Model-Based Contrastive Explanations for XAIP: Towards a General Model and Prototype

Planning is an important sub-field of artificial intelligence (AI) focusing on letting intelligent agents deliberate on the most adequate course of action to attain their goals. Thanks to the recent boost in the number of critical domains and systems which exploit planning for their internal procedures, there is an increasing need for planning systems to become more transparent and trustworthy. Along this line, planning systems are now required to produce not only plans but also explanations about those plans, or the way they were attained. To address this issue, a new research area is emerging in the AI panorama: eXplainable AI (XAI), within which explainable planning (XAIP) is a pivotal sub-field. As a recent domain, XAIP is far from mature. No consensus has been reached in the literature about what explanations are, how they should be computed, and what they should explain in the first place. Furthermore, existing contributions are mostly theoretical, and software implementations are rarely more than preliminary. To overcome such issues, in this thesis we design an explainable planning framework bridging the gap between theoretical contributions from literature and software implementations. More precisely, taking inspiration from the state of the art, we develop a formal model for XAIP, and the software tool enabling its practical exploitation. Accordingly, the contribution of this thesis is four-folded. First, we review the state of the art of XAIP, supplying an outline of its most significant contributions from the literature. We then generalise the aforementioned contributions into a unified model for XAIP, aimed at supporting model-based contrastive explanations. Next, we design and implement an algorithm-agnostic library for XAIP based on our model. Finally, we validate our library from a technological perspective, via an extensive testing suite. Furthermore, we assess its performance and usability through a set of benchmarks and end-to-end examples.

Desarrollo de un modelo integral de evaluación para instituciones de educación superior y análisis de su impacto en la planificación y gestión institucional

El propósito de esta tesis doctoral es el desarrollo de un modelo integral de evaluación de la gestión para instituciones de educación superior (IES), fundamentado en valorar la gestión de diferentes subsistemas que la integran, así como estudiar el impacto en la planificación y gestión institucional. Este Modelo de Evaluación Institucional fue denominado Modelo Integral de Evaluación de Gestión de las IES (MIEGIES), que incorpora la gestión de la complejidad, los aspectos gerenciales, el compromiso o responsabilidad social, los recursos, además de los procesos propios universitarios con una visión integral de la gestión. Las bases conceptuales se establecen por una revisión del contexto mundial de la educación superior, pasando por un análisis sobre evaluación y calidad en entornos universitarios. La siguiente reflexión conceptual versó sobre la gestión de la complejidad, de la gestión gerencial, de la gestión de responsabilidad social universitaria, de la gestión de los recursos y de la gestión de los procesos, seguida por un aporte sobre modelaje y modelos. Para finalizar, se presenta un resumen teórico sobre el alcance de la aplicación de ecuaciones estructurales para la validación de modelos. El desarrollo del modelo conceptual, dimensiones e indicadores, fue efectuado aplicando los principios de la metodología de sistemas suaves –SSM. Para ello, se identifica la definición raíz (DR), la razón sistémica de ser del modelo, para posteriormente desarrollar sus componentes y principios conceptuales. El modelo quedó integrado por cinco subsistemas, denominados: de la Complejidad, de la Responsabilidad Social Universitaria, Gerencial, de Procesos y de Recursos. Los subsistemas se consideran como dimensiones e indicadores para el análisis y son los agentes críticos para el funcionamiento de una IES. Los aspectos referidos a lo Epistemetodológico, comenzó por identificar el enfoque epistemológico que sustenta el abordaje metodológico escogido. A continuación se identifican los elementos clásicos que se siguieron para llevar a cabo la investigación: Alcance o profundidad, población y muestra, instrumentos de recolección de información y su validación, para finalizar con la explicación procedimental para validar el modelo MIEGIES. La población considerada para el estudio empírico de validación fueron 585 personas distribuidas entre alumnos, docentes, personal administrativo y directivos de una Universidad Pública Venezolana. La muestra calculada fue de 238 individuos, número considerado representativo de la población. La aplicación de los instrumentos diseñados y validados permitió la obtención de un conjunto de datos, a partir de los cuales se validó el modelo MIEGIES. La validación del Modelo MIGEIES parte de sugerencias conceptuales para el análisis de los datos. Para ello se identificaron las variables relevantes, que pueden ser constructos o conceptos, las variables latentes que no pueden ser medidas directamente, sino que requiere seleccionar los indicadores que mejor las representan. Se aplicó la estrategia de modelación confirmatoria de los Modelos de Ecuaciones Estructurales (SEM). Para ello se parte de un análisis descriptivo de los datos, estimando la fiabilidad. A continuación se aplica un análisis factorial exploratorio y un análisis factorial confirmatorio. Para el análisis de la significancia del modelo global y el impacto en la planificación y gestión, se consideran el análisis de coeficientes de regresión y la tabla de ANOVA asociada, la cual de manera global especifica que el modelo planteado permite explicar la relación entre las variables definidas para la evaluación de la gestión de las IES. Así mismo, se encontró que este resultado de manera global explica que en la evaluación institucional tiene mucha importancia la gestión de la calidad y las finanzas. Es de especial importancia destacar el papel que desarrolla la planificación estratégica como herramienta de gestión que permite apoyar la toma de decisiones de las organizaciones en torno al quehacer actual y al camino que deben recorrer en el futuro para adecuarse a los cambios y a las demandas que les impone el entorno. El contraste estadístico de los dos modelos ajustados, el teórico y el empírico, permitió a través de técnicas estadísticas multivariables, demostrar de manera satisfactoria, la validez y aplicación del modelo propuesto en las IES. Los resultados obtenidos permiten afirmar que se pueden estimar de manera significativa los constructos que definen la evaluación de las instituciones de educación superior mediante el modelo elaborado. En el capítulo correspondiente a Conclusiones, se presenta en una de las primeras instancias, la relación conceptual propuesta entre los procesos de evaluación de la gestión institucional y de los cinco subsistemas que la integran. Posteriormente se encuentra que los modelos de ecuaciones estructurales con base en la estrategia de modelación confirmatoria es una herramienta estadística adecuada en la validación del modelo teórico, que fue el procedimiento propuesto en el marco de la investigación. En cuanto al análisis del impacto del Modelo en la Planificación y la Gestión, se concluye que ésta es una herramienta útil para cerrar el círculo de evaluación institucional. La planificación y la evaluación institucional son procesos inherentes a la filosofía de gestión. Es por ello que se recomienda su práctica como de necesario cumplimiento en todas las instancias funcionales y operativas de las Instituciones de Educación Superior. ABSTRACT The purpose of this dissertation is the development of a comprehensive model of management evaluation for higher education institutions (HEIs), based on evaluating the management of different subsystems and study the impact on planning and institutional management. This model was named Institutional Assessment Comprehensive Evaluation Model for the Management of HEI (in Spanish, MIEGIES). The model incorporates the management of complexity, management issues, commitment and social responsibility and resources in addition to the university's own processes with a comprehensive view of management. The conceptual bases are established by a review of the global context of higher education, through analysis and quality assessment in university environments. The following conceptual discussions covered the management of complexity, management practice, management of university social responsibility, resources and processes, followed by a contribution of modeling and models. Finally, a theoretical overview of the scope of application of structural equation model (SEM) validation is presented. The development of the conceptual model, dimensions and indicators was carried out applying the principles of soft systems methodology (SSM). For this, the root definition (RD), the systemic rationale of the model, to further develop their components and conceptual principles are identified. The model was composed of five subsystems, called: Complexity, University Social Responsibility, Management, Process and Resources. The subsystems are considered as dimensions and measures for analysis and are critical agents for the functioning of HEIs. In matters relating to epistemology and methodology we began to identify the approach that underpins the research: Scope, population and sample and data collection instruments. The classic elements that were followed to conduct research are identified. It ends with the procedural explanation to validate the MIEGIES model. The population considered for the empirical validation study was composed of 585 people distributed among students, faculty, staff and authorities of a public Venezuelan university. The calculated sample was 238 individuals, number considered representative of the population. The application of designed and validated instruments allowed obtaining a data set, from which the MIEGIES model was validated. The MIGEIES Model validation is initiated by the theoretical analysis of concepts. For this purpose the relevant variables that can be concepts or constructs were identified. The latent variables cannot be measured directly, but require selecting indicators that best represent them. Confirmatory modeling strategy of Structural Equation Modeling (SEM) was applied. To do this, we start from a descriptive analysis of the data, estimating reliability. An exploratory factor analysis and a confirmatory factor analysis were applied. To analyze the significance of the overall models the analysis of regression coefficients and the associated ANOVA table are considered. This comprehensively specifies that the proposed model can explain the relationship between the variables defined for evaluating the management of HEIs. It was also found that this result comprehensively explains that for institutional evaluation quality management and finance are very important. It is especially relevant to emphasize the role developed by strategic planning as a management tool that supports the decision making of organizations around their usual activities and the way they should evolve in the future in order to adapt to changes and demands imposed by the environment. The statistical test of the two fitted models, the theoretical and the empirical, enabled through multivariate statistical techniques to demonstrate satisfactorily the validity and application of the proposed model for HEIs. The results confirm that the constructs that define the evaluation of HEIs in the developed model can be estimated. In the Conclusions section the conceptual relationship between the processes of management evaluation and the five subsystems that comprise it are shown. Subsequently, it is indicated that structural equation models based on confirmatory modeling strategy is a suitable statistical tool in validating the theoretical model, which was proposed in the framework of the research procedure. The impact of the model in Planning and Management indicates that this is a useful tool to complete the institutional assessment. Planning and institutional assessment processes are inherent in management philosophy. That is why its practice is recommended as necessary compliance in all functional and operational units of HEIs.

A Planning Model with Resources in e-Learning

This work proposes a model for planning of education based on resources and layers. Each learning material or concept is determined by certain characteristics: a layer and a list of resources and resource values. Models of studied subject domain, learner, information and verification unit, learning material, plan of education and education have been defined. The plan of education can be conventional, statical, author’s and dynamic. Algorithms for course generation, dynamic plan generation and carrying out education are presented. The proposed model for planning of education based on resources and layers has been included in the system PeU.

Tales of the City. Storytelling as a contemporary tool of urban planning and design

Almost thirty years ago, as the social sciences underwent their 'discursive turn', Bernardo Secchi (1984) drew, in what he called the 'urban planning narrative', the attention of planners to the production of myths, turning an activity often seen as primarily technical into one centred around the production of images and ideas. This conception of planning practice gave rise to a powerful current of research in English-speaking countries. Efforts were made to both combine the urban planning narrative with storytelling and to establish storytelling as a prescriptive or descriptive model for planning practice. Thus, just as storytelling is supposed to have led democratic communication off track through a pronounced concern for a good story, storytelling applied to the field of urban production may have led to an increasing preoccupation with staging and showmanship for projects to the detriment of their real inclusion in political debate. It is this possible transformation of the territorial action that will be the focus of the articles collected in this special issue of Articulo - Journal of urban research.

Transmission network expansion planning considering uncertainty in generation and demand

This paper presents a mathematical model and a methodology to solve a transmission network expansion planning problem considering uncertainty in demand and generation. The methodology used to solve the problem, finds the optimal transmission network expansion plan that allows the power system to operate adequately in an environment with uncertainty. The model presented results in an optimization problem that is solved using a specialized genetic algorithm. The results obtained for known systems from the literature show that cheaper plans can be found satisfying the uncertainty in demand and generation. ©2008 IEEE.

Integrated planning of electric power distribution networks

In this work, a heuristic model for integrated planning of primary distribution network and secondary distribution circuits is proposed. A Tabu Search (TS) algorithm is employed to solve the planning of primary distribution networks. Evolutionary Algorithms (EA) are used to solve the planning model of secondary networks. The planning integration of both networks is carried out by means a constructive heuristic taking into account a set of integration alternatives between these networks. These integration alternatives are treated in a hierarchical way. The planning of primary networks and secondary distribution circuits is carried out based on assessment of the effects of the alternative solutions in the expansion costs of both networks simultaneously. In order to evaluate this methodology, tests were performed for a real-life distribution system taking into account the primary and secondary networks.

2D-3D Reconstruction-Based Planning of Total Hip Arthroplasty

This chapter proposed a personalized X-ray reconstruction-based planning and post-operative treatment evaluation framework called iJoint for advancing modern Total Hip Arthroplasty (THA). Based on a mobile X-ray image calibration phantom and a unique 2D-3D reconstruction technique, iJoint can generate patient-specific models of hip joint by non-rigidly matching statistical shape models to the X-ray radiographs. Such a reconstruction enables a true 3D planning and treatment evaluation of hip arthroplasty from just 2D X-ray radiographs whose acquisition is part of the standard diagnostic and treatment loop. As part of the system, a 3D model-based planning environment provides surgeons with hip arthroplasty related parameters such as implant type, size, position, offset and leg length equalization. With this newly developed system, we are able to provide true 3D solutions for computer assisted planning of THA using only 2D X-ray radiographs, which is not only innovative but also cost-effective.

Homeless access to dental health care: Creation of a dental service model

This project involved developing a model for planning a dental emergency treatment center that could function as an embedded component of a shelter for the homeless population. The dental services provided by such a clinic should include treatment for tooth pain, dental caries or cavities, chipped or broken teeth, broken partials, abscessed teeth, emergency cleanings, periodontal disease or gum disease and fillings. These are the dental services that are most often sought by homeless people in hospital emergency rooms.^ The underlying assumption for this project was that the oral health needs of the homeless community can most effectively be addressed by implementing small dental clinics in existing facilities that provide shelter and other services for this population. The model described in this project identifies oral health care services that would be provided by the clinic, facility (physical plant) requirements and associated infrastructure to operate an embedded dental clinic, methods for obtaining funding, strategies of recruiting dental professionals to staff the facility, and methods to assess the outcomes of the embedded clinic strategy. As an example, this project describes a strategy for developing such an embedded clinic at San Antonio Metropolitan Ministries SAMM shelter based on recommendations from community health care leaders, managers of homeless shelters, members of the homeless community and dental professionals^

Intelligent execution monitoring and error analysis in planning involving processes

An intelligent agent, operating in an external world which cannot be fully described in its internal world model, must be able to monitor the success of a previously generated plan and to respond to any errors which may have occurred. The process of error analysis requires the ability to reason in an expert fashion about time and about processes occurring in the world. Reasoning about time is needed to deal with causality. Reasoning about processes is needed since the direct effects of a plan action can be completely specified when the plan is generated, but the indirect effects cannot. For example, the action `open tap' leads with certainty to `tap open', whereas whether there will be a fluid flow and how long it might last is more difficult to predict. The majority of existing planning systems cannot handle these kinds of reasoning, thus limiting their usefulness. This thesis argues that both kinds of reasoning require a complex internal representation of the world. The use of Qualitative Process Theory and an interval-based representation of time are proposed as a representation scheme for such a world model. The planning system which was constructed has been tested on a set of realistic planning scenarios. It is shown that even simple planning problems, such as making a cup of coffee, require extensive reasoning if they are to be carried out successfully. The final Chapter concludes that the planning system described does allow the correct solution of planning problems involving complex side effects, which planners up to now have been unable to solve.

Managing Interval Resources in Automated Planning

In this paper RDPPLan, a model for planning with quantitative resources specified as numerical intervals, is presented. Nearly all existing models of planning with resources require to specify exact values for updating resources modified by actions execution. In other words these models cannot deal with more realistic situations in which the resources quantities are not completely known but are bounded by intervals. The RDPPlan model allow to manage domains more tailored to real world, where preconditions and effects over quantitative resources can be specified by intervals of values, in addition mixed logical/quantitative and pure numerical goals can be posed. RDPPlan is based on non directional search over a planning graph, like DPPlan, from which it derives, it uses propagation rules which have been appropriately extended to the management of resource intervals. The propagation rules extended with resources must verify invariant properties over the planning graph which have been proven by the authors and guarantee the correctness of the approach. An implementation of the RDPPlan model is described with search strategies specifically developed for interval resources.

An estimate of biometric parameters in eucalyptus clone plantations in Southern Bahia: an application of the airborne laser scanning (ALS) technology

The application of airborne laser scanning (ALS) technologies in forest inventories has shown great potential to improve the efficiency of forest planning activities. Precise estimates, fast assessment and relatively low complexity can explain the good results in terms of efficiency. The evolution of GPS and inertial measurement technologies, as well as the observed lower assessment costs when these technologies are applied to large scale studies, can explain the increasing dissemination of ALS technologies. The observed good quality of results can be expressed by estimates of volumes and basal area with estimated error below the level of 8.4%, depending on the size of sampled area, the quantity of laser pulses per square meter and the number of control plots. This paper analyzes the potential of an ALS assessment to produce certain forest inventory statistics in plantations of cloned Eucalyptus spp with precision equal of superior to conventional methods. The statistics of interest in this case were: volume, basal area, mean height and dominant trees mean height. The ALS flight for data assessment covered two strips of approximately 2 by 20 Km, in which clouds of points were sampled in circular plots with a radius of 13 m. Plots were sampled in different parts of the strips to cover different stand ages. The clouds of points generated by the ALS assessment: overall height mean, standard error, five percentiles (height under which we can find 10%, 30%, 50%,70% and 90% of the ALS points above ground level in the cloud), and density of points above ground level in each percentile were calculated. The ALS statistics were used in regression models to estimate mean diameter, mean height, mean height of dominant trees, basal area and volume. Conventional forest inventory sample plots provided real data. For volume, an exploratory assessment involving different combinations of ALS statistics allowed for the definition of the most promising relationships and fitting tests based on well known forest biometric models. The models based on ALS statistics that produced the best results involved: the 30% percentile to estimate mean diameter (R(2)=0,88 and MQE%=0,0004); the 10% and 90% percentiles to estimate mean height (R(2)=0,94 and MQE%=0,0003); the 90% percentile to estimate dominant height (R(2)=0,96 and MQE%=0,0003); the 10% percentile and mean height of ALS points to estimate basal area (R(2)=0,92 and MQE%=0,0016); and, to estimate volume, age and the 30% and 90% percentiles (R(2)=0,95 MQE%=0,002). Among the tested forest biometric models, the best fits were provided by the modified Schumacher using age and the 90% percentile, modified Clutter using age, mean height of ALS points and the 70% percentile, and modified Buckman using age, mean height of ALS points and the 10% percentile.

Planeamento e controlo da instalação e manutenção de sistemas AVAC

Atualmente a elaboração de plano de trabalhos (utilizando a ferramenta Microsoft Project) é prática corrente em muitas empresas, como são o caso das empresas de construção civil, como meio de auxílio para o estabelecimento de prazos e articulação com as diversas subempreitadas que uma qualquer obra tem ao longo da sua execução. O presente trabalho visa demonstrar a importância da elaboração de um plano de trabalhos para o planeamento e controlo de uma instalação de Aquecimento, Ventilação e Ar Condicionado (AVAC). Recorrendo ao uso do software Microsoft Project foi elaborado um modelo de planeamento e controlo, utilizando dados reais, para o planeamento e controlo de uma instalação de AVAC. Foi ainda elaborado, através do Microsoft Excel, um plano anual de manutenções preventivas para os contratos de manutenção em vigor. No final, são indicadas as principais conclusões e as vantagens da utilização deste modelo na obtenção de melhores resultados no cumprimento de prazos e redução de custos com o pessoal afeto à empresa. São ainda perspetivados futuros desenvolvimentos utilizando essa ferramenta. Com o modelo criado, a empresa tem meios para planear e controlar uma série de parâmetros, como os prazos de entrega dos equipamentos a instalar, o número de trabalhadores necessários para desempenhar uma determinada função, os custos associados à mão-de-obra e/ou equipamento, o cumprimento de prazos estabelecidos pela empresa de construção civil e outros dados que possam vir a ser relevantes para a melhoria da rentabilidade dos projetos.

Toiminnanohjauksen kehittäminen toteutusaikataulun avulla

Tässä työssä tarkastellaan projektiin liittyvää toiminnanohjausta. Tutkimusongelmana on, kuinka toiminnanohjausta voitaisiin kehittää aikataulutuksen avulla. Työn yhtenä tavoitteena on luoda kokonaiskuva aikataulutuksen nykytilanteesta. Toisena tavoitteena on kartoittaa ajankohtaiset ongelmat. Kolmantena tavoitteena on luoda malli aikataulun laatimiselle kalanterin valmistuksessa.Kirjallisuustutkimuksen perusteella luotiin kuva projektiliiketoiminnasta ja projektinhallintaan sisältyvästä aikataulunhallinnasta. Haastattelututkimuksen perusteella luotiin kokonaiskuva projektin aikataulutuksesta ja siihen liittyvistä ongelmista. Tutkimusta täydennettiin analysoimalla erään projektin suunniteltua aikataulua ja toteutunutta aikataulua. Toiminnanohjauksen aikataulutustehtävä on muuttumassa yhä tarkeämmäksi toimittaessa resurssien äärirajoilla. Ajankohtaisia kysymyksiä ovat: toteutusaikataulun laatimistekniikka, projektin osittaminen, optimi suunnittelujärjestys ja kriittinen polku.Tutkimuksen perusteella kehitettiin toteutusaikataulun suunnittelun malli ja malliaikataulu. Mallin keskeiset osat ovat: projektin osittaminen tehtäviin, tärkeimpien tehtävien määrittäminen tarkempaa ohjausta varten, tehtävien välisten riippuvuuksien määrittäminen, tehtävien limitysmahdollisuuksien ja ajallisen keston arviointi, aikataulun esittäminen janakaaviona ja ulkoisten tekijöiden, kuten hankinnan asettamien vaatimusten huomioiminen malliaikataulussa. Malliaikataulua tullaan soveltamaan seuraavaan sopivaan toimitusprojektiin.

Rakennuksen tietomallien hyödyntämisen edellytykset rakentamisen valmistelu- ja rakennusvaiheessa

Työn tavoitteena oli tutkia, kuinka rakennuksen tietomallia voidaan hyödyntää rakennusliikkeen tuotannonsuunnittelu- ja rakennusvaiheessa sekä mitä sen hyödyntäminen edellyttää rakennuksen tietomallin informaatiosisällöltä. Tavoitteena oli myös tunnistaa tuotannonohjauksen ”pullonkauloja”, joissa kohdin toimintaa voitaisiin tietomallien avulla tehostaa. Työn teoreettisena taustana on aineettoman pääoman merkitys yrityksen kilpailuedun luojana, tietämyksen hallinta ja teknologian hyödyntäminen tietämyksen hallinnassa. Työssä tutkittiin rakennustuotannon johtamista ja tietomallintamisen hyödyntämistä rakentamisessa sekä tietomallien hyödynnettävyyden varmistamista yleisellä tasolla. Työssä tutustuttiin kohdeyritykseen tuotannonohjaukseen ja rakennusvaiheen tietomallien hyödyntämisen nykytilaan. Tuloksina voidaan todeta, että tietomallien tuotannonsuunnittelu- ja rakennusvaiheen tietomallien hyödyntämisen perusedellytys on tietomallien oikeellisuus sekä tietomallien ja perinteisten suunnitteludokumenttien yhdenmukainen tietosisältö. Tämän lisäksi tarvitaan suunnitellut toimintatavat ja toimivat tiedonjakelukanavat sekä kyky hyödyntää tieto- ja viestintäteknologiaa. Tietomallit eivät tämän tutkimuksen perusteella näytä luoneen tarvetta uudenlaisille tuotanto-organisaation roolituksille. Tietomalleilla uskotaan olevan positiivisia vaikutuksia rakennusvaiheen muutostenhallinnassa. Tuotantoorganisaation henkilöillä oli positiivisia odotuksia tietomallien hyödyntämisestä tuotannonohjauksessa. Tietomallien odotetaan tukevan erillisten suunnitelmien muodostamien kokonaisuuksien hahmottamista, rakennusvaiheen osapuolten yhteistyötä ja töiden yhteensovitusta sekä logistiikan suunnittelua ja vaikutusten havainnointia.

Approche sociolinguistique de la production terminologique au Québec : application du modèle glottopolitique

Thèse numérisée par la Division de la gestion de documents et des archives de l'Université de Montréal.