Gérer les installations septiques individuelles québécoises : lignes directrices pour un programme de gestion municipal

Au Québec, environ un million de résidences sont isolées des réseaux d’égouts municipaux et doivent traiter leurs eaux usées à l’aide de systèmes d’assainissement autonomes. Également nommés installations septiques, ces systèmes sont susceptibles de contaminer l’environnement lorsqu’ils sont défaillants, désuets ou non conformes. Les épisodes de cyanobactéries survenus de 2006 à 2012 dans les plans d’eau québécois ont été attribués à d’importants apports de phosphore, que libèrent notamment les installations septiques polluantes. Les municipalités, les municipalités régionales de comté et les régies intermunicipales ont des compétences et des obligations en vertu du Règlement sur l’évacuation et le traitement des eaux usées des résidences isolées et de l’article 25.1 de la Loi sur les compétences municipales portant sur une gestion adéquate des installations septiques. Une gestion optimale de celles-ci permettant de protéger l’environnement et de prévenir la contamination est possible par la mise en place d’un programme de gestion qui concerne plus spécifiquement la vidange des fosses septiques et l’inspection des systèmes. L’objectif de cet essai est de faire une analyse des éléments constituants d’un programme de gestion des installations septiques et de discuter de leur mise en oeuvre pour assurer leur fonctionnement optimal. L’essai a été rédigé de manière à présenter l’information sous forme de lignes directrices pour guider les gestionnaires de programme ainsi que les preneurs de décisions. Un programme de gestion optimal se traduit par la prise en charge de la vidange des fosses septiques d’un territoire par une des trois entités municipales qui peuvent exercer un meilleur contrôle des systèmes d’épuration autonomes par la vidange des fosses septiques et de leur inspection régulière. Les inspections prennent la forme de relevés sanitaires qui permettent de classifier les installations en fonction de leur performance et d’inspections sommaires qui visent à faire un diagnostic simple et rapide d’une installation. Plusieurs autres éléments doivent être pris en compte dans le cadre d’un programme de gestion. Soit l’application d’un règlement municipal, la fréquence et le type de vidange, le recours à des experts, la sensibilisation des propriétaires des systèmes d’épuration autonomes, une démarche d’acceptabilité sociale et l’emploi de logiciels de suivi. L’efficacité d’un programme de gestion dépendra de la rigueur avec laquelle les instances municipales appliqueront les éléments de gestion. Il est recommandé au ministère du Développement durable, de l’Environnement et de la Lutte contre les changements climatiques d’apporter des modifications au règlement provincial dans le but de le rendre plus contraignant pour les systèmes vecteurs de contamination indirecte et ceux antérieurs à 1981. Le ministère des Affaires municipales et de l’Occupation du territoire devrait offrir son soutien aux municipalités dans le cadre de leur programme de gestion en offrant de l’aide financière, de la documentation et des formations.

Classification multi-échelle d'images à très haute résolution spatiale basée sur une nouvelle approche texturale

Résumé : Face à l’accroissement de la résolution spatiale des capteurs optiques satellitaires, de nouvelles stratégies doivent être développées pour classifier les images de télédétection. En effet, l’abondance de détails dans ces images diminue fortement l’efficacité des classifications spectrales; de nombreuses méthodes de classification texturale, notamment les approches statistiques, ne sont plus adaptées. À l’inverse, les approches structurelles offrent une ouverture intéressante : ces approches orientées objet consistent à étudier la structure de l’image pour en interpréter le sens. Un algorithme de ce type est proposé dans la première partie de cette thèse. Reposant sur la détection et l’analyse de points-clés (KPC : KeyPoint-based Classification), il offre une solution efficace au problème de la classification d’images à très haute résolution spatiale. Les classifications effectuées sur les données montrent en particulier sa capacité à différencier des textures visuellement similaires. Par ailleurs, il a été montré dans la littérature que la fusion évidentielle, reposant sur la théorie de Dempster-Shafer, est tout à fait adaptée aux images de télédétection en raison de son aptitude à intégrer des concepts tels que l’ambiguïté et l’incertitude. Peu d’études ont en revanche été menées sur l’application de cette théorie à des données texturales complexes telles que celles issues de classifications structurelles. La seconde partie de cette thèse vise à combler ce manque, en s’intéressant à la fusion de classifications KPC multi-échelle par la théorie de Dempster-Shafer. Les tests menés montrent que cette approche multi-échelle permet d’améliorer la classification finale dans le cas où l’image initiale est de faible qualité. De plus, l’étude effectuée met en évidence le potentiel d’amélioration apporté par l’estimation de la fiabilité des classifications intermédiaires, et fournit des pistes pour mener ces estimations.

Spam campaign detection, analysis, and formalization

Les courriels Spams (courriels indésirables ou pourriels) imposent des coûts annuels extrêmement lourds en termes de temps, d’espace de stockage et d’argent aux utilisateurs privés et aux entreprises. Afin de lutter efficacement contre le problème des spams, il ne suffit pas d’arrêter les messages de spam qui sont livrés à la boîte de réception de l’utilisateur. Il est obligatoire, soit d’essayer de trouver et de persécuter les spammeurs qui, généralement, se cachent derrière des réseaux complexes de dispositifs infectés, ou d’analyser le comportement des spammeurs afin de trouver des stratégies de défense appropriées. Cependant, une telle tâche est difficile en raison des techniques de camouflage, ce qui nécessite une analyse manuelle des spams corrélés pour trouver les spammeurs. Pour faciliter une telle analyse, qui doit être effectuée sur de grandes quantités des courriels non classés, nous proposons une méthodologie de regroupement catégorique, nommé CCTree, permettant de diviser un grand volume de spams en des campagnes, et ce, en se basant sur leur similarité structurale. Nous montrons l’efficacité et l’efficience de notre algorithme de clustering proposé par plusieurs expériences. Ensuite, une approche d’auto-apprentissage est proposée pour étiqueter les campagnes de spam en se basant sur le but des spammeur, par exemple, phishing. Les campagnes de spam marquées sont utilisées afin de former un classificateur, qui peut être appliqué dans la classification des nouveaux courriels de spam. En outre, les campagnes marquées, avec un ensemble de quatre autres critères de classement, sont ordonnées selon les priorités des enquêteurs. Finalement, une structure basée sur le semiring est proposée pour la représentation abstraite de CCTree. Le schéma abstrait de CCTree, nommé CCTree terme, est appliqué pour formaliser la parallélisation du CCTree. Grâce à un certain nombre d’analyses mathématiques et de résultats expérimentaux, nous montrons l’efficience et l’efficacité du cadre proposé.

IntelliChair: a non-intrusive sitting posture and sitting activity recognition system

Current Ambient Intelligence and Intelligent Environment research focuses on the interpretation of a subject’s behaviour at the activity level by logging the Activity of Daily Living (ADL) such as eating, cooking, etc. In general, the sensors employed (e.g. PIR sensors, contact sensors) provide low resolution information. Meanwhile, the expansion of ubiquitous computing allows researchers to gather additional information from different types of sensor which is possible to improve activity analysis. Based on the previous research about sitting posture detection, this research attempts to further analyses human sitting activity. The aim of this research is to use non-intrusive low cost pressure sensor embedded chair system to recognize a subject’s activity by using their detected postures. There are three steps for this research, the first step is to find a hardware solution for low cost sitting posture detection, second step is to find a suitable strategy of sitting posture detection and the last step is to correlate the time-ordered sitting posture sequences with sitting activity. The author initiated a prototype type of sensing system called IntelliChair for sitting posture detection. Two experiments are proceeded in order to determine the hardware architecture of IntelliChair system. The prototype looks at the sensor selection and integration of various sensor and indicates the best for a low cost, non-intrusive system. Subsequently, this research implements signal process theory to explore the frequency feature of sitting posture, for the purpose of determining a suitable sampling rate for IntelliChair system. For second and third step, ten subjects are recruited for the sitting posture data and sitting activity data collection. The former dataset is collected byasking subjects to perform certain pre-defined sitting postures on IntelliChair and it is used for posture recognition experiment. The latter dataset is collected by asking the subjects to perform their normal sitting activity routine on IntelliChair for four hours, and the dataset is used for activity modelling and recognition experiment. For the posture recognition experiment, two Support Vector Machine (SVM) based classifiers are trained (one for spine postures and the other one for leg postures), and their performance evaluated. Hidden Markov Model is utilized for sitting activity modelling and recognition in order to establish the selected sitting activities from sitting posture sequences.2. After experimenting with possible sensors, Force Sensing Resistor (FSR) is selected as the pressure sensing unit for IntelliChair. Eight FSRs are mounted on the seat and back of a chair to gather haptic (i.e., touch-based) posture information. Furthermore, the research explores the possibility of using alternative non-intrusive sensing technology (i.e. vision based Kinect Sensor from Microsoft) and find out the Kinect sensor is not reliable for sitting posture detection due to the joint drifting problem. A suitable sampling rate for IntelliChair is determined according to the experiment result which is 6 Hz. The posture classification performance shows that the SVM based classifier is robust to “familiar” subject data (accuracy is 99.8% with spine postures and 99.9% with leg postures). When dealing with “unfamiliar” subject data, the accuracy is 80.7% for spine posture classification and 42.3% for leg posture classification. The result of activity recognition achieves 41.27% accuracy among four selected activities (i.e. relax, play game, working with PC and watching video). The result of this thesis shows that different individual body characteristics and sitting habits influence both sitting posture and sitting activity recognition. In this case, it suggests that IntelliChair is suitable for individual usage but a training stage is required.

Implementação de classificador de tarefas mentais baseado em EEG

BARBOSA, André F. ; SOUZA, Bryan C. ; PEREIRA JUNIOR, Antônio ; MEDEIROS, Adelardo A. D.de, . Implementação de Classificador de Tarefas Mentais Baseado em EEG. In: CONGRESSO BRASILEIRO DE REDES NEURAIS, 9., 2009, Ouro Preto, MG. Anais... Ouro Preto, MG, 2009

Burned Area Mapping in the North American Boreal Forest Using Terra-MODIS LTDR (2001–2011): A Comparison with the MCD45A1, MCD64A1 and BA GEOLAND-2 Products

An algorithm based on a Bayesian network classifier was adapted to produce 10-day burned area (BA) maps from the Long Term Data Record Version 3 (LTDR) at a spatial resolution of 0.05° (~5 km) for the North American boreal region from 2001 to 2011. The modified algorithm used the Brightness Temperature channel from the Moderate Resolution Imaging Spectroradiometer (MODIS) band 31 T31 (11.03 μm) instead of the Advanced Very High Resolution Radiometer (AVHRR) band T3 (3.75 μm). The accuracy of the BA-LTDR, the Collection 5.1 MODIS Burned Area (MCD45A1), the MODIS Collection 5.1 Direct Broadcast Monthly Burned Area (MCD64A1) and the Burned Area GEOLAND-2 (BA GEOLAND-2) products was assessed using reference data from the Alaska Fire Service (AFS) and the Canadian Forest Service National Fire Database (CFSNFD). The linear regression analysis of the burned area percentages of the MCD64A1 product using 40 km × 40 km grids versus the reference data for the years from 2001 to 2011 showed an agreement of R2 = 0.84 and a slope = 0.76, while the BA-LTDR showed an agreement of R2 = 0.75 and a slope = 0.69. These results represent an improvement over the MCD45A1 product, which showed an agreement of R2 = 0.67 and a slope = 0.42. The MCD64A1, BA-LTDR and MCD45A1 products underestimated the total burned area in the study region, whereas the BA GEOLAND-2 product overestimated it by approximately five-fold, with an agreement of R2 = 0.05. Despite MCD64A1 showing the best overall results, the BA-LTDR product proved to be an alternative for mapping burned areas in the North American boreal forest region compared with the other global BA products, even those with higher spatial/spectral resolution

Fiber Bragg Gratings Strain Sensors for Railway Applications

Fiber optical sensors have played an important role in applications for monitoring the health of civil infrastructures, such as bridges, oil rigs, and railroads. Due to the reduction in cost of fiber-optic components and systems, fiber optical sensors have been studied extensively for their higher sensitivity, precision and immunity to electrical interference compared to their electrical counterparts. A fiber Bragg grating (FBG) strain sensor has been employed for this study to detect and distinguish normal and lateral loads on rail tracks. A theoretical analysis of the relationship between strain and displacement under vertical and horizontal strains on an aluminum beam has been performed, and the results are in excellent agreement with the measured strain data. Then a single FBG sensor system with erbium-doped fiber amplifier broadband source has been carried out. Force and temperature applied on the system have resulted in changes of 0.05 nm per 50 με and 0.094 nm per 10 oC at the center wavelength of the FBG. Furthermore, a low cost fiber-optic sensor system with a distributed feedback (DFB) laser as the light source has been implemented. We show that it has superior noise and sensitivity performances compared to strain gauge sensors. The design has been extended to accommodate multiple sensors with negligible cross talk. When two cascaded sensors on a rail track section are tested, strain readings of the sensor 20 inches away from the position of applied force decay to one seventh of the data of the sensor at the applied force location. The two FBG sensor systems can detect 1 ton of vertical load with a square wave pattern and 0.1 ton of lateral loads (3 tons and 0.5 ton, respectively, for strain gauges). Moreover, a single FBG sensor has been found capable of detecting and distinguishing lateral and normal strains applied at different frequencies. FBG sensors are promising alternatives to electrical sensors for their high sensitivity,ease of installation, and immunity to electromagnetic interferences.

'Explicit Learning: an Effort towards Human Scheduling Algorithms'

Abstract Scheduling problems are generally NP-hard combinatorial problems, and a lot of research has been done to solve these problems heuristically. However, most of the previous approaches are problem-specific and research into the development of a general scheduling algorithm is still in its infancy. Mimicking the natural evolutionary process of the survival of the fittest, Genetic Algorithms (GAs) have attracted much attention in solving difficult scheduling problems in recent years. Some obstacles exist when using GAs: there is no canonical mechanism to deal with constraints, which are commonly met in most real-world scheduling problems, and small changes to a solution are difficult. To overcome both difficulties, indirect approaches have been presented (in [1] and [2]) for nurse scheduling and driver scheduling, where GAs are used by mapping the solution space, and separate decoding routines then build solutions to the original problem. In our previous indirect GAs, learning is implicit and is restricted to the efficient adjustment of weights for a set of rules that are used to construct schedules. The major limitation of those approaches is that they learn in a non-human way: like most existing construction algorithms, once the best weight combination is found, the rules used in the construction process are fixed at each iteration. However, normally a long sequence of moves is needed to construct a schedule and using fixed rules at each move is thus unreasonable and not coherent with human learning processes. When a human scheduler is working, he normally builds a schedule step by step following a set of rules. After much practice, the scheduler gradually masters the knowledge of which solution parts go well with others. He can identify good parts and is aware of the solution quality even if the scheduling process is not completed yet, thus having the ability to finish a schedule by using flexible, rather than fixed, rules. In this research we intend to design more human-like scheduling algorithms, by using ideas derived from Bayesian Optimization Algorithms (BOA) and Learning Classifier Systems (LCS) to implement explicit learning from past solutions. BOA can be applied to learn to identify good partial solutions and to complete them by building a Bayesian network of the joint distribution of solutions [3]. A Bayesian network is a directed acyclic graph with each node corresponding to one variable, and each variable corresponding to individual rule by which a schedule will be constructed step by step. The conditional probabilities are computed according to an initial set of promising solutions. Subsequently, each new instance for each node is generated by using the corresponding conditional probabilities, until values for all nodes have been generated. Another set of rule strings will be generated in this way, some of which will replace previous strings based on fitness selection. If stopping conditions are not met, the Bayesian network is updated again using the current set of good rule strings. The algorithm thereby tries to explicitly identify and mix promising building blocks. It should be noted that for most scheduling problems the structure of the network model is known and all the variables are fully observed. In this case, the goal of learning is to find the rule values that maximize the likelihood of the training data. Thus learning can amount to 'counting' in the case of multinomial distributions. In the LCS approach, each rule has its strength showing its current usefulness in the system, and this strength is constantly assessed [4]. To implement sophisticated learning based on previous solutions, an improved LCS-based algorithm is designed, which consists of the following three steps. The initialization step is to assign each rule at each stage a constant initial strength. Then rules are selected by using the Roulette Wheel strategy. The next step is to reinforce the strengths of the rules used in the previous solution, keeping the strength of unused rules unchanged. The selection step is to select fitter rules for the next generation. It is envisaged that the LCS part of the algorithm will be used as a hill climber to the BOA algorithm. This is exciting and ambitious research, which might provide the stepping-stone for a new class of scheduling algorithms. Data sets from nurse scheduling and mall problems will be used as test-beds. It is envisaged that once the concept has been proven successful, it will be implemented into general scheduling algorithms. It is also hoped that this research will give some preliminary answers about how to include human-like learning into scheduling algorithms and may therefore be of interest to researchers and practitioners in areas of scheduling and evolutionary computation. References 1. Aickelin, U. and Dowsland, K. (2003) 'Indirect Genetic Algorithm for a Nurse Scheduling Problem', Computer & Operational Research (in print). 2. Li, J. and Kwan, R.S.K. (2003), 'Fuzzy Genetic Algorithm for Driver Scheduling', European Journal of Operational Research 147(2): 334-344. 3. Pelikan, M., Goldberg, D. and Cantu-Paz, E. (1999) 'BOA: The Bayesian Optimization Algorithm', IlliGAL Report No 99003, University of Illinois. 4. Wilson, S. (1994) 'ZCS: A Zeroth-level Classifier System', Evolutionary Computation 2(1), pp 1-18.

Leveraging Multiple Features for Image Retrieval and Matching

The goal of image retrieval and matching is to find and locate object instances in images from a large-scale image database. While visual features are abundant, how to combine them to improve performance by individual features remains a challenging task. In this work, we focus on leveraging multiple features for accurate and efficient image retrieval and matching. We first propose two graph-based approaches to rerank initially retrieved images for generic image retrieval. In the graph, vertices are images while edges are similarities between image pairs. Our first approach employs a mixture Markov model based on a random walk model on multiple graphs to fuse graphs. We introduce a probabilistic model to compute the importance of each feature for graph fusion under a naive Bayesian formulation, which requires statistics of similarities from a manually labeled dataset containing irrelevant images. To reduce human labeling, we further propose a fully unsupervised reranking algorithm based on a submodular objective function that can be efficiently optimized by greedy algorithm. By maximizing an information gain term over the graph, our submodular function favors a subset of database images that are similar to query images and resemble each other. The function also exploits the rank relationships of images from multiple ranked lists obtained by different features. We then study a more well-defined application, person re-identification, where the database contains labeled images of human bodies captured by multiple cameras. Re-identifications from multiple cameras are regarded as related tasks to exploit shared information. We apply a novel multi-task learning algorithm using both low level features and attributes. A low rank attribute embedding is joint learned within the multi-task learning formulation to embed original binary attributes to a continuous attribute space, where incorrect and incomplete attributes are rectified and recovered. To locate objects in images, we design an object detector based on object proposals and deep convolutional neural networks (CNN) in view of the emergence of deep networks. We improve a Fast RCNN framework and investigate two new strategies to detect objects accurately and efficiently: scale-dependent pooling (SDP) and cascaded rejection classifiers (CRC). The SDP improves detection accuracy by exploiting appropriate convolutional features depending on the scale of input object proposals. The CRC effectively utilizes convolutional features and greatly eliminates negative proposals in a cascaded manner, while maintaining a high recall for true objects. The two strategies together improve the detection accuracy and reduce the computational cost.

Applied web traffic analysis for numerical encoding of SQL Injection attack features.

SQL Injection Attack (SQLIA) remains a technique used by a computer network intruder to pilfer an organisation’s confidential data. This is done by an intruder re-crafting web form’s input and query strings used in web requests with malicious intent to compromise the security of an organisation’s confidential data stored at the back-end database. The database is the most valuable data source, and thus, intruders are unrelenting in constantly evolving new techniques to bypass the signature’s solutions currently provided in Web Application Firewalls (WAF) to mitigate SQLIA. There is therefore a need for an automated scalable methodology in the pre-processing of SQLIA features fit for a supervised learning model. However, obtaining a ready-made scalable dataset that is feature engineered with numerical attributes dataset items to train Artificial Neural Network (ANN) and Machine Leaning (ML) models is a known issue in applying artificial intelligence to effectively address ever evolving novel SQLIA signatures. This proposed approach applies numerical attributes encoding ontology to encode features (both legitimate web requests and SQLIA) to numerical data items as to extract scalable dataset for input to a supervised learning model in moving towards a ML SQLIA detection and prevention model. In numerical attributes encoding of features, the proposed model explores a hybrid of static and dynamic pattern matching by implementing a Non-Deterministic Finite Automaton (NFA). This combined with proxy and SQL parser Application Programming Interface (API) to intercept and parse web requests in transition to the back-end database. In developing a solution to address SQLIA, this model allows processed web requests at the proxy deemed to contain injected query string to be excluded from reaching the target back-end database. This paper is intended for evaluating the performance metrics of a dataset obtained by numerical encoding of features ontology in Microsoft Azure Machine Learning (MAML) studio using Two-Class Support Vector Machines (TCSVM) binary classifier. This methodology then forms the subject of the empirical evaluation.

Active Data Collection Techniques to Understand Online Scammers and Cybercriminals

Nigerian scam, also known as advance fee fraud or 419 scam, is a prevalent form of online fraudulent activity that causes financial loss to individuals and businesses. Nigerian scam has evolved from simple non-targeted email messages to more sophisticated scams targeted at users of classifieds, dating and other websites. Even though such scams are observed and reported by users frequently, the community’s understanding of Nigerian scams is limited since the scammers operate “underground”. To better understand the underground Nigerian scam ecosystem and seek effective methods to deter Nigerian scam and cybercrime in general, we conduct a series of active and passive measurement studies. Relying upon the analysis and insight gained from the measurement studies, we make four contributions: (1) we analyze the taxonomy of Nigerian scam and derive long-term trends in scams; (2) we provide an insight on Nigerian scam and cybercrime ecosystems and their underground operation; (3) we propose a payment intervention as a potential deterrent to cybercrime operation in general and evaluate its effectiveness; and (4) we offer active and passive measurement tools and techniques that enable in-depth analysis of cybercrime ecosystems and deterrence on them. We first created and analyze a repository of more than two hundred thousand user-reported scam emails, stretching from 2006 to 2014, from four major scam reporting websites. We select ten most commonly observed scam categories and tag 2,000 scam emails randomly selected from our repository. Based upon the manually tagged dataset, we train a machine learning classifier and cluster all scam emails in the repository. From the clustering result, we find a strong and sustained upward trend for targeted scams and downward trend for non-targeted scams. We then focus on two types of targeted scams: sales scams and rental scams targeted users on Craigslist. We built an automated scam data collection system and gathered large-scale sales scam emails. Using the system we posted honeypot ads on Craigslist and conversed automatically with the scammers. Through the email conversation, the system obtained additional confirmation of likely scam activities and collected additional information such as IP addresses and shipping addresses. Our analysis revealed that around 10 groups were responsible for nearly half of the over 13,000 total scam attempts we received. These groups used IP addresses and shipping addresses in both Nigeria and the U.S. We also crawled rental ads on Craigslist, identified rental scam ads amongst the large number of benign ads and conversed with the potential scammers. Through in-depth analysis of the rental scams, we found seven major scam campaigns employing various operations and monetization methods. We also found that unlike sales scammers, most rental scammers were in the U.S. The large-scale scam data and in-depth analysis provide useful insights on how to design effective deterrence techniques against cybercrime in general. We study underground DDoS-for-hire services, also known as booters, and measure the effectiveness of undermining a payment system of DDoS Services. Our analysis shows that the payment intervention can have the desired effect of limiting cybercriminals’ ability and increasing the risk of accepting payments.

An investigation of Hebbian phase sequences as assembly graphs

Hebb proposed that synapses between neurons that fire synchronously are strengthened, forming cell assemblies and phase sequences. The former, on a shorter scale, are ensembles of synchronized cells that function transiently as a closed processing system; the latter, on a larger scale, correspond to the sequential activation of cell assemblies able to represent percepts and behaviors. Nowadays, the recording of large neuronal populations allows for the detection of multiple cell assemblies. Within Hebb's theory, the next logical step is the analysis of phase sequences. Here we detected phase sequences as consecutive assembly activation patterns, and then analyzed their graph attributes in relation to behavior. We investigated action potentials recorded from the adult rat hippocampus and neocortex before, during and after novel object exploration (experimental periods). Within assembly graphs, each assembly corresponded to a node, and each edge corresponded to the temporal sequence of consecutive node activations. The sum of all assembly activations was proportional to firing rates, but the activity of individual assemblies was not. Assembly repertoire was stable across experimental periods, suggesting that novel experience does not create new assemblies in the adult rat. Assembly graph attributes, on the other hand, varied significantly across behavioral states and experimental periods, and were separable enough to correctly classify experimental periods (Naïve Bayes classifier; maximum AUROCs ranging from 0.55 to 0.99) and behavioral states (waking, slow wave sleep, and rapid eye movement sleep; maximum AUROCs ranging from 0.64 to 0.98). Our findings agree with Hebb's view that assemblies correspond to primitive building blocks of representation, nearly unchanged in the adult, while phase sequences are labile across behavioral states and change after novel experience. The results are compatible with a role for phase sequences in behavior and cognition.

'Explicit Learning: an Effort towards Human Scheduling Algorithms'

Abstract Scheduling problems are generally NP-hard combinatorial problems, and a lot of research has been done to solve these problems heuristically. However, most of the previous approaches are problem-specific and research into the development of a general scheduling algorithm is still in its infancy. Mimicking the natural evolutionary process of the survival of the fittest, Genetic Algorithms (GAs) have attracted much attention in solving difficult scheduling problems in recent years. Some obstacles exist when using GAs: there is no canonical mechanism to deal with constraints, which are commonly met in most real-world scheduling problems, and small changes to a solution are difficult. To overcome both difficulties, indirect approaches have been presented (in [1] and [2]) for nurse scheduling and driver scheduling, where GAs are used by mapping the solution space, and separate decoding routines then build solutions to the original problem. In our previous indirect GAs, learning is implicit and is restricted to the efficient adjustment of weights for a set of rules that are used to construct schedules. The major limitation of those approaches is that they learn in a non-human way: like most existing construction algorithms, once the best weight combination is found, the rules used in the construction process are fixed at each iteration. However, normally a long sequence of moves is needed to construct a schedule and using fixed rules at each move is thus unreasonable and not coherent with human learning processes. When a human scheduler is working, he normally builds a schedule step by step following a set of rules. After much practice, the scheduler gradually masters the knowledge of which solution parts go well with others. He can identify good parts and is aware of the solution quality even if the scheduling process is not completed yet, thus having the ability to finish a schedule by using flexible, rather than fixed, rules. In this research we intend to design more human-like scheduling algorithms, by using ideas derived from Bayesian Optimization Algorithms (BOA) and Learning Classifier Systems (LCS) to implement explicit learning from past solutions. BOA can be applied to learn to identify good partial solutions and to complete them by building a Bayesian network of the joint distribution of solutions [3]. A Bayesian network is a directed acyclic graph with each node corresponding to one variable, and each variable corresponding to individual rule by which a schedule will be constructed step by step. The conditional probabilities are computed according to an initial set of promising solutions. Subsequently, each new instance for each node is generated by using the corresponding conditional probabilities, until values for all nodes have been generated. Another set of rule strings will be generated in this way, some of which will replace previous strings based on fitness selection. If stopping conditions are not met, the Bayesian network is updated again using the current set of good rule strings. The algorithm thereby tries to explicitly identify and mix promising building blocks. It should be noted that for most scheduling problems the structure of the network model is known and all the variables are fully observed. In this case, the goal of learning is to find the rule values that maximize the likelihood of the training data. Thus learning can amount to 'counting' in the case of multinomial distributions. In the LCS approach, each rule has its strength showing its current usefulness in the system, and this strength is constantly assessed [4]. To implement sophisticated learning based on previous solutions, an improved LCS-based algorithm is designed, which consists of the following three steps. The initialization step is to assign each rule at each stage a constant initial strength. Then rules are selected by using the Roulette Wheel strategy. The next step is to reinforce the strengths of the rules used in the previous solution, keeping the strength of unused rules unchanged. The selection step is to select fitter rules for the next generation. It is envisaged that the LCS part of the algorithm will be used as a hill climber to the BOA algorithm. This is exciting and ambitious research, which might provide the stepping-stone for a new class of scheduling algorithms. Data sets from nurse scheduling and mall problems will be used as test-beds. It is envisaged that once the concept has been proven successful, it will be implemented into general scheduling algorithms. It is also hoped that this research will give some preliminary answers about how to include human-like learning into scheduling algorithms and may therefore be of interest to researchers and practitioners in areas of scheduling and evolutionary computation. References 1. Aickelin, U. and Dowsland, K. (2003) 'Indirect Genetic Algorithm for a Nurse Scheduling Problem', Computer & Operational Research (in print). 2. Li, J. and Kwan, R.S.K. (2003), 'Fuzzy Genetic Algorithm for Driver Scheduling', European Journal of Operational Research 147(2): 334-344. 3. Pelikan, M., Goldberg, D. and Cantu-Paz, E. (1999) 'BOA: The Bayesian Optimization Algorithm', IlliGAL Report No 99003, University of Illinois. 4. Wilson, S. (1994) 'ZCS: A Zeroth-level Classifier System', Evolutionary Computation 2(1), pp 1-18.

The role of context in image annotation and recommendation

With the rise of smart phones, lifelogging devices (e.g. Google Glass) and popularity of image sharing websites (e.g. Flickr), users are capturing and sharing every aspect of their life online producing a wealth of visual content. Of these uploaded images, the majority are poorly annotated or exist in complete semantic isolation making the process of building retrieval systems difficult as one must firstly understand the meaning of an image in order to retrieve it. To alleviate this problem, many image sharing websites offer manual annotation tools which allow the user to “tag” their photos, however, these techniques are laborious and as a result have been poorly adopted; Sigurbjörnsson and van Zwol (2008) showed that 64% of images uploaded to Flickr are annotated with < 4 tags. Due to this, an entire body of research has focused on the automatic annotation of images (Hanbury, 2008; Smeulders et al., 2000; Zhang et al., 2012a) where one attempts to bridge the semantic gap between an image’s appearance and meaning e.g. the objects present. Despite two decades of research the semantic gap still largely exists and as a result automatic annotation models often offer unsatisfactory performance for industrial implementation. Further, these techniques can only annotate what they see, thus ignoring the “bigger picture” surrounding an image (e.g. its location, the event, the people present etc). Much work has therefore focused on building photo tag recommendation (PTR) methods which aid the user in the annotation process by suggesting tags related to those already present. These works have mainly focused on computing relationships between tags based on historical images e.g. that NY and timessquare co-exist in many images and are therefore highly correlated. However, tags are inherently noisy, sparse and ill-defined often resulting in poor PTR accuracy e.g. does NY refer to New York or New Year? This thesis proposes the exploitation of an image’s context which, unlike textual evidences, is always present, in order to alleviate this ambiguity in the tag recommendation process. Specifically we exploit the “what, who, where, when and how” of the image capture process in order to complement textual evidences in various photo tag recommendation and retrieval scenarios. In part II, we combine text, content-based (e.g. # of faces present) and contextual (e.g. day-of-the-week taken) signals for tag recommendation purposes, achieving up to a 75% improvement to precision@5 in comparison to a text-only TF-IDF baseline. We then consider external knowledge sources (i.e. Wikipedia & Twitter) as an alternative to (slower moving) Flickr in order to build recommendation models on, showing that similar accuracy could be achieved on these faster moving, yet entirely textual, datasets. In part II, we also highlight the merits of diversifying tag recommendation lists before discussing at length various problems with existing automatic image annotation and photo tag recommendation evaluation collections. In part III, we propose three new image retrieval scenarios, namely “visual event summarisation”, “image popularity prediction” and “lifelog summarisation”. In the first scenario, we attempt to produce a rank of relevant and diverse images for various news events by (i) removing irrelevant images such memes and visual duplicates (ii) before semantically clustering images based on the tweets in which they were originally posted. Using this approach, we were able to achieve over 50% precision for images in the top 5 ranks. In the second retrieval scenario, we show that by combining contextual and content-based features from images, we are able to predict if it will become “popular” (or not) with 74% accuracy, using an SVM classifier. Finally, in chapter 9 we employ blur detection and perceptual-hash clustering in order to remove noisy images from lifelogs, before combining visual and geo-temporal signals in order to capture a user’s “key moments” within their day. We believe that the results of this thesis show an important step towards building effective image retrieval models when there lacks sufficient textual content (i.e. a cold start).