EmotiBlog: a finer-grained and more precise learning of subjectivity expression models

The exponential growth of the subjective information in the framework of the Web 2.0 has led to the need to create Natural Language Processing tools able to analyse and process such data for multiple practical applications. They require training on specifically annotated corpora, whose level of detail must be fine enough to capture the phenomena involved. This paper presents EmotiBlog – a fine-grained annotation scheme for subjectivity. We show the manner in which it is built and demonstrate the benefits it brings to the systems using it for training, through the experiments we carried out on opinion mining and emotion detection. We employ corpora of different textual genres –a set of annotated reported speech extracted from news articles, the set of news titles annotated with polarity and emotion from the SemEval 2007 (Task 14) and ISEAR, a corpus of real-life self-expressed emotion. We also show how the model built from the EmotiBlog annotations can be enhanced with external resources. The results demonstrate that EmotiBlog, through its structure and annotation paradigm, offers high quality training data for systems dealing both with opinion mining, as well as emotion detection.

Integration of different models in the design of chemical processes: Application to the design of a power plant

With advances in the synthesis and design of chemical processes there is an increasing need for more complex mathematical models with which to screen the alternatives that constitute accurate and reliable process models. Despite the wide availability of sophisticated tools for simulation, optimization and synthesis of chemical processes, the user is frequently interested in using the ‘best available model’. However, in practice, these models are usually little more than a black box with a rigid input–output structure. In this paper we propose to tackle all these models using generalized disjunctive programming to capture the numerical characteristics of each model (in equation form, modular, noisy, etc.) and to deal with each of them according to their individual characteristics. The result is a hybrid modular–equation based approach that allows synthesizing complex processes using different models in a robust and reliable way. The capabilities of the proposed approach are discussed with a case study: the design of a utility system power plant that has been decomposed into its constitutive elements, each treated differently numerically. And finally, numerical results and conclusions are presented.

Finitely coordinated models for low-temperature phases of amorphous systems

We introduce models of heterogeneous systems with finite connectivity defined on random graphs to capture finite-coordination effects on the low-temperature behaviour of finite-dimensional systems. Our models use a description in terms of small deviations of particle coordinates from a set of reference positions, particularly appropriate for the description of low-temperature phenomena. A Born-von Karman-type expansion with random coefficients is used to model effects of frozen heterogeneities. The key quantity appearing in the theoretical description is a full distribution of effective single-site potentials which needs to be determined self-consistently. If microscopic interactions are harmonic, the effective single-site potentials turn out to be harmonic as well, and the distribution of these single-site potentials is equivalent to a distribution of localization lengths used earlier in the description of chemical gels. For structural glasses characterized by frustration and anharmonicities in the microscopic interactions, the distribution of single-site potentials involves anharmonicities of all orders, and both single-well and double-well potentials are observed, the latter with a broad spectrum of barrier heights. The appearance of glassy phases at low temperatures is marked by the appearance of asymmetries in the distribution of single-site potentials, as previously observed for fully connected systems. Double-well potentials with a broad spectrum of barrier heights and asymmetries would give rise to the well-known universal glassy low-temperature anomalies when quantum effects are taken into account. © 2007 IOP Publishing Ltd.

Adaptation and gain pool summation:Alternative models and masking data

Foley [J. Opt. Soc. Am. A 11 (1994) 1710] has proposed an influential psychophysical model of masking in which mask components in a contrast gain pool are raised to an exponent before summation and divisive inhibition. We tested this summation rule in experiments in which contrast detection thresholds were measured for a vertical 1 c/deg (or 2 c/deg) sine-wave component in the presence of a 3 c/deg (or 6 c/deg) mask that had either a single component oriented at -45° or a pair of components oriented at ±45°. Contrary to the predictions of Foley's model 3, we found that for masks of moderate contrast and above, threshold elevation was predicted by linear summation of the mask components in the inhibitory stage of the contrast gain pool. We built this feature into two new models, referred to as the early adaptation model and the hybrid model. In the early adaptation model, contrast adaptation controls a threshold-like nonlinearity on the output of otherwise linear pathways that provide the excitatory and inhibitory inputs to a gain control stage. The hybrid model involves nonlinear and nonadaptable routes to excitatory and inhibitory stages as well as an adaptable linear route. With only six free parameters, both models provide excellent fits to the masking and adaptation data of Foley and Chen [Vision Res. 37 (1997) 2779] but unlike Foley and Chen's model, are able to do so with only one adaptation parameter. However, only the hybrid model is able to capture the features of Foley's (1994) pedestal plus orthogonal fixed mask data. We conclude that (1) linear summation of inhibitory components is a feature of contrast masking, and (2) that the main aftereffect of spatial adaptation on contrast increment thresholds can be assigned to a single site. © 2002 Elsevier Science Ltd. All rights reserved.

GIS linked environmental process models

In recent years there has been a great effort to combine the technologies and techniques of GIS and process models. This project examines the issues of linking a standard current generation 2½d GIS with several existing model codes. The focus for the project has been the Shropshire Groundwater Scheme, which is being developed to augment flow in the River Severn during drought periods by pumping water from the Shropshire Aquifer. Previous authors have demonstrated that under certain circumstances pumping could reduce the soil moisture available for crops. This project follows earlier work at Aston in which the effects of drawdown were delineated and quantified through the development of a software package that implemented a technique which brought together the significant spatially varying parameters. This technique is repeated here, but using a standard GIS called GRASS. The GIS proved adequate for the task and the added functionality provided by the general purpose GIS - the data capture, manipulation and visualisation facilities - were of great benefit. The bulk of the project is concerned with examining the issues of the linkage of GIS and environmental process models. To this end a groundwater model (Modflow) and a soil moisture model (SWMS2D) were linked to the GIS and a crop model was implemented within the GIS. A loose-linked approach was adopted and secondary and surrogate data were used wherever possible. The implications of which relate to; justification of a loose-linked versus a closely integrated approach; how, technically, to achieve the linkage; how to reconcile the different data models used by the GIS and the process models; control of the movement of data between models of environmental subsystems, to model the total system; the advantages and disadvantages of using a current generation GIS as a medium for linking environmental process models; generation of input data, including the use of geostatistic, stochastic simulation, remote sensing, regression equations and mapped data; issues of accuracy, uncertainty and simply providing adequate data for the complex models; how such a modelling system fits into an organisational framework.

Wealth effects of mergers and acquisitions for US firms: using alternative pricing models

This empirical study employs a different methodology to examine the change in wealth associated with mergers and acquisitions (M&As) for US firms. Specifically, we employ the standard CAPM, the Fama-French three-factor model and the Carhart four-factor models within the OLS and GJR-GARCH estimation methods to test the behaviour of the cumulative abnormal returns (CARs). Whilst the standard CAPM captures the variability of stock returns with the overall market, the Fama-French factors capture the risk factors that are important to investors. Additionally, augmenting the Fama-French three-factor model with the Carhart momentum factor to generate the four-factor captures additional pricing elements that may affect stock returns. Traditionally, estimates of abnormal returns (ARs) in M&As situations rely on the standard OLS estimation method. However, the standard OLS will provide inefficient estimates of the ARs if the data contain ARCH and asymmetric effects. To minimise this problem of estimation efficiency we re-estimated the ARs using GJR-GARCH estimation method. We find that there is variation in the results both as regards the choice models and estimation methods. Besides these variations in the estimated models and the choice of estimation methods, we also tested whether the ARs are affected by the degree of liquidity of the stocks and the size of the firm. We document significant positive post-announcement cumulative ARs (CARs) for target firm shareholders under both the OLS and GJR-GARCH methods across all three methodologies. However, post-event CARs for acquiring firm shareholders were insignificant for both sets of estimation methods under the three methodologies. The GJR-GARCH method seems to generate larger CARs than those of the OLS method. Using both market capitalization and trading volume as a measure of liquidity and the size of the firm, we observed strong return continuations in the medium firms relative to small and large firms for target shareholders. We consistently observed market efficiency in small and large firm. This implies that target firms for small and large firms overreact to new information resulting in a more efficient market. For acquirer firms, our measure of liquidity captures strong return continuations for small firms under the OLS estimates for both CAPM and Fama-French three-factor models, whilst under the GJR-GARCH estimates only for Carhart model. Post-announcement bootstrapping simulated CARs confirmed our earlier results.

A comparison of sales response predictions from demand models applied to store-level versus panel data

In order to generate sales promotion response predictions, marketing analysts estimate demand models using either disaggregated (consumer-level) or aggregated (store-level) scanner data. Comparison of predictions from these demand models is complicated by the fact that models may accommodate different forms of consumer heterogeneity depending on the level of data aggregation. This study shows via simulation that demand models with various heterogeneity specifications do not produce more accurate sales response predictions than a homogeneous demand model applied to store-level data, with one major exception: a random coefficients model designed to capture within-store heterogeneity using store-level data produced significantly more accurate sales response predictions (as well as better fit) compared to other model specifications. An empirical application to the paper towel product category adds additional insights. This article has supplementary material online.

The role of models@run.time in supporting on-the-fly interoperability

Models at runtime can be defined as abstract representations of a system, including its structure and behaviour, which exist in tandem with the given system during the actual execution time of that system. Furthermore, these models should be causally connected to the system being modelled, offering a reflective capability. Significant advances have been made in recent years in applying this concept, most notably in adaptive systems. In this paper we argue that a similar approach can also be used to support the dynamic generation of software artefacts at execution time. An important area where this is relevant is the generation of software mediators to tackle the crucial problem of interoperability in distributed systems. We refer to this approach as emergent middleware, representing a fundamentally new approach to resolving interoperability problems in the complex distributed systems of today. In this context, the runtime models are used to capture meta-information about the underlying networked systems that need to interoperate, including their interfaces and additional knowledge about their associated behaviour. This is supplemented by ontological information to enable semantic reasoning. This paper focuses on this novel use of models at runtime, examining in detail the nature of such runtime models coupled with consideration of the supportive algorithms and tools that extract this knowledge and use it to synthesise the appropriate emergent middleware.

Automatic labelling of topic models learned from Twitter by summarisation

Latent topics derived by topic models such as Latent Dirichlet Allocation (LDA) are the result of hidden thematic structures which provide further insights into the data. The automatic labelling of such topics derived from social media poses however new challenges since topics may characterise novel events happening in the real world. Existing automatic topic labelling approaches which depend on external knowledge sources become less applicable here since relevant articles/concepts of the extracted topics may not exist in external sources. In this paper we propose to address the problem of automatic labelling of latent topics learned from Twitter as a summarisation problem. We introduce a framework which apply summarisation algorithms to generate topic labels. These algorithms are independent of external sources and only rely on the identification of dominant terms in documents related to the latent topic. We compare the efficiency of existing state of the art summarisation algorithms. Our results suggest that summarisation algorithms generate better topic labels which capture event-related context compared to the top-n terms returned by LDA. © 2014 Association for Computational Linguistics.

Estimation of global beta and tests of capital asset pricing models

With the rapid globalization and integration of world capital markets, more and more stocks are listed in multiple markets. With multi-listed stocks, the traditional measurement of systematic risk, the domestic beta, is not appropriate since it only contain information from one market. ^ Prakash et al. (1993) developed a technique, the global beta, to capture information from multiple markets wherein the stocks are listed. In this study, the global betas are obtained as well as domestic betas for 704 multi-listed stocks from 59 world equity markets. Welch tests show that domestic betas are not equal across markets, therefore, global beta is more appropriate in a global investment setting. ^ The traditional Capital Asset Pricing Models (CAPM) is also tested with regards to both domestic beta and global beta. The results generally support the positive relationship between stocks returns and global beta while tend to reject this relationship between stocks returns and domestic beta. Further tests of International CAPM with domestic beta and global beta strengthen the conclusion.^

Data to Decision in a Dynamic Ocean: Robust Species Distribution Models and Spatial Decision Frameworks

Human use of the oceans is increasingly in conflict with conservation of endangered species. Methods for managing the spatial and temporal placement of industries such as military, fishing, transportation and offshore energy, have historically been post hoc; i.e. the time and place of human activity is often already determined before assessment of environmental impacts. In this dissertation, I build robust species distribution models in two case study areas, US Atlantic (Best et al. 2012) and British Columbia (Best et al. 2015), predicting presence and abundance respectively, from scientific surveys. These models are then applied to novel decision frameworks for preemptively suggesting optimal placement of human activities in space and time to minimize ecological impacts: siting for offshore wind energy development, and routing ships to minimize risk of striking whales. Both decision frameworks relate the tradeoff between conservation risk and industry profit with synchronized variable and map views as online spatial decision support systems.

For siting offshore wind energy development (OWED) in the U.S. Atlantic (chapter 4), bird density maps are combined across species with weights of OWED sensitivity to collision and displacement and 10 km2 sites are compared against OWED profitability based on average annual wind speed at 90m hub heights and distance to transmission grid. A spatial decision support system enables toggling between the map and tradeoff plot views by site. A selected site can be inspected for sensitivity to a cetaceans throughout the year, so as to capture months of the year which minimize episodic impacts of pre-operational activities such as seismic airgun surveying and pile driving.

Routing ships to avoid whale strikes (chapter 5) can be similarly viewed as a tradeoff, but is a different problem spatially. A cumulative cost surface is generated from density surface maps and conservation status of cetaceans, before applying as a resistance surface to calculate least-cost routes between start and end locations, i.e. ports and entrance locations to study areas. Varying a multiplier to the cost surface enables calculation of multiple routes with different costs to conservation of cetaceans versus cost to transportation industry, measured as distance. Similar to the siting chapter, a spatial decisions support system enables toggling between the map and tradeoff plot view of proposed routes. The user can also input arbitrary start and end locations to calculate the tradeoff on the fly.

Essential to the input of these decision frameworks are distributions of the species. The two preceding chapters comprise species distribution models from two case study areas, U.S. Atlantic (chapter 2) and British Columbia (chapter 3), predicting presence and density, respectively. Although density is preferred to estimate potential biological removal, per Marine Mammal Protection Act requirements in the U.S., all the necessary parameters, especially distance and angle of observation, are less readily available across publicly mined datasets.

In the case of predicting cetacean presence in the U.S. Atlantic (chapter 2), I extracted datasets from the online OBIS-SEAMAP geo-database, and integrated scientific surveys conducted by ship (n=36) and aircraft (n=16), weighting a Generalized Additive Model by minutes surveyed within space-time grid cells to harmonize effort between the two survey platforms. For each of 16 cetacean species guilds, I predicted the probability of occurrence from static environmental variables (water depth, distance to shore, distance to continental shelf break) and time-varying conditions (monthly sea-surface temperature). To generate maps of presence vs. absence, Receiver Operator Characteristic (ROC) curves were used to define the optimal threshold that minimizes false positive and false negative error rates. I integrated model outputs, including tables (species in guilds, input surveys) and plots (fit of environmental variables, ROC curve), into an online spatial decision support system, allowing for easy navigation of models by taxon, region, season, and data provider.

For predicting cetacean density within the inner waters of British Columbia (chapter 3), I calculated density from systematic, line-transect marine mammal surveys over multiple years and seasons (summer 2004, 2005, 2008, and spring/autumn 2007) conducted by Raincoast Conservation Foundation. Abundance estimates were calculated using two different methods: Conventional Distance Sampling (CDS) and Density Surface Modelling (DSM). CDS generates a single density estimate for each stratum, whereas DSM explicitly models spatial variation and offers potential for greater precision by incorporating environmental predictors. Although DSM yields a more relevant product for the purposes of marine spatial planning, CDS has proven to be useful in cases where there are fewer observations available for seasonal and inter-annual comparison, particularly for the scarcely observed elephant seal. Abundance estimates are provided on a stratum-specific basis. Steller sea lions and harbour seals are further differentiated by ‘hauled out’ and ‘in water’. This analysis updates previous estimates (Williams & Thomas 2007) by including additional years of effort, providing greater spatial precision with the DSM method over CDS, novel reporting for spring and autumn seasons (rather than summer alone), and providing new abundance estimates for Steller sea lion and northern elephant seal. In addition to providing a baseline of marine mammal abundance and distribution, against which future changes can be compared, this information offers the opportunity to assess the risks posed to marine mammals by existing and emerging threats, such as fisheries bycatch, ship strikes, and increased oil spill and ocean noise issues associated with increases of container ship and oil tanker traffic in British Columbia’s continental shelf waters.

Starting with marine animal observations at specific coordinates and times, I combine these data with environmental data, often satellite derived, to produce seascape predictions generalizable in space and time. These habitat-based models enable prediction of encounter rates and, in the case of density surface models, abundance that can then be applied to management scenarios. Specific human activities, OWED and shipping, are then compared within a tradeoff decision support framework, enabling interchangeable map and tradeoff plot views. These products make complex processes transparent for gaming conservation, industry and stakeholders towards optimal marine spatial management, fundamental to the tenets of marine spatial planning, ecosystem-based management and dynamic ocean management.

Improved Accuracy in Optical Motion Capture Measurements Leads to Increased Sensitivity to Kinematic Changes

Clinical optical motion capture allows us to obtain kinematic and kinetic outcome measures that aid clinicians in diagnosing and treating different pathologies affecting healthy gait. The long term aim for gait centres is for subject-specific analyses that can predict, prevent, or reverse the effects of pathologies through gait retraining. To track the body, anatomical segment coordinate systems are commonly created by applying markers to the surface of the skin over specific, bony anatomy that is manually palpated. The location and placement of these markers is subjective and precision errors of up to 25mm have been reported [1]. Additionally, the selection of which anatomical landmarks to use in segment models can result in large angular differences; for example angular differences in the trunk can range up to 53o for the same motion depending on marker placement [2]. These errors can result in erroneous kinematic outcomes that either diminish or increase the apparent effects of a treatment or pathology compared to healthy data. Our goal was to improve the accuracy and precision of optical motion capture outcome measures. This thesis describes two separate studies. In the first study we aimed to establish an approach that would allow us to independently quantify the error among trunk models. Using this approach we determined if there was a best model to accurately track trunk motion. In the second study we designed a device to improve precision for test, re-test protocols that would also reduce the set-up time for motion capture experiments. Our method to compare a kinematically derived centre of mass velocity to one that was derived kinetically was successful in quantifying error among trunk models. Our findings indicate that models that use lateral shoulder markers as well as limit the translational degrees of freedom of the trunk through shared pelvic markers result in the least amount of error for the tasks we studied. We also successfully reduced intra- and inter-operator anatomical marker placement errors using a marker alignment device. The improved accuracy and precision resulting from the methods established in this thesis may lead to increased sensitivity to changes in kinematics, and ultimately result in more consistent treatment outcomes.

Bayesian nonparametric latent variable models

L’un des problèmes importants en apprentissage automatique est de déterminer la complexité du modèle à apprendre. Une trop grande complexité mène au surapprentissage, ce qui correspond à trouver des structures qui n’existent pas réellement dans les données, tandis qu’une trop faible complexité mène au sous-apprentissage, c’est-à-dire que l’expressivité du modèle est insuffisante pour capturer l’ensemble des structures présentes dans les données. Pour certains modèles probabilistes, la complexité du modèle se traduit par l’introduction d’une ou plusieurs variables cachées dont le rôle est d’expliquer le processus génératif des données. Il existe diverses approches permettant d’identifier le nombre approprié de variables cachées d’un modèle. Cette thèse s’intéresse aux méthodes Bayésiennes nonparamétriques permettant de déterminer le nombre de variables cachées à utiliser ainsi que leur dimensionnalité. La popularisation des statistiques Bayésiennes nonparamétriques au sein de la communauté de l’apprentissage automatique est assez récente. Leur principal attrait vient du fait qu’elles offrent des modèles hautement flexibles et dont la complexité s’ajuste proportionnellement à la quantité de données disponibles. Au cours des dernières années, la recherche sur les méthodes d’apprentissage Bayésiennes nonparamétriques a porté sur trois aspects principaux : la construction de nouveaux modèles, le développement d’algorithmes d’inférence et les applications. Cette thèse présente nos contributions à ces trois sujets de recherches dans le contexte d’apprentissage de modèles à variables cachées. Dans un premier temps, nous introduisons le Pitman-Yor process mixture of Gaussians, un modèle permettant l’apprentissage de mélanges infinis de Gaussiennes. Nous présentons aussi un algorithme d’inférence permettant de découvrir les composantes cachées du modèle que nous évaluons sur deux applications concrètes de robotique. Nos résultats démontrent que l’approche proposée surpasse en performance et en flexibilité les approches classiques d’apprentissage. Dans un deuxième temps, nous proposons l’extended cascading Indian buffet process, un modèle servant de distribution de probabilité a priori sur l’espace des graphes dirigés acycliques. Dans le contexte de réseaux Bayésien, ce prior permet d’identifier à la fois la présence de variables cachées et la structure du réseau parmi celles-ci. Un algorithme d’inférence Monte Carlo par chaîne de Markov est utilisé pour l’évaluation sur des problèmes d’identification de structures et d’estimation de densités. Dans un dernier temps, nous proposons le Indian chefs process, un modèle plus général que l’extended cascading Indian buffet process servant à l’apprentissage de graphes et d’ordres. L’avantage du nouveau modèle est qu’il admet les connections entres les variables observables et qu’il prend en compte l’ordre des variables. Nous présentons un algorithme d’inférence Monte Carlo par chaîne de Markov avec saut réversible permettant l’apprentissage conjoint de graphes et d’ordres. L’évaluation est faite sur des problèmes d’estimations de densité et de test d’indépendance. Ce modèle est le premier modèle Bayésien nonparamétrique permettant d’apprendre des réseaux Bayésiens disposant d’une structure complètement arbitraire.

Understanding Co-translational Protein Targeting and Lithium Dendrite Formation through Free Energy Simulations and Coarse-Grained Models

We describe the application of alchemical free energy methods and coarse-grained models to study two key problems: (i) co-translational protein targeting and insertion to direct membrane proteins to the endoplasmic reticulum for proper localization and folding, (ii) lithium dendrite formation during recharging of lithium metal batteries. We show that conformational changes in the signal recognition particle, a central component of the protein targeting machinery, confer additional specificity during the the recognition of signal sequences. We then develop a three-dimensional coarse-grained model to study the long-timescale dynamics of membrane protein integration at the translocon and a framework for the calculation of binding free energies between the ribosome and translocon. Finally, we develop a coarse-grained model to capture the dynamics of lithium deposition and dissolution at the electrode interface with time-dependent voltages to show that pulse plating and reverse pulse plating methods can mitigate dendrite growth.

Investigating the mesh dependency and upscaling of 3D grain-based models for the simulation of brittle fracture processes in low-porosity crystalline rock

The application of 3D grain-based modelling techniques is investigated in both small and large scale 3DEC models, in order to simulate brittle fracture processes in low-porosity crystalline rock. Mesh dependency in 3D grain-based models (GBMs) is examined through a number of cases to compare Voronoi and tetrahedral grain assemblages. Various methods are used in the generation of tessellations, each with a number of issues and advantages. A number of comparative UCS test simulations capture the distinct failure mechanisms, strength profiles, and progressive damage development using various Voronoi and tetrahedral GBMs. Relative calibration requirements are outlined to generate similar macro-strength and damage profiles for all the models. The results confirmed a number of inherent model behaviors that arise due to mesh dependency. In Voronoi models, inherent tensile failure mechanisms are produced by internal wedging and rotation of Voronoi grains. This results in a combined dependence on frictional and cohesive strength. In tetrahedral models, increased kinematic freedom of grains and an abundance of straight, connected failure pathways causes a preference for shear failure. This results in an inability to develop significant normal stresses causing cohesional strength dependence. In general, Voronoi models require high relative contact tensile strength values, with lower contact stiffness and contact cohesional strength compared to tetrahedral tessellations. Upscaling of 3D GBMs is investigated for both Voronoi and tetrahedral tessellations using a case study from the AECL’s Mine-by-Experiment at the Underground Research Laboratory. An upscaled tetrahedral model was able to reasonably simulate damage development in the roof forming a notch geometry by adjusting the cohesive strength. An upscaled Voronoi model underestimated the damage development in the roof and floor, and overestimated the damage in the side-walls. This was attributed to the discretization resolution limitations.