an unsupervised manner. Then, a map conditioned on a label is assigned to a continuous subset of a point cloud, similar to a chart of a manifold. This enables our proposed model to preserve the topological structure with clear boundaries, whereas previous approaches tend to generate blurry point clouds and fail to generate holes. The experimental results demonstrate that ChartPointFlow achieves state-of-the-art performance in terms of generation and reconstruction compared with other point cloud generators. Moreover, ChartPointFlow divides an object into semantic subparts using charts, and it demonstrates superior performance in case of unsupervised segmentation.

With the significant development of black-box machine learning algorithms, particularly deep neural networks, the practical demand for reliability assessment is rapidly increasing. On the basis of the concept that “Bayesian deep learning knows what it does not know,” the uncertainty of deep neural network outputs has been investigated as a reliability measure for classification and regression tasks. By considering an embedding task as a regression task, several existing studies have quantified the uncertainty of embedded features and improved the retrieval performance of cutting-edge models by model averaging. However, in image-caption embedding-and-retrieval tasks, well-known samples are not always easy to retrieve. This study shows that the existing method has poor performance in reliability assessment and investigates another aspect of image-caption embedding-and-retrieval tasks. We propose posterior uncertainty by considering the retrieval task as a classification task, which can accurately assess the reliability of retrieval results. The consistent performance of the two uncertainty measures is observed with different datasets (MS-COCO and Flickr30k), different deep-learning architectures (dropout and batch normalization), and different similarity functions. To the best of our knowledge, this is the first study to perform a reliability assessment on image-caption embedding-and-retrieval tasks.

Neural architecture search (NAS) is a promising method to automatically identify neural network architectures. Differentiable architecture search (DARTS) is method that significantly reduces search time and finds architectures that can achieve state-of-the-art performance. For computer vision tasks, DARTS searches convolutional neural networks (CNNs) via stacking convolution layers and pooling operations. Recent studies on neural architectures indicate that attention modules can improve the performances of CNNs by discarding information of no interest, while existing NAS methods have put little focus on it. In this study, we propose Att-DARTS, which searches attention modules as well as convolution and pooling operations simultaneously. In our experiments on CIFAR-10 and CIFAR-100 datasets, we demonstrate that Att-DARTS can find architectures that achieve lower classification error rates and require fewer parameters compared to those found by DARTS.

Deep convolutional neural networks (CNNs) have achieved remarkable results in image processing tasks. However, their high expression ability risks overfitting. Consequently, data augmentation techniques have been proposed to prevent overfitting while enriching datasets. Recent CNN architectures with more parameters are rendering traditional data augmentation techniques insufficient. In this study, we propose a new data augmentation technique called random image cropping and patching (RICAP) which randomly crops four images and patches them to create a new training image. Moreover, RICAP mixes the class labels of the four images, resulting in an advantage of the soft labels. We evaluated RICAP with current state-of-the-art CNNs (e.g., the shake-shake regularization model) by comparison with competitive data augmentation techniques such as cutout and mixup. RICAP achieves a new state-of-the-art test error of 2.19% on CIFAR-10. We also confirmed that deep CNNs with RICAP achieve better results on classification tasks using CIFAR-100 and ImageNet, an image-caption retrieval task using Microsoft COCO, and other computer vision tasks.

Reinforcement learning (RL) builds an effective agent that handles tasks in complex and uncertain environments by maximizing future reward. However, the efficiency is insufficient for practical use such as game AI and autonomous driving. An effective but selfish agent conflicts with other humans, and hence the demand of a human-like behavior arises. Imitation learning (IL) has been employed to train an agent to mimic the actions of expert behaviors provided as training data. However, IL tends to build an agent limited in performance by the expert skill, and even worse, the agent exhibits an inconsistent behavior since IL is not goal-oriented. In this paper, we propose a training scheme by mixing RL and IL for both discrete and continuous action space problems. The proposed scheme builds an agent that achieves a performance higher than an agent trained by only IL and exhibits a more human-like behavior than agents trained by RL or IL, validated by human sensitivity.

An early and accurate diagnosis of psychiatric disorders is critical for patients' quality of life and deep understanding of the disorders. For this reason, many studies have proposed machine learning-based diagnostic procedures for functional magnetic resonance imaging (fMRI) data. Especially, these procedures often employed temporal models due to the time-varying nature of the brain activities and probabilistic generative models for understanding the underlying mechanism of the disorders. For leveraging the recent advantage of deep learning, we proposed a state-space model of fMRI images based on deep learning. The proposed deep state-space model is more flexible than conventional models and less likely to suffer from overfitting than a straightforward deep learning-based classifier. The proposed model estimates the subjects' conditions more accurately than existing diagnostic procedures. Also, the proposed model potentially identifies brain regions related to the disorders.

Image-based anomaly segmentation is a fundamental topic for image analysis. For medical use, it supports treatments via refined diagnosis and growth rate evaluation of tumors and lesions. Especially, an unsupervised training is expected to generalize to unknown anomalies. Probabilistic models have been used for this purpose, whereby these models are trained to maximize the likelihood of known samples and detect anomalous samples by assigning low likelihoods. Recent studies have proposed a probabilistic model based on deep neural networks (DNNs) called AEs and they achieved significant performance thanks to their flexibility. However, AEs are sensitive to complex structure (e.g., ridges and grooves of a brain) rather than semantic anomalies (e.g., tumors and lesions). We decomposed the approximated log-likelihood into two terms; predictable uncertainty and normalized error. We found that the former represents the complexity of structure. Hence, we propose the normalized error as a novel uncertainty-sensitive score by removing the predictable uncertainty. We evaluated our score by experiments with head magnetic resonance imaging (MRI) datasets and demonstrate the robustness of the proposed normalized error to data complexity.

Convolutional neural networks (CNNs) have demonstrated remarkable results in image classification for benchmark tasks and practical applications. The CNNs with deeper architectures have achieved even higher performance recently thanks to their robustness to the parallel shift of objects in images and their numerous parameters and the resulting high expression ability. However, CNNs have a limited robustness to other geometric transformations such as scaling and rotation. This limits the performance improvement of the deep CNNs, but there is no established solution. This paper focuses on scale transformation and proposes a network architecture called the weight-shared multi-stage network (WSMS-Net), which consists of multiple stages of CNNs. The proposed WSMS-Net is easily combined with existing deep CNNs such as residual network and densely connected convolutional network and enables them to acquire robustness to object scaling. Experimental results on the CIFAR-10, CIFAR-100, and ImageNet datasets demonstrate that existing deep CNNs combined with the proposed WSMS-Net achieve higher accuracies for image classification tasks with only a minor increase in the number of parameters and computation time.

Deep convolutional neural networks (CNNs) have demonstrated remarkable results in image recognition owing to their rich expression ability and numerous parameters. However, an excessive expression ability compared to the variety of training images often has a risk of overfitting. Data augmentation techniques have been proposed to address this problem as they enrich datasets by flipping, cropping, resizing, and color-translating images. They enable deep CNNs to achieve an impressive performance. In this study, we propose a new data augmentation technique called \emph{random image cropping and patching} (\emph{RICAP}), which randomly crops four images and patches them to construct a new training image. Hence, RICAP randomly picks up subsets of original features among the four images and discard others, enriching the variety of training images. Also, RICAP mixes the class labels of the four images and enjoys a benefit similar to label smoothing. We evaluated RICAP with current state-of-the-art CNNs (e.g., shake-shake regularization model) and achieved a new state-of-the-art test error of \textcolor{red}{$2.23%$} on CIFAR-10 among competitive data augmentation techniques such as cutout and mixup. We also confirmed that deep CNNs with RICAP achieved better results on CIFAR-100 and ImageNet than those results obtained by other techniques.

Deep neural networks have a rich ability to learn complex representations and achieved remarkable results in various tasks. However, they are prone to overfitting due to the limited number of training samples; regularizing the learning process of neural networks is critical. In this paper, we propose a novel regularization method, which estimates parameters of a large convolutional neural network as implicit probabilistic distributions generated by a hypernetwork. Also, we can perform model averaging to improve the network performance. Experimental results demonstrate our regularization method outperformed the commonly-used maximum a posterior (MAP) estimation.

One of the most common needs in manufacturing plants is rejecting products not coincident with the standards as anomalies. Accurate and automatic anomaly detection improves product reliability and reduces inspection cost. Probabilistic models have been employed to detect test samples with lower likelihoods as anomalies in unsupervised manner. Recently, a probabilistic model called deep generative model (DGM) has been proposed for end-to-end modeling of natural images and already achieved a certain success. However, anomaly detection of machine components with complicated structures is still challenging because they produce a wide variety of normal image patches with low likelihoods. For overcoming this difficulty, we propose unregularized score for the DGM. As its name implies, the unregularized score is the anomaly score of the DGM without the regularization terms. The unregularized score is robust to the inherent complexity of a sample and has a smaller risk of rejecting a sample appearing less frequently but being coincident with the standards.

Deep convolutional neural networks (CNNs) have become one of the most successful methods for image processing tasks in past few years. Recent studies on modern residual architectures, enabling CNNs to be much deeper, have achieved much better results thanks to their high expressive ability by numerous parameters. In general, CNNs are known to have the robustness to the small parallel shift of objects in images by their local receptive fields, weight parameters shared by each unit, and pooling layers sandwiching them. However, CNNs have a limited robustness to the other geometric transformations such as scaling and rotation, and this lack becomes an obstacle to performance improvement even now. This paper proposes a novel network architecture, the \emphweight-shared multi-stage network (WSMS-Net), and focuses on acquiring the scale invariance by constructing of multiple stages of CNNs. The WSMS-Net is easily combined with existing deep CNNs, enables existing deep CNNs to acquire a robustness to the scaling, and therefore, achieves higher classification accuracy on CIFAR-10, CIFAR-100 and ImageNet datasets.

Many comic books are now published as digital books, which easily provide color contents compared to the physical books. The motivation of automatic colorization of comic books now arises. Previous studies colorize sketches without other clues or with spatial color annotations. They are expected to reduce workloads of comic artists but still require spatial color annotations for desirable colorizations. This study introduces a color style information and combines it with conditional adversarially learned inference. The experimental results demonstrate that the objects are painted with colors depending on the color style information and that the color style information extracted from a color image supports to painting an object with a desirable color.

Artificial intelligence (AI) agent created with Deep Q-Networks (DQN) can defeat human agents in video games. Despite its high performance, DQN often exhibits odd behaviors, which could be immersion-breaking against the purpose of creating game AI. Moreover, DQN is capable of reacting to the game environment much faster than humans, making itself invincible (thus not fun to play with) in certain types of games. On the other hand, supervised learning framework trains an AI agent using historical play data of human agents as training data. Supervised learning agent exhibits a more human-like behavior than reinforcement learning agents because of imitating training data. However, its performance is often no better than human agents. The ultimate purpose of AI agents is to entertain human players. A good performance and a humanlike behavior are important factors of the AI agents, and both of them should be achieved simultaneously. This study proposes frameworks combining reinforcement learning and supervised learning and we call then separated network model and shared network model. We evaluated their performances by the game scores and behaviors by Turing test. The experimental results demonstrate that the proposed frameworks develop an AI agent of better performance than human agent and natural behavior than reinforcement learning agents.

This paper proposes a novel application of deep learning models, Paragraph Vector, and Long Short-Term Memory (LSTM), to financial time series forecasting. Investors make decisions according to various factors, including consumer price index, price-earnings ratio, and miscellaneous events reported in newspapers. In order to assist their decisions in a timely manner, many automatic ways to analyze those information have been proposed in the last decade. However, many of them used either numerical or textual information, but not both for a single company. In this paper, we propose an approach that converts newspaper articles into their distributed representations via Paragraph Vector and models the temporal effects of past events on opening prices about multiple companies with LSTM. The performance of the proposed approach is demonstrated on real-world data of fifty companies listed on Tokyo Stock Exchange.

For image generation, deep neural networks are trained to extract high-level features on natural images and to reconstruct the images from the features. However it is difficult to learn to generate images containing enormous contents. To overcome this difficulty, a network with an attention mechanism has been proposed. It is trained to attend to parts of the image and to generate images step by step. This enables the network to deal with the details of a part of the image and the rough structure of the entire image. The attention mechanism is implemented by recurrent neural networks. Additionally, the Generative Adversarial Networks (GANs) approach has been proposed to generate more realistic images. In this study, we present image generation where leverages effectiveness of attention mechanism and the GANs approach. We show our method enables the iterative construction of images and more realistic image generation than standard GANs and the attention mechanism of DRAW.

Semi-supervised learning is a topic of practical importance because of the difficulty of obtaining numerous labeled data. In this paper, we apply an extension of adversarial autoencoder to semi-supervised learning tasks. In attempt to separate style and content, we divide the latent representation of the autoencoder into two parts. We regularize the autoencoder by imposing a prior distribution on both parts to make them independent. As a result, one of the latent representations is associated with content, which is useful to classify the images. We demonstrate that our method disentangles style and content of the input images and achieves less test error rate than vanilla autoencoder on MNIST semi-supervised classification tasks.

Homeostatic plasticity in mammalian central nervous system is considered to maintain activity in neuronal circuits within a functional range. In the absence of homeostatic plasticity neuronal activity is prone to be destabilized because correlation-based synaptic modification, Hebbian plasticity, induces positive feedback change. Several studies on homeostatic plasticity assumed the existence of a process for monitoring neuronal activity and adjusting synaptic efficacy on a time scale of hours, but its biological mechanism still remains unclear. Excitatory synaptic efficacy is associated with the size of a post-synaptic element, dendritic spine, and the size of the dendritic spine fluctuates even after neuronal activity is silenced. These fluctuations could be a non-Hebbian form of synaptic plasticity that serves such a homeostatic function. This study proposed and analyzed a synaptic plasticity model incorporating random fluctuations and Hebbian plasticity at excitatory synapses, and found that it prevents excessive changes in neuronal activity by adjusting synaptic efficacy. Random fluctuations do not monitor neuronal activity, but their relative influence depends on neuronal activity. The proposed synaptic plasticity model acts as a form of homeostatic plasticity, regardless of neuronal activity monitoring. Thus, random fluctuations play an important role in homeostatic plasticity and contribute to development and functions of neural networks.

This paper introduces a method for Multimedia Event Detection (MED). Given training videos for a certain event, a classifier is constructed to identify videos displaying it. In particular, the problems of the weakly supervised setting and the unclear event structure are addressed in this paper. The first issue is associated with the loosely annotated training videos that usually contain many irrelevant shots. The second one is the difficulty of assuming the event structure in advance, because videos can be created by arbitrary camera and editing techniques. To overcome these problems, a Hidden Conditional Random Field (HCRF) is used where hidden states work as an intermediate layer to discriminate between relevant and irrelevant shots to the event. In addition, the relation among hidden states characterises the event structure. Thus, the above problems are managed by optimising hidden states and their relation, so as to distinguish videos where the event occurs from the rest of videos. Experimental results on TRECVID video data validate the effectiveness of HCRFs in this context.

An application of Query-By-Example (QBE) is presented where shots that are visually similar to provided example shots are retrieved. To implement QBE, counter-example shots are required to accurately distinguish shots that are relevant to the query from those that are not (Li and Snoek ( 2009 ), Yu et al. ( 2004 )). However, there are usually a huge number of shots, not relevant to a particular query, which can serve as counter-example shots. It is difficult for a user to provide counter-example shots that would aid retrieval. Thus, we developed a QBE method based on partially supervised learning where a retrieval model is constructed by selecting counter-example shots from shots without user supervision. To ensure the speed and accuracy of the QBE method, we select a small number of counter-example shots that are visually similar to given example shots but irrelevant to the query. Such shots are useful for characterizing the boundary between relevant and irrelevant shots. For our method, we first filter shots that are visually dissimilar to example shots based on SVMs on a visual feature. Then we filter shots relevant to the query based on concept detection results from pre-constructed classifiers. Shots that pass the above two tests are considered as counter-example shots. Experimental results obtained using TRECVID 2009 video data validate the effectiveness of our method.

This paper proposes an approach to clustering synonymous noun phrases focusing on two types of predicate argument relations extracted from potentially big textual data. One is associated with common effects, the other with common influences. Based on the context represented by those relations, a matrix is constructed with rows being noun phrases and columns being a pair of a noun phrase and a verb phrase. Following the distribution hypothesis often adopted in the literature, it is assumed that rows (i.e., noun phrases) with similar distributions share similar meanings. Due to the inherent sparsity of the matrix, however, two strategies are taken to group noun phrases having similar distributions. One strategy is to simply use a large-scale corpus, which however results in an even larger matrix. To handle the large matrix, a parallel distributed programming model, MapReduce, is employed. The other is to adopt hierarchical agglomerative co-clustering and approximates its computation in a way suited to the MapReduce programming model. The proposed approach is evaluated based on a series of experiments in terms of the validity of our underlying assumptions, processing time, quality of the resulting clusters, and effect of parallelization.

Query expansion methods using pseudo-relevance feedback have been shown effective for microblog search because they can solve vocabulary mismatch problems often seen in searching short documents such as Twitter messages (tweets), which are limited to 140 characters. Pseudo-relevance feedback assumes that the top ranked documents in the initial search results are relevant and that they contain topic-related words appropriate for relevance feedback. However, those assumptions do not always hold in reality because the initial search results often contain many irrelevant documents. In such a case, only a few of the suggested expansion words may be useful with many others being useless or even harmful. To overcome the limitation of pseudo-relevance feedback for microblog search, we propose a novel query expansion method based on two-stage relevance feedback that models search interests by manual tweet selection and integration of lexical and temporal evidence into its relevance model. Our experiments using a corpus of microblog data (the Tweets2011 corpus) demonstrate that the proposed two-stage relevance feedback approaches considerably improve search result relevance over almost all topics.

This paper proposes a new approach to trajectory pattern mining, which attempts to discover frequent movement patterns from the trajectories of moving objects. For dealing with a large volume of trajectory data, traditional approaches quantize them by a grid with a fixed resolution. However, an appropriate resolution often varies across different areas of trajectories. Simply increasing the resolution cannot capture broad patterns and consumes unnecessarily large computational resources. To solve the problem, we propose a hierarchical grid-based approach with quadtree search. The approach initially searches for frequent patterns with a coarse grid and drills down into a finer grid level to discover more minute patterns. The algorithm is naturally parallelized and implemented in the MapReduce programming model to accelerate the computation. Our evaluative experiments on real-word data show the effectiveness of our approach in mining complex patterns with lower computational cost than the previous work.

This paper develops a query-by-example method for retrieving shots of an event (event shots) using example shots provided by a user. The following three problems are mainly addressed. Firstly, event shots cannot be retrieved using a single model as they contain significantly different features due to varied camera techniques, settings and so forth. This is overcome by using rough set theory to extract multiple classification rules with each rule specialized to retrieve a portion of event shots. Secondly, since a user can only provide a small number of example shots, the amount of event shots retrieved by extracted rules is inevitably limited. We thus incorporate bagging and the random subspace method. Classifiers characterize significantly different event shots depending on example shots and feature dimensions. However, this can result in the potential retrieval of many unnecessary shots. Rough set theory is used to combine classifiers into rules which provide greater retrieval accuracy. Lastly, counter example shots, which are a necessity for rough set theory, are not provided by the user. Hence, a partially supervised learning method is used to collect these from shots other than example shots. Counter example shots, which are as similar to example shots as possible, are collected because they are useful for characterizing the boundary between event shots and the remaining shots. The proposed method is tested on TRECVID 2009 video data.

In this paper, we extend the traditional `Query-By-Example' (QBE) approach where example shots are provided to represent a query, and then shots similar to them are retrieved. One crucial problem of QBE is that when example shots for the query are unavailable, the retrieval cannot be performed. To overcome this, we propose an innovative approach, named `Query-By-Virtual-Example' (QBVE), where example shots for any arbitrary query can be created by using virtual reality techniques. We call such example shots `virtual example shots'. In our system, virtual example shots are created by synthesizing user's gesture in front of a video camera, 3D object models (3DCGs) and background images. Experimental results on TRECVID 2009 video data show the validity of substituting virtual example shots with real example shots used in QBE.

In recent years, a number of machine learning approaches to literature-based gene function annotation have been proposed. However, due to issues such as lack of labeled data, class imbalance and computational cost, they have usually been unable to surpass simpler approaches based on string-matching. In this paper, we propose a principled machine learning approach based on kernel classifiers.

This article proposes a novel application of a statistical language model to opinionated document retrieval targeting weblogs (blogs). In particular, we explore the use of the trigger model—originally developed for incorporating distant word dependencies—in order to model the characteristics of personal opinions that cannot be properly modeled by standard n-grams. Our primary assumption is that there are two constituents to form a subjective opinion. One is the subject of the opinion or the object that the opinion is about, and the other is a subjective expression; the former is regarded as a triggering word and the latter as a triggered word. We automatically identify those subjective trigger patterns to build a language model from a corpus of product customer reviews. Experimental results on the Text Retrieval Conference Blog track test collections show that, when used for reranking initial search results, our proposed model significantly improves opinionated document retrieval. In addition, we report on an experiment on dynamic adaptation of the model to a given query, which is found effective for most of the difficult queries categorized under politics and organizations. We also demonstrate that, without any modification to the proposed model itself, it can be effectively applied to polarized opinion retrieval.

Traditionally, HMM-based approaches to online Kanji handwriting recognition have relied on a hand-made dictionary, mapping characters to primitives such as strokes or substrokes. We present an unsupervised way to learn a stroke tagger from data, which we eventually use to automatically generate such a dictionary. In addition to not requiring a prior hand-made dictionary, our approach can improve the recognition accuracy by exploiting unlabeled data when the amount of labeled data is limited.

This paper proposes a novel application of a statistical language model to opinionated document retrieval targeting weblogs (blogs). In particular, we explore the use of the trigger model---originally developed for incorporating distant word dependencies---in order to model the characteristics of personal opinions that cannot be properly modeled by standard n-grams. Our primary assumption is that there are two constituents to form a subjective opinion. One is the subject of the opinion or the object that the opinion is about, and the other is a subjective expression; the former is regarded as a triggering word and the latter as a triggered word. We automatically identify those subjective trigger patterns to build a language model from a corpus of product customer reviews. Experimental results on the TREC Blog Track test collections show that, when used for reranking initial search results, our proposed model significantly improves opinionated document retrieval by over 20% in MAP. In addition, we report on an experiment on dynamic adaptation of the model to a given query, which is found effective for most of difficult queries categorized under politics and organizations.

This paper presents a method for acquiring a concession strategy of an agent in multi-issue negotiation. This method learns how to make a concession to an opponent for realizing win-win negotiation. To learn the concession strategy, we adopt reinforc

In Japanese, it is quite common for the same word to be written in several different ways. This is especially true for katakana words which are typically used for transliterating foreign languages. This ambiguity becomes critical for automatic processing such as information retrieval (IR). To tackle this problem, we propose a simple but effective approach to generating katakana variants by considering phonemic representation of the original language for a given word. The proposed approach is evaluated through an assessment of the variants it generates. Also, the impact of the generated variants on IR is studied in comparison to an existing approach using katakana rewriting rules.

Recently, many learning systems, such as e-learning and WBT (Web based teaching) systems have been developed. In these systems, users can get educational contents and graphical material in a remote place through the network. However, the learning is always one-sided. That is, all users learn the same contents and cannot convey their states like "what is he or she doing now?" to these systems. Therefore, these systems are not necessarily satisfying all the users' demands. In this paper, we propose an interactive learning application of cooking in mixed reality (MR) environment. In an MR environment, a system should not only provide the useful knowledge, but also recognize users' state. Furthermore, by making a parallel transition model about cooking, it is possible to control the user's cooking process. Therefore, users can learn details about cooking more flexibly and effectively

In this paper, we propose an online training-oriented video shooting navigation system focused on camerawork based on video grammar by real-time camerawork evaluation to train users shooting nice shots for the later editing work. In this system, the processing speed must be very high so that we use a luminance projection correlation and a structure tensor method to extract the camerawork parameters in real-time. From the results of camerawork analysis, the results of each frame are classified into 7 camerawork types and the system issues 6 types of alarms and navigates users along the specified shot depending on camerawork based on video grammar in real-time while shooting the shot. Thereby, users can naturally acquire shooting style by trying to decrease alarms of improper camerawork without a consideration of the video grammar

Using multimedia techniques in systems that provide assistance to users has become a common interest in recent research. To achieve a supreme level of interactivity in these systems, taking into account users' states as well as users' preferences is extremely important. However, most actual systems do not consider users' preferences when providing assistance to users. We design a system that selects a desirable recipe for a user to cook based on preference, then guide him through the cooking. The recipe selection relies on people's common interpretation of cooking recipes. We store users' feedback (interpretations) in the form of collaborative filtering, and this feedback is used to identify properties of other recipes. Furthermore, users can learn more flexibly and effectively by controlling their process using a parallel transition model for cooking. Using a parallel transition model, users are able to perform parallel actions in any order they want and the system advise them according to the action taken

This paper describes a method for extracting reliable reputation on the Web. In this research, reliable reputation is the information that has an opposite polarity value of contributor's stance (positive or negative). We call this information "fair reputation". In order to extract fair reputations, we develop the following two tasks. The first task is classification of feedback documents into positive or negative classes. For this task, we propose a classification method using 'document level reputation' that can determine the polarity value of the document. The second task is extraction and classification of reputations. Using the classified reputations, we extract "fair reputations". The experimental results using movie reviews showed that the proposed method could classify feedback documents more correctly than the previous method, and that fair reputations are useful for evaluating reputations

This paper presents a method for acquiring a strategy of an agent in multi-issue negotiation. This method learns how to make a concession to an opponent for realizing win-win negotiation. To learn the concession strategy, we adopt reinforcement learning. First, an agent receives a proposal from an opponent. The agent recognizes a negotiation state using the difference between their proposals and difference between their concessions. According to the state, the agent makes a proposal by reinforcement learning. A reward of the learning is a profit of an agreement and punishment of negotiation breakdown. The experimental results showed that agents could acquire a negotiation strategy that avoids negotiation breakdown and increases profits of an agreement. As a result, agents can acquire the action policy that strikes a balance between cooperation and competition

For efficient video data management, 'video data mining' is required to discover 'semantic patterns' which are not only previously unknown and interesting, but also associated with semantically relevant events ('semantic events') in movies. In order to extract semantic patterns from a movie, we firstly represent it as a multi-stream of raw level metadata that abstracts the semantic information of the movie. Then, regarding to the temporal characteristic of the semantic event of the movie, we extract sequential patterns which are obtained by connecting temporally close and strongly associated symbols in the multi-stream of raw level metadata. We also propose a parallel data mining method in order to reduce the expensive computational cost. Finally, we verify whether the extracted patterns can be considered as semantic patterns or not.

The task to discover useful editing patterns from a professional video, such as a movie, is one of the main purposes of video data mining. These patterns successfully convey an editor's intentions to the viewer. However, data mining on multimedia data like a movie is a challenging task, due to the complicated contents of multimedia data. Particularly, the discovered patterns need to be supported by their semantic features, because these features tell amateur editors how to use the corresponding patterns during the process of editing a new video. We focus on the rhythm in a movie, consisting of the durations of a target character's appearance and disappearance. Based on this rhythm, we divide the movie into topics. Each topic corresponds to one meaningful episode of the character. By investigating such topics, we can discover useful editing patterns of a character's rhythm, supported by their semantic features. Also, these rhythms can be used to annotate certain types of topics.

In this paper, we propose a video shooting navigation system by real-time useful shot discrimination based on video grammar to support users shooting nice shots for later editing work. In this system, the processing speed must be very high so that we use a gray value projection method to extract the camerawork parameters in real-time. From the result of camerawork and gray value analysis, the shots are classified into 16 states and the system issues alarms and instructions about the usefulness and uselessness of the shots in real-time just after shooting the shot. Thereby, users can retake a shot necessary for later editing efficiently

Soon all people will have mobile terminals and they will collect information based on their interests and purposes in usual situations. We propose the method of information delivery for communications in temporary groups which are built in mobile environments in this paper. In our proposed method, profiles of a temporary group are automatically generated from contents which are stored into mobile terminals, and then, information is delivered to a temporary group based on profiles. It is possible to support communication in mobile environments by, using our proposed method.

This paper proposes a new kind of visual marker for a wearable computer based augmented reality (AR) system. Various marker systems have already been developed for using with an AR system. However, the improvement is needed to use these marker systems on a wearable AR system in terms of the amount of information or the width of the recognizable area. The main feature of our visual marker is a large amount of information that it can possess. Our marker consists of 32 bits data area with error detection code. This feature enables a wearable AR system to be used in various ways.

In recent years, digital video is rapidly becoming important for education, entertainment, and a host of multimedia applications. With the increasing size of video collections, technology is needed to effectively browse a video in a short time without losing any important contents. We propose a system which automatically summarizes video by analyzing the video color information and utterance information. To do so, we use the color histogram of a shot as color information. We discover important intervals having several color change patterns by using the probability model. Furthermore, we extract utterance information by using closed caption. We find the dialogue structure by analyzing the connectivity between utterances. Finally, we integrate them so as to skim meaningful portions in a video.

Data mining is a technique to discover useful patterns or patterns of special interest as explicit knowledge from a vast quantity of data. In video editing, there are a lot of editing patterns. According to the editor's preferences, different editing patterns give the opportunity to achieve a variety of effects. Discovering the editing patterns is required, because it is useful to find each editor's skills and to use them for editing new video material. In this paper, we propose the methods of extracting editing rules from video stream by introducing data mining technique. We can edit a video material by applying the extracted rules. The edited video may produce the same quality as the video from which we extracted the patterns.

Current video database systems are able to manipulate syntactic aspects of temporal media like video and audio. However, few of these systems are able to manipulate video information according to its semantic content. Furthermore, manually describing video content is tedious and not practical for the video data due to the temporal, spatial, and massive characteristic of the video, meanwhile content description always depends on user interpretation. Our video description model is based on "object-oriented model based on prototypes" which is proposed in the object-oriented programming field. A prototype is a prefabricated object with a predefined structure, content and behavior from which new objects can be constructed from other prototypes. By using prototypes, the user is able to describe video stories easily implementing and modifying prototypes of typical scenes that usually appear in several stories. Our method uses two hierarchical structures, one based on prototypes' inheritance structure and another based on prototype internal structure. From this point of view, our system includes a hierarchical prototype menu, which guides the user to reach easily requested prototypes; hence, the story remains the same regardless of different user interpretation.

No abstract available.