Eigenvoice modeling for cross likelihood ratio based speaker clustering : a Bayesian approach

This paper proposes the use of Bayesian approaches with the cross likelihood ratio (CLR) as a criterion for speaker clustering within a speaker diarization system, using eigenvoice modeling techniques. The CLR has previously been shown to be an effective decision criterion for speaker clustering using Gaussian mixture models. Recently, eigenvoice modeling has become an increasingly popular technique, due to its ability to adequately represent a speaker based on sparse training data, as well as to provide an improved capture of differences in speaker characteristics. The integration of eigenvoice modeling into the CLR framework to capitalize on the advantage of both techniques has also been shown to be beneficial for the speaker clustering task. Building on that success, this paper proposes the use of Bayesian methods to compute the conditional probabilities in computing the CLR, thus effectively combining the eigenvoice-CLR framework with the advantages of a Bayesian approach to the diarization problem. Results obtained on the 2002 Rich Transcription (RT-02) Evaluation dataset show an improved clustering performance, resulting in a 33.5% relative improvement in the overall Diarization Error Rate (DER) compared to the baseline system.

A non-linear dynamic analysis of heavy haul track line sleepers under wheel flat

Significant wheel-rail dynamic forces occur because of imperfections in the wheels and/or rail. One of the key responses to the transmission of these forces down through the track is impact force on the sleepers. Dynamic analysis of nonlinear systems is very complicated and does not lend itself easily to a classical solution of multiple equations. Trying to deduce the behaviour of track components from experimental data is very difficult because such data is hard to obtain and applies to only the particular conditions of the track being tested. The finite element method can be the best solution to this dilemma. This paper describes a finite element model using the software package ANSYS for various sized flat defects in the tread of a wheel rolling at a typical speed on heavy haul track. The paper explores the dynamic response of a prestressed concrete sleeper to these defects.

A dynamic wheel-rail impact analysis of railway track under wheel flat by Finite Element Analysis

Wheel–rail interaction is one of the most important research topics in railway engineering. It involves track impact response, track vibration and track safety. Track structure failures caused by wheel–rail impact forces can lead to significant economic loss for track owners through damage to rails and to the sleepers beneath. Wheel–rail impact forces occur because of imperfections in the wheels or rails such as wheel flats, irregular wheel profiles, rail corrugations and differences in the heights of rails connected at a welded joint. A wheel flat can cause a large dynamic impact force as well as a forced vibration with a high frequency, which can cause damage to the track structure. In the present work, a three-dimensional (3-D) finite element (FE) model for the impact analysis induced by the wheel flat is developed by use of the finite element analysis (FEA) software package ANSYS and validated by another validated simulation. The effect of wheel flats on impact forces is thoroughly investigated. It is found that the presence of a wheel flat will significantly increase the dynamic impact force on both rail and sleeper. The impact force will monotonically increase with the size of wheel flats. The relationships between the impact force and the wheel flat size are explored from this finite element analysis and they are important for track engineers to improve their understanding of the design and maintenance of the track system.

Speaker diarization : "who spoke when"

Speaker diarization is the process of annotating an input audio with information that attributes temporal regions of the audio signal to their respective sources, which may include both speech and non-speech events. For speech regions, the diarization system also specifies the locations of speaker boundaries and assign relative speaker labels to each homogeneous segment of speech. In short, speaker diarization systems effectively answer the question of ‘who spoke when’. There are several important applications for speaker diarization technology, such as facilitating speaker indexing systems to allow users to directly access the relevant segments of interest within a given audio, and assisting with other downstream processes such as summarizing and parsing. When combined with automatic speech recognition (ASR) systems, the metadata extracted from a speaker diarization system can provide complementary information for ASR transcripts including the location of speaker turns and relative speaker segment labels, making the transcripts more readable. Speaker diarization output can also be used to localize the instances of specific speakers to pool data for model adaptation, which in turn boosts transcription accuracies. Speaker diarization therefore plays an important role as a preliminary step in automatic transcription of audio data. The aim of this work is to improve the usefulness and practicality of speaker diarization technology, through the reduction of diarization error rates. In particular, this research is focused on the segmentation and clustering stages within a diarization system. Although particular emphasis is placed on the broadcast news audio domain and systems developed throughout this work are also trained and tested on broadcast news data, the techniques proposed in this dissertation are also applicable to other domains including telephone conversations and meetings audio. Three main research themes were pursued: heuristic rules for speaker segmentation, modelling uncertainty in speaker model estimates, and modelling uncertainty in eigenvoice speaker modelling. The use of heuristic approaches for the speaker segmentation task was first investigated, with emphasis placed on minimizing missed boundary detections. A set of heuristic rules was proposed, to govern the detection and heuristic selection of candidate speaker segment boundaries. A second pass, using the same heuristic algorithm with a smaller window, was also proposed with the aim of improving detection of boundaries around short speaker segments. Compared to single threshold based methods, the proposed heuristic approach was shown to provide improved segmentation performance, leading to a reduction in the overall diarization error rate. Methods to model the uncertainty in speaker model estimates were developed, to address the difficulties associated with making segmentation and clustering decisions with limited data in the speaker segments. The Bayes factor, derived specifically for multivariate Gaussian speaker modelling, was introduced to account for the uncertainty of the speaker model estimates. The use of the Bayes factor also enabled the incorporation of prior information regarding the audio to aid segmentation and clustering decisions. The idea of modelling uncertainty in speaker model estimates was also extended to the eigenvoice speaker modelling framework for the speaker clustering task. Building on the application of Bayesian approaches to the speaker diarization problem, the proposed approach takes into account the uncertainty associated with the explicit estimation of the speaker factors. The proposed decision criteria, based on Bayesian theory, was shown to generally outperform their non- Bayesian counterparts.

Modelling users' contextual querying behaviour for web image searching

The rapid growth of visual information on Web has led to immense interest in multimedia information retrieval (MIR). While advancement in MIR systems has achieved some success in specific domains, particularly the content-based approaches, general Web users still struggle to find the images they want. Despite the success in content-based object recognition or concept extraction, the major problem in current Web image searching remains in the querying process. Since most online users only express their needs in semantic terms or objects, systems that utilize visual features (e.g., color or texture) to search images create a semantic gap which hinders general users from fully expressing their needs. In addition, query-by-example (QBE) retrieval imposes extra obstacles for exploratory search because users may not always have the representative image at hand or in mind when starting a search (i.e. the page zero problem). As a result, the majority of current online image search engines (e.g., Google, Yahoo, and Flickr) still primarily use textual queries to search. The problem with query-based retrieval systems is that they only capture users’ information need in terms of formal queries;; the implicit and abstract parts of users’ information needs are inevitably overlooked. Hence, users often struggle to formulate queries that best represent their needs, and some compromises have to be made. Studies of Web search logs suggest that multimedia searches are more difficult than textual Web searches, and Web image searching is the most difficult compared to video or audio searches. Hence, online users need to put in more effort when searching multimedia contents, especially for image searches. Most interactions in Web image searching occur during query reformulation. While log analysis provides intriguing views on how the majority of users search, their search needs or motivations are ultimately neglected. User studies on image searching have attempted to understand users’ search contexts in terms of users’ background (e.g., knowledge, profession, motivation for search and task types) and the search outcomes (e.g., use of retrieved images, search performance). However, these studies typically focused on particular domains with a selective group of professional users. General users’ Web image searching contexts and behaviors are little understood although they represent the majority of online image searching activities nowadays. We argue that only by understanding Web image users’ contexts can the current Web search engines further improve their usefulness and provide more efficient searches. In order to understand users’ search contexts, a user study was conducted based on university students’ Web image searching in News, Travel, and commercial Product domains. The three search domains were deliberately chosen to reflect image users’ interests in people, time, event, location, and objects. We investigated participants’ Web image searching behavior, with the focus on query reformulation and search strategies. Participants’ search contexts such as their search background, motivation for search, and search outcomes were gathered by questionnaires. The searching activity was recorded with participants’ think aloud data for analyzing significant search patterns. The relationships between participants’ search contexts and corresponding search strategies were discovered by Grounded Theory approach. Our key findings include the following aspects: - Effects of users' interactive intents on query reformulation patterns and search strategies - Effects of task domain on task specificity and task difficulty, as well as on some specific searching behaviors - Effects of searching experience on result expansion strategies A contextual image searching model was constructed based on these findings. The model helped us understand Web image searching from user perspective, and introduced a context-aware searching paradigm for current retrieval systems. A query recommendation tool was also developed to demonstrate how users’ query reformulation contexts can potentially contribute to more efficient searching.

Determining process model precision and generalization with weighted artificial negative events

Process mining encompasses the research area which is concerned with knowledge discovery from event logs. One common process mining task focuses on conformance checking, comparing discovered or designed process models with actual real-life behavior as captured in event logs in order to assess the “goodness” of the process model. This paper introduces a novel conformance checking method to measure how well a process model performs in terms of precision and generalization with respect to the actual executions of a process as recorded in an event log. Our approach differs from related work in the sense that we apply the concept of so-called weighted artificial negative events towards conformance checking, leading to more robust results, especially when dealing with less complete event logs that only contain a subset of all possible process execution behavior. In addition, our technique offers a novel way to estimate a process model’s ability to generalize. Existing literature has focused mainly on the fitness (recall) and precision (appropriateness) of process models, whereas generalization has been much more difficult to estimate. The described algorithms are implemented in a number of ProM plugins, and a Petri net conformance checking tool was developed to inspect process model conformance in a visual manner.

The suitability of using accumulated plastic strain to assess the damage at the rail-wheel interfaces

Numerous Abaqus [1] finite element analyses have been carried out using various plasticity models to investigate the effect of friction force on the rail head in relation to both the development of the accumulated plastic strain (PEEQ) and the changes in the depth of PEEQ distribution in the wheel-rail contact. The normal force distribution on the rail head was assumed to be Hertzian. The tangential force was implemented as a fraction of the normal force in the subroutine. Each analysis was carried out for a single pass and the effect of various friction coefficient values has been observed.

Second-order elastic finite element analysis of steel structures using a single element per member

Finite element frame analysis programs targeted for design office application necessitate algorithms which can deliver reliable numerical convergence in a practical timeframe with comparable degrees of accuracy, and a highly desirable attribute is the use of a single element per member to reduce computational storage, as well as data preparation and the interpretation of the results. To this end, a higher-order finite element method including geometric non-linearity is addressed in the paper for the analysis of elastic frames for which a single element is used to model each member. The geometric non-linearity in the structure is handled using an updated Lagrangian formulation, which takes the effects of the large translations and rotations that occur at the joints into consideration by accumulating their nodal coordinates. Rigid body movements are eliminated from the local member load-displacement relationship for which the total secant stiffness is formulated for evaluating the large member deformations of an element. The influences of the axial force on the member stiffness and the changes in the member chord length are taken into account using a modified bowing function which is formulated in the total secant stiffness relationship, for which the coupling of the axial strain and flexural bowing is included. The accuracy and efficiency of the technique is verified by comparisons with a number of plane and spatial structures, whose structural response has been reported in independent studies.

Second-order elastic finite element analysis of structures using a single element per member

Finite element frame analysis programs targeted for design office application necessitate algorithms which can deliver reliable numerical convergence in a practical timeframe with comparable degrees of accuracy, and a highly desirable attribute is the use of a single element per member to reduce computational storage, as well as data preparation and the interpretation of the results. To this end, a higher-order finite element method including geometric non-linearity is addressed in the paper for the analysis of elastic frames for which a single element is used to model each member. The geometric non-linearity in the structure is handled using an updated Lagrangian formulation, which takes the effects of the large translations and rotations that occur at the joints into consideration by accumulating their nodal coordinates. Rigid body movements are eliminated from the local member load-displacement relationship for which the total secant stiffness is formulated for evaluating the large member deformations of an element. The influences of the axial force on the member stiffness and the changes in the member chord length are taken into account using a modified bowing function which is formulated in the total secant stiffness relationship, for which the coupling of the axial strain and flexural bowing is included.

Manoeuvring control of underactuated surface vessels using manifold regulation for Port-Hamiltonian systems

This paper considers the manoeuvring of underactuated surface vessels. The control objective is to steer the vessel to reach a manifold which encloses a waypoint. A transformation of configuration variables and a potential field are used in a Port-Hamiltonian framework to design an energy-based controller. With the proposed controller, the geometric task associated with the manoeuvring problem depends on the desired potential energy (closed-loop) and the dynamic task depends on the total energy and damping. Therefore, guidance and motion control are addressed jointly, leading to model-energy-based trajectory generation.

The feasibility of a home-based moderate-intensity physical activity intervention in obese children and adolescents

Objective To explore the feasibility of conducting a 10-week home-based physical activity (PA) programme and evaluate the changes in insulin sensitivity (S I) commensurate with the programme in obese young people. Design Open-labelled intervention. Setting Home-based intervention with clinical assessments at a tertiary paediatric hospital. Subjects 18 obese (body mass index (BMI)>International Obesity Task Force age and sex-specifi c cut-offs) children and adolescents (8-18 years, 11 girls/7 boys) were recruited. 15 participants (nine girls/six boys, mean±SE age 11.8±0.6 years, BMI-SD scores (BMI-SDS) 3.5±0.1, six prepubertal/nine pubertal) completed the intervention. Intervention The programme comprised biweekly home visits over 10 weeks with personalised plans implemented aiming to increase moderate-intensity PA. Pedometers and PA diaries were used as self-monitoring tools. The goals were to (1) teach participants behavioural skills related to adopting and maintaining an active lifestyle and (2) increase daily participation in PA. Outcome measures Mean steps/day were assessed. SI assessed by the frequently sampled intravenous glucose tolerance test and other components of the insulin resistance syndrome were measured. Results Mean steps/day increased significantly from 10 363±927 (baseline) to 13 013±1131 (week 10) (p<0.05). S I was also significantly increased, despite no change in BMI-SDS, and remained so after an additional 10-week follow-up. Conclusions The results suggest that such a homebased PA programme is feasible. S I improved without changes in BMI-SDS. More rigorous evaluations of such programmes are warranted.

Enhanced fast handover triggering mechanism for fast proxy mobile IPv6

Handover performance is critical to support real-time traffic applications in wireless network communications. The longer the handover delay is, the longer an Mobile Node (MN) is prevented from sending and receiving any data packet. In real-time network communication applications, such as VoIP and video-conference, a long handover delay is often unacceptable. In order to achieve better handover performance, Fast Proxy Mobile IPv6 (FPMIPv6) has been standardised as an improvement to the original Proxy Mobile IPv6 (PMIPv6) in the Internet Engineering Task Force (IETF). The FPMIPv6 adopts a link layer triggering mechanism to perform two modes of operation: predictive and reactive modes. Using the link layer triggering, the handover performance of the FPMIPv6 can be improved in the predictive mode. However, an unsuccessful predictive handover operation will lead to activation of a reactive handover. In the reactive mode, MNs still experience long handover delays and a large amount of packet loss, which significantly degrade the handover performance of the FPMIPv6. Addressing this problem, this thesis presents an Enhanced Triggering Mechanism (ETM) in the FPMIPv6 to form an enhanced FPMIPv6 (eFPMIPv6). The ETM reduces the most time consuming processes in the reactive handover: the failed Handover Initiate (HO-Initiate) delay and bidirectional tunnel establishment delay. Consequently, the overall handover performance of the FPMIPv6 is enhanced in the eFPMIPv6. To show the advantages of the proposed eFPMIPv6, a theoretical analysis is carried out to mathematically model the performance of PMIPv6, FPMIPv6 and eFPMIPv6. Extensive case studies are conducted to validate the effectiveness of the presented eFPMIPv6 mechanism. They are carried out under various scenarios with changes in network link delay, traffic load, number of hops and MN moving velocity. The case studies show that the proposed mechanism ETM reduces the reactive handover delay, and the presented eFPMIPv6 outperforms the PMIPv6 and FPMIPv6 in terms of the overall handover performance.

Overview of the ShARe/CLEF eHealth Evaluation Lab 2014

This paper reports on the 2nd ShARe/CLEFeHealth evaluation lab which continues our evaluation resource building activities for the medical domain. In this lab we focus on patients' information needs as opposed to the more common campaign focus of the specialised information needs of physicians and other healthcare workers. The usage scenario of the lab is to ease patients and next-of-kins' ease in understanding eHealth information, in particular clinical reports. The 1st ShARe/CLEFeHealth evaluation lab was held in 2013. This lab consisted of three tasks. Task 1 focused on named entity recognition and normalization of disorders; Task 2 on normalization of acronyms/abbreviations; and Task 3 on information retrieval to address questions patients may have when reading clinical reports. This year's lab introduces a new challenge in Task 1 on visual-interactive search and exploration of eHealth data. Its aim is to help patients (or their next-of-kin) in readability issues related to their hospital discharge documents and related information search on the Internet. Task 2 then continues the information extraction work of the 2013 lab, specifically focusing on disorder attribute identification and normalization from clinical text. Finally, this year's Task 3 further extends the 2013 information retrieval task, by cleaning the 2013 document collection and introducing a new query generation method and multilingual queries. De-identified clinical reports used by the three tasks were from US intensive care and originated from the MIMIC II database. Other text documents for Tasks 1 and 3 were from the Internet and originated from the Khresmoi project. Task 2 annotations originated from the ShARe annotations. For Tasks 1 and 3, new annotations, queries, and relevance assessments were created. 50, 79, and 91 people registered their interest in Tasks 1, 2, and 3, respectively. 24 unique teams participated with 1, 10, and 14 teams in Tasks 1, 2 and 3, respectively. The teams were from Africa, Asia, Canada, Europe, and North America. The Task 1 submission, reviewed by 5 expert peers, related to the task evaluation category of Effective use of interaction and targeted the needs of both expert and novice users. The best system had an Accuracy of 0.868 in Task 2a, an F1-score of 0.576 in Task 2b, and Precision at 10 (P@10) of 0.756 in Task 3. The results demonstrate the substantial community interest and capabilities of these systems in making clinical reports easier to understand for patients. The organisers have made data and tools available for future research and development.

Behaviour of corrugated steel roof cladding under combined wind uplift and racking

During high wind events, crest-fixed profiled steel roof claddings in low-rise buildings can be subjected to combined cyclic wind uplift and in-plane racking (shear) forces. Static and cyclic tests of corrugated steel roof claddings were carried out to investigate the effect of in-plane racking force on the uplift strength, in particular, in relation to the fatigue cracking commonly observed under cyclic wind uplift. The presence of racking force appeared to have insignificant effect on the static and cyclic wind uplift strength. It may therefore be possible to include the diaphragm strength of these claddings in the design of low-rise buildings in a similar manner to valley-fixed claddings. This may lead to a reduction in bracing requirements.