A reference based analysis framework for understanding anomaly detection techniques for symbolic sequences

Anomaly detection for symbolic sequence data is a highly important area of research and is relevant in many application domains. While several techniques have been proposed within different domains, understanding of their relative strengths and weaknesses is limited. The key factor for this is that the nature of sequence data varies significantly across domains, and hence while a technique might perform well in its original domain, its performance is not guaranteed in a different domain. In this paper, we aim at establishing this understanding for a wide variety of anomaly detection techniques for symbolic sequences. We present a comparative evaluation of a large number of anomaly detection techniques on a variety of publicly available as well as artificially generated data sets. Many of these are existing techniques while some are slight variants and/or adaptations of traditional anomaly detection techniques to sequence data. The analysis presented in this paper allows relative comparison of the different anomaly detection techniques and highlights their strengths and weaknesses. We extend the reference based analysis (RBA) framework, which was originally proposed to analyze multivariate categorical data, to analyze symbolic sequence data sets. We visualize the symbolic sequences using the characteristics provided by the RBA framework and use the visualization to understand various aspects of the sequence data. We then use the characterization done by RBA to understand the performance of the different techniques. Using the RBA framework, we propose two anomaly detection techniques for symbolic sequences, which show consistently superior performance over the existing techniques across the different data sets.     

Animation toolkit based on a database approach for reusing motions and models

As animations become more readily available, simultaneously the complexity of creating animations has also increased. In this paper, we address the issue by describing an animation toolkit based on a database approach for reusing geometric animation models and their motion sequences. The aim of our approach is to create a framework aimed for novice animators. Here, we use an alternative notion of a VRML scene graph to describe a geometric model, specifically intended for reuse. We represent this scene graph model as a relational database. A set of spatial, temporal, and motion operations are then used to manipulate the models and motions in an animation database. Spatial operations help in inserting/deleting geometric models in a new animation scene. Temporal and motion operations help in generating animation sequences in a variety of ways. For instance, motion information of one geometric model can be applied to another model or a motion sequence can be retargeted to meet additional constraints (e.g., wiping action on a table can be retargeted with constraints that reduce the size of the table). We present the design and implementation of this toolkit along with several interesting examples of animation sequences that can be generated using this toolkit.     

Unsupervised Activity Perception in Crowded and Complicated Scenes Using Hierarchical Bayesian Models

We propose a novel unsupervised learning framework to model activities and interactions in crowded and complicated scenes. Hierarchical Bayesian models are used to connect three elements in visual surveillance: low-level visual features, simple "atomic" activities, and interactions. Atomic activities are modeled as distributions over low-level visual features, and multi-agent interactions are modeled as distributions over atomic activities. These models are learnt in an unsupervised way. Given a long video sequence, moving pixels are clustered into different atomic activities and short video clips are clustered into different interactions. In this paper, we propose three hierarchical Bayesian models, Latent Dirichlet Allocation (LDA) mixture model, Hierarchical Dirichlet Process (HDP) mixture model, and Dual Hierarchical Dirichlet Processes (Dual-HDP) model. They advance existing language models, such as LDA [1] and HDP [2]. Our data sets are challenging video sequences from crowded traffic scenes and train station scenes with many kinds of activities co-occurring. Without tracking and human labeling effort, our framework completes many challenging visual surveillance tasks of board interest such as: (1) discovering typical atomic activities and interactions; (2) segmenting long video sequences into different interactions; (3) segmenting motions into different activities; (4) detecting abnormality; and (5) supporting high-level queries on activities and interactions.     

A case study in mapping conceptual designs to object‐relational schemas

The recent emergence of object‐relational technology into the commercial database market has caused new challenges for the implementation of conceptual database designs. This paper presents our experience with using the Oracle 8 object‐relational data model in the implementation of an engineering application described using the EXPRESS conceptual modeling language. EXPRESS is part of the engineering community's Standard for the Exchange of Product Data and can be characterized as a structurally object‐oriented modeling language, supporting the notion of entities, entity hierarchies, complex constraints on entity hierarchies, relationships and inverse relationships between entities, and user‐defined types. As a result, EXPRESS provides an excellent framework for studying the mapping of conceptual modeling concepts into an object‐relational model. In this paper, we describe the way in which the features of EXPRESS can be mapped into object‐relational features such as object tables, object references, and nested tables. We also describe the manner in which features such as member functions on object types, triggers, and stored procedures can be used to support the implementation of constraints associated with a conceptual schema. Although the mappings presented are specific to EXPRESS and Oracle 8, the mappings are generalizable to conceptual modeling languages and object‐relational models with similar features. Our work defines how traditional mapping concepts must be revised in order to make adequate use of the features now found in object‐relational models. As part of this paper, we also compare our mapping approach using Oracle 8 to mapping issues for the PostgreSQL object‐relational model and the Objectivity/DB object‐oriented data model. Copyright © 2000 John Wiley & Sons, Ltd.     

Online learning and fusion of orientation appearance models for robust rigid object tracking

We introduce a robust framework for learning and fusing of orientation appearance models based on both texture and depth information for rigid object tracking. Our framework fuses data obtained from a standard visual camera and dense depth maps obtained by low-cost consumer depth cameras such as the Kinect. To combine these two completely different modalities, we propose to use features that do not depend on the data representation: angles. More specifically, our framework combines image gradient orientations as extracted from intensity images with the directions of surface normals computed from dense depth fields. We propose to capture the correlations between the obtained orientation appearance models using a fusion approach motivated by the original Active Appearance Models (AAMs). To incorporate these features in a learning framework, we use a robust kernel based on the Euler representation of angles which does not require off-line training, and can be efficiently implemented online. The robustness of learning from orientation appearance models is presented both theoretically and experimentally in this work. This kernel enables us to cope with gross measurement errors, missing data as well as other typical problems such as illumination changes and occlusions. By combining the proposed models with a particle filter, the proposed framework was used for performing 2D plus 3D rigid object tracking, achieving robust performance in very difficult tracking scenarios including extreme pose variations.     

Efficient Relational Techniques for Processing Graph Queries

Graphs are widely used for modeling complicated data such as social networks, chemical compounds, protein interactions and semantic web. To effectively understand and utilize any collection of graphs, a graph database that efficiently supports elementary querying mechanisms is crucially required. For example, Subgraph and Supergraph queries are important types of graph queries which have many applications in practice. A primary challenge in computing the answers of graph queries is that pair-wise comparisons of graphs are usually hard problems. Relational database management systems (RDBMSs) have repeatedly been shown to be able to efficiently host different types of data such as complex objects and XML data. RDBMSs derive much of their performance from sophisticated optimizer components which make use of physical properties that are specific to the relational model such as sortedness, proper join ordering and powerful indexing mechanisms. In this article, we study the problem of indexing and querying graph databases using the relational infrastructure. We present a purely relational framework for processing graph queries. This framework relies on building a layer of graph features knowledge which capture metadata and summary features of the underlying graph database. We describe different querying mechanisms which make use of the layer of graph features knowledge to achieve scalable performance for processing graph queries. Finally, we conduct an extensive set of experiments on real and synthetic datasets to demonstrate the efficiency and the scalability of our techniques.     

Perceptual Organization Based Computational Model for Robust Segmentation of Moving Objects

The role of perceptual organization in motion analysis has heretofore been minimal. In this work we present a simple but powerful computational model and associated algorithms based on the use of perceptual organizational principles, such as temporal coherence (or common fate) and spatial proximity, for motion segmentation. The computational model does not use the traditional frame by frame motion analysis; rather it treats an image sequence as a single 3D spatio-temporal volume. It endeavors to find organizations in this volume of data over three levels—signal, primitive, and structural. The signal level is concerned with detecting individual image pixels that are probably part of a moving object. The primitive level groups these individual pixels into planar patches, which we call the temporal envelopes. Compositions of these temporal envelopes describe the spatio-temporal surfaces that result from object motion. At the structural level, we detect these compositions of temporal envelopes by utilizing the structure and organization among them. The algorithms employed to realize the computational model include 3D edge detection, Hough transformation, and graph based methods to group the temporal envelopes based on Gestalt principles. The significance of the Gestalt relationships between any two temporal envelopes is expressed in probabilistic terms. One of the attractive features of the adopted algorithm is that it does not require the detection of special 2D features or the tracking of these features across frames. We demonstrate that even with simple grouping strategies, we can easily handle drastic illumination changes, occlusion events, and multiple moving objects, without the use of training and specific object or illumination models. We present results on a large variety of motion sequences to demonstrate this robustness.     

Data modeling in the NoSQL world

NoSQL systems have gained their popularity for many reasons, including the flexibility they provide in organizing data, as they relax the rigidity provided by the relational model and by the other structured models. This flexibility and the heterogeneity that has emerged in the area have led to a little use of traditional modeling techniques, as opposed to what has happened with databases for decades.In this paper, we argue how traditional notions related to data modeling can be useful in this context as well. Specifically, we propose NoAM (NoSQL Abstract Model), a novel abstract data model for NoSQL databases, which exploits the commonalities of various NoSQL systems. We also propose a database design methodology for NoSQL systems based on NoAM, with initial activities that are independent of the specific target system. NoAM is used to specify a system-independent representation of the application data and, then, this intermediate representation can be implemented in target NoSQL databases, taking into account their specific features. Overall, the methodology aims at supporting scalability, performance, and consistency, as needed by next-generation web applications.     

Occurrence Counting Analysis for the π-calculus

We propose an abstract interpretation-based analysis for automatically proving non-trivial properties of mobile systems of processes. We focus on properties relying on the number of occurrences of processes during computation sequences, such as mutual exclusion and non-exhaustion of resources.We design a non-standard semantics for the π-calculus in order to explicitly trace the origin of channels and to solve efficiently problems set by α-conversion and non-deterministic choices. We abstract this semantics into an approximate one. The use of a relational domain for counting the occurrences of processes allows us to prove quickly and efficiently properties such as mutual exclusion and non-exhaustion of resources. At last, dynamic partitioning allows us to detect some configurations by which no infinite computation sequences can pass.     

Scalable keyword search on large data streams

It is widely recognized that the integration of information retrieval (IR) and database (DB) techniques provides users with a broad range of high quality services. Along this direction, IR-styled m-keyword query processing over a relational database in an rdbms framework has been well studied. It finds all hidden interconnected tuple structures, for example connected trees that contain keywords and are interconnected by sequences of primary/foreign key relationships among tuples. A new challenging issue is how to monitor events that are implicitly interrelated over an open-ended relational data stream for a user-given m-keyword query. Such a relational data stream is a sequence of tuple insertion/deletion operations. The difficulty of the problem is related to the number of costly joins to be processed over time when tuples are inserted and/or deleted. Such cost is mainly affected by three parameters, namely, the number of keywords, the maximum size of interconnected tuple structures, and the complexity of the database schema when it is viewed as a schema graph. In this paper, we propose new approaches. First, we propose a novel algorithm to efficiently determine all the joins that need to be processed for answering an m-keyword query. Second, we propose a new demand-driven approach to process such a query over a high speed relational data stream. We show that we can achieve high efficiency by significantly reducing the number of intermediate results when processing joins over a relational data stream. The proposed new techniques allow us to achieve high scalability in terms of both query plan generation and query plan execution. We conducted extensive experimental studies using synthetic data and real data to simulate a relational data stream. Our approach significantly outperforms existing algorithms.     

Inverse Procedural Modelling of Trees

Procedural tree models have been popular in computer graphics for their ability to generate a variety of output trees from a set of input parameters and to simulate plant interaction with the environment for a realistic placement of trees in virtual scenes. However, defining such models and their parameters is a difficult task. We propose an inverse modelling approach for stochastic trees that takes polygonal tree models as input and estimates the parameters of a procedural model so that it produces trees similar to the input. Our framework is based on a novel parametric model for tree generation and uses Monte Carlo Markov Chains to find the optimal set of parameters. We demonstrate our approach on a variety of input models obtained from different sources, such as interactive modelling systems, reconstructed scans of real trees and developmental models.     

Markov constraints: steerable generation of Markov sequences

Markov chains are a well known tool to model temporal properties of many phenomena, from text structure to fluctuations in economics. Because they are easy to generate, Markovian sequences, i.e. temporal sequences having the Markov property, are also used for content generation applications such as text or music generation that imitate a given style. However, Markov sequences are traditionally generated using greedy, left-to-right algorithms. While this approach is computationally cheap, it is fundamentally unsuited for interactive control. This paper addresses the issue of generating steerable Markovian sequences. We target interactive applications such as games, in which users want to control, through simple input devices, the way the system generates a Markovian sequence, such as a text, a musical sequence or a drawing. To this aim, we propose to revisit Markov sequence generation as a branch and bound constraint satisfaction problem (CSP). We propose a CSP formulation of the basic Markovian hypothesis as elementary Markov Constraints (EMC). We propose algorithms that achieve domain-consistency for the propagators of EMCs, in an event-based implementation of CSP. We show how EMCs can be combined to estimate the global Markovian probability of a whole sequence, and accommodate for different species of Markov generation such as fixed order, variable-order, or smoothing. Such a formulation, although more costly than traditional greedy generation algorithms, yields the immense advantage of being naturally steerable, since control specifications can be represented by arbitrary additional constraints, without any modification of the generation algorithm. We illustrate our approach on simple yet combinatorial chord sequence and melody generation problems and give some performance results.     

Maximum entropy direct models for speech recognition

Traditional statistical models for speech recognition have mostly been based on a Bayesian framework using generative models such as hidden Markov models (HMMs). This paper focuses on a new framework for speech recognition using maximum entropy direct modeling, where the probability of a state or word sequence given an observation sequence is computed directly from the model. In contrast to HMMs, features can be asynchronous and overlapping. This model therefore allows for the potential combination of many different types of features, which need not be statistically independent of each other. In this paper, a specific kind of direct model, the maximum entropy Markov model (MEMM), is studied. Even with conventional acoustic features, the approach already shows promising results for phone level decoding. The MEMM significantly outperforms traditional HMMs in word error rate when used as stand-alone acoustic models. Preliminary results combining the MEMM scores with HMM and language model scores show modest improvements over the best HMM speech recognizer.     

Fusion of audio and motion information on HMM-based highlight extraction for baseball games

This paper aims to extract baseball game highlights based on audio-motion integrated cues. In order to better describe different audio and motion characteristics in baseball game highlights, we propose a novel representation method based on likelihood models. The proposed likelihood models measure the "likeliness" of low-level audio features and motion features to a set of predefined audio types and motion categories, respectively. Our experiments show that using the proposed likelihood representation is more robust than using low-level audio/motion features to extract the highlight. With the proposed likelihood models, we then construct an integrated feature representation by symmetrically fusing the audio and motion likelihood models. Finally, we employ a hidden Markov model (HMM) to model and detect the transition of the integrated representation for highlight segments. A series of experiments have been conducted on a 12-h video database to demonstrate the effectiveness of our proposed method and show that the proposed framework achieves promising results over a variety of baseball game sequences.     

Thermo-visual feature fusion for object tracking using multiple spatiogram trackers

In this paper, we propose a framework that can efficiently combine features for robust tracking based on fusing the outputs of multiple spatiogram trackers. This is achieved without the exponential increase in storage and processing that other multimodal tracking approaches suffer from. The framework allows the features to be split arbitrarily between the trackers, as well as providing the flexibility to add, remove or dynamically weight features. We derive a mean-shift type algorithm for the framework that allows efficient object tracking with very low computational overhead. We especially target the fusion of thermal infrared and visible spectrum features as the most useful features for automated surveillance applications. Results are shown on multimodal video sequences clearly illustrating the benefits of combining multiple features using our framework.     

A bias/variance decomposition for models using collective inference

Bias/variance analysis is a useful tool for investigating the performance of machine learning algorithms. Conventional analysis decomposes loss into errors due to aspects of the learning process, but in relational domains, the inference process used for prediction introduces an additional source of error. Collective inference techniques introduce additional error, both through the use of approximate inference algorithms and through variation in the availability of test-set information. To date, the impact of inference error on model performance has not been investigated. We propose a new bias/variance framework that decomposes loss into errors due to both the learning and inference processes. We evaluate the performance of three relational models on both synthetic and real-world datasets and show that (1) inference can be a significant source of error, and (2) the models exhibit different types of errors as data characteristics are varied.     

Asynchronous communication model based on linear logic

We propose a new framework called ACL for concurrent computation based on linear logic. ACL is a kind oflinear logic programming framework, where its operational semantics is described in terms ofproof construction in linear logic. We also give a model-theoretic semantics based onphase semantics, a model of linear logic. Our framework well captures concurrent computation based on asynchronous communication. It will, therefore, provide us with a new insight into other models of asynchronous concurrent computation from alogical point of view. We also expect ACL to become a formal framework for analysis, synthesis and transformation of concurrent programs by the use of techniques for traditional logic programming. ACL's attractive features for concurrent programming paradigms are also discussed.     

An Algebraic Baseline for Automatic Transformations in MDA

Software evolution can be supported at two levels: models and programs. The model-based software development approach allows the application of a more abstract process of software evolution, in accordance with the OMG's MDA initiative. We describe a framework for model management, called MOMENT, that supports automatic formal model transformations in MDA. Our model transformation approach is based on the algebraic specification of models and benefits from mature term rewriting system technology to perform model transformation using rewriting logic. In this paper, we present how we apply this formal transformation mechanism between platformindependent models, such as UML models and relational schemas. Our approach enhances the integration between formal environments and industrial technologies such as .NET technology, and exploits the best features of both.     

An inductive database system based on virtual mining views

Inductive databases integrate database querying with database mining. In this article, we present an inductive database system that does not rely on a new data mining query language, but on plain SQL. We propose an intuitive and elegant framework based on virtual mining views, which are relational tables that virtually contain the complete output of data mining algorithms executed over a given data table. We show that several types of patterns and models that are implicitly present in the data, such as itemsets, association rules, and decision trees, can be represented and queried with SQL using a unifying framework. As a proof of concept, we illustrate a complete data mining scenario with SQL queries over the mining views, which is executed in our system.