Topic transition detection using hierarchical hidden Markov and semi-Markov models

In this paper we introduce a probabilistic framework to exploit hierarchy, structure sharing and duration information for topic transition detection in videos. Our probabilistic detection framework is a combination of a shot classification step and a detection phase using hierarchical probabilistic models. We consider two models in this paper: the extended Hierarchical Hidden Markov Model (HHMM) and the Coxian Switching Hidden semi-Markov Model (S-HSMM) because they allow the natural decomposition of semantics in videos, including shared structures, to be modeled directly, and thus enabling efficient inference and reducing the sample complexity in learning. Additionally, the S-HSMM allows the duration information to be incorporated, consequently the modeling of long-term dependencies in videos is enriched through both hierarchical and duration modeling. Furthermore, the use of the Coxian distribution in the S-HSMM makes it tractable to deal with long sequences in video. Our experimentation of the proposed framework on twelve educational and training videos shows that both models outperform the baseline cases (flat HMM and HSMM) and performances reported in earlier work in topic detection. The superior performance of the S-HSMM over the HHMM verifies our belief that duration information is an important factor in video content modeling.

Polymorphic malware detection using Hierarchical Hidden Markov Model

Binary signatures have been widely used to detect malicious software on the current Internet. However, this approach is unable to achieve the accurate identification of polymorphic malware variants, which can be easily generated by the malware authors using code generation engines. Code generation engines randomly produce varying code sequences but perform the same desired malicious functions. Previous research used flow graph and signature tree to identify polymorphic malware families. The key difficulty of previous research is the generation of precisely defined state machine models from polymorphic variants. This paper proposes a novel approach, using Hierarchical Hidden Markov Model (HHMM), to provide accurate inductive inference of the malware family. This model can capture the features of self-similar and hierarchical structure of polymorphic malware family signature sequences. To demonstrate the effectiveness and efficiency of this approach, we evaluate it with real malware samples. Using more than 15,000 real malware, we find our approach can achieve high true positives, low false positives, and low computational cost.

Discovery of activity structures using the hierarchical hidden Markov model

In building a surveillance system for monitoring people behaviours, it is important to understand the typical patterns of people's movement in the environment. This task is difficult when dealing with high-level behaviours. The flat model such as the hidden Markov model (HMM) is inefficient in differentiating between signatures of such behaviours. This paper examines structure learning for high-level behaviours using the hierarchical hidden Markov model (HHMM).We propose a two-phase learning algorithm in which the parameters of the behaviours at low levels are estimated first and then the structures and parameters of the behaviours at high levels are learned from multi-camera training data. Our algorithm is then evaluated using data from a real environment, demonstrating the robustness of the learned structure in recognising people's behaviour.

Hierarchical summarization techniques for network traffic

In today’s high speed networks it is becoming increasingly challenging for network managers to understand the nature of the traffic that is carried in their network. A major problem for traffic analysis in this context is how to extract a concise yet accurate summary of the relevant aggregate traffic flows that are present in network traces. In this paper, we present two summarization techniques to minimize the size of the traffic flow report that is generated by a hierarchical cluster analysis tool. By analyzing the accuracy and compaction gain of our approach on a standard benchmark dataset, we demonstrate that our approach achieves more accurate summaries than those of an existing tool that is based on frequent itemset mining.

Controlled synthesis of TiO2 hierarchical nanofibre structures via electrospinning and solvothermal processes : photocatalytic activity for degradation of methylene blue

The present article describes a new titanium oxide‐based (TiO2) photocatalyst that shows promise for acceleration of dye degradation. A hierarchical TiO2 nanostructure comprising nanorods on‐nanofibres has been prepared using a sol–gel route and electrospinning. Calcination of electrospun nanobre mats was performed in air at 500 °C. The TiO2 nanofibre surface was then exploited as a ‘seeding ground’ to grow TiO2 nanorods by a solvothermal process in NaOH. The nanofibres had a diameter of approximately 100 nm while the nanorods were evenly distributed on the nanofibre surface with a mean diameter of around 50–80 nm. The hierarchical nanostructure showed enhanced photocatalytic activity when compared to pure TiO2 nanofibres. This improved efficiency in degrading methylene blue through the photocatalytic process was attributed to the larger specific surface area of the TiO2 nanostructures, as well as high surface‐to‐volume ratio and higher reactive surface resulting in enhanced surface adsorption and interfacial redox reaction.

Document clustering in correlation similarity measure space

This paper presents a new spectral clustering method called correlation preserving indexing (CPI), which is performed in the correlation similarity measure space. In this framework, the documents are projected into a low-dimensional semantic space in which the correlations between the documents in the local patches are maximized while the correlations between the documents outside these patches are minimized simultaneously. Since the intrinsic geometrical structure of the document space is often embedded in the similarities between the documents, correlation as a similarity measure is more suitable for detecting the intrinsic geometrical structure of the document space than euclidean distance. Consequently, the proposed CPI method can effectively discover the intrinsic structures embedded in high-dimensional document space. The effectiveness of the new method is demonstrated by extensive experiments conducted on various data sets and by comparison with existing document clustering methods.

Hierarchical workflow management in wireless sensor network

To build the service-oriented applications in a wireless sensor network (WSN), the workflow can be utilized to compose a set of atomic services and execute the corresponding pre-designed processes. In general, WSN applications rely closely on the sensor data which are usually inaccurate or even incomplete in the resource-constrained WSN. Then, the erroneous sensor data will affect the execution of atomic services and furthermore the workflows, which form an important part in the bottom-to-up dynamics of WSN applications. In order to alleviate this issue, it is necessary to manage the workflow hierarchically. However, the hierarchical workflow management remains an open and challenging problem. In this paper, by adopting the Bloom filter as an effective connection between the sensor node layer and the upper application layer, a hierarchical workflow management approach is proposed to ensure the QoS of workflow-based WSN application . The case study and experimental evaluations demonstrate the capability of the proposed approach.

Sparse subspace representation for spectral document clustering

We present a novel method for document clustering using sparse representation of documents in conjunction with spectral clustering. An ℓ1-norm optimization formulation is posed to learn the sparse representation of each document, allowing us to characterize the affinity between documents by considering the overall information instead of traditional pair wise similarities. This document affinity is encoded through a graph on which spectral clustering is performed. The decomposition into multiple subspaces allows documents to be part of a sub-group that shares a smaller set of similar vocabulary, thus allowing for cleaner clusters. Extensive experimental evaluations on two real-world datasets from Reuters-21578 and 20Newsgroup corpora show that our proposed method consistently outperforms state-of-the-art algorithms. Significantly, the performance improvement over other methods is prominent for this datasets.

A slice sampler for restricted hierarchical beta process with applications to shared subspace learning

Hierarchical beta process has found interesting applications in recent years. In this paper we present a modified hierarchical beta process prior with applications to hierarchical modeling of multiple data sources. The novel use of the prior over a hierarchical factor model allows factors to be shared across different sources. We derive a slice sampler for this model, enabling tractable inference even when the likelihood and the prior over parameters are non-conjugate. This allows the application of the model in much wider contexts without restrictions. We present two different data generative models – a linear Gaussian-Gaussian model for real valued data and a linear Poisson-gamma model for count data. Encouraging transfer learning results are shown for two real world applications – text modeling and content based image retrieval.