Research on automated anomaly detection in complex systems by using log files has been on an upswing with the introduction of new deep-learning natural language processing methods. However, manually identifying and labelling anomalous logs is time-consuming, error-prone, and labor-intensive. This thesis instead uses an existing state-of-the-art method which learns from PU data as a baseline and evaluates three extensions to it. The first extension provides insight into the performance of the choice of word em-beddings on the downstream task. The second extension applies a re-labelling strategy to reduce problems from pseudo-labelling. The final extension removes the need for pseudo-labelling by applying a state-of-the-art loss function from the field of PU learning. The findings show that FastText and GloVe embeddings are viable options, with FastText providing faster training times but mixed results in terms of performance. It is shown that several of the methods studied in this thesis suffer from sporadically poor performances on one of the datasets studied. Finally, it is shown that using modified risk functions from the field of PU learning provides new state-of-the-art performances on the datasets considered in this thesis.

The RoboCup Rescue Robot and Simulation competitions have been held since 2000. The experience gained during these competitions has increased the maturity level of the field, which allowed deploying robots after real disasters (e.g. Fukushima Daiichi nuclear disaster). This article provides an overview of these competitions and highlights the state of the art and the lessons learned.

Over the years, the Artificial Intelligence and Integrated Computer Systems (AIICS) Division at Linköping University has developed a high-level architecture for collaborative robotic systems that includes a delegation system capable of defining complex missions to be executed by a team of agents. This architecture has been used as a part of a research arena for developing and improving public safety and security using ground, aerial, surfaceand underwater robotic systems. Recently, the division decided to purchase a Boston Dynamics Spot robot to further progress into the public safety and security research area.The robot has a robotic arm and navigation functionalities such as map building, motion planning, and obstacle avoidance. This thesis investigates how the Boston Dynamics Spot robot can be integrated into the high-level architecture for collaborative robotic systems from the AIICS division. Additionally, how the robot’s functionalities can be extended so that it is capable of determining which actions it should take to achieve high-level behavioursconsidering its capabilities and current state. In this context, higher-level behaviours include picking up and delivering first aid supplies, which can be beneficial in specific emergency situations. The study was divided and done in an iterative approach.The system was tested in various scenarios that represent its intended future use. The result demonstrated the robot’s ability to plan and accomplish the desired high-level behaviours. However, there were instances when achieving the desired behaviours proved challenging due to various limiting factors, including limitations posed by the robot’s internal controller.

Robots are increasingly expected to go beyond controlled environments in laboratories and factories, to act autonomously in real-world workplaces and public spaces. Autonomous robots navigating the real world have to contend with a great deal of uncertainty, which poses additional challenges. Uncertainty in the real world accrues from several sources. Some of it may originate from imperfect internal models of reality. Other uncertainty is inherent, a direct side effect of partial observability induced by sensor limitations and occlusions. Regardless of the source, the resulting decision problem is unfortunately computationally intractable under uncertainty. This poses a great challenge as the real world is also dynamic. It  will not pause while the robot computes a solution. Autonomous robots navigating among people, for example in traffic, need to be able to make split-second decisions. Uncertainty is therefore often neglected in practice, with potentially catastrophic consequences when something unexpected happens. The aim of this thesis is to leverage recent advances in machine learning to compute safe real-time approximations to decision-making under uncertainty for real-world robots. We explore a range of methods, from probabilistic to deep learning, as well as different combinations with optimization-based methods from robotics, planning and control. Driven by applications in robot navigation, and grounded in experiments with real autonomous quadcopters, we address several parts of this problem. From reducing uncertainty by learning better models, to directly approximating the decision problem itself, all the while attempting to satisfy both the safety and real-time requirements of real-world autonomy.

Robots are increasingly expected to go beyond controlled environments in laboratories and factories, to enter real-world public spaces and homes. However, robot behavior is still usually engineered for narrowly defined scenarios. To manually encode robot behavior that works within complex real world environments, such as busy work places or cluttered homes, can be a daunting task. In addition, such robots may require a high degree of autonomy to be practical, which imposes stringent requirements on safety and robustness. \setlength{\parindent}{2em}\setlength{\parskip}{0em}The aim of this thesis is to examine methods for automatically learning safe robot behavior, lowering the costs of synthesizing behavior for complex real-world situations. To avoid task-specific assumptions, we approach this from a data-driven machine learning perspective. The strength of machine learning is its generality, given sufficient data it can learn to approximate any task. However, being embodied agents in the real-world, robots pose a number of difficulties for machine learning. These include real-time requirements with limited computational resources, the cost and effort of operating and collecting data with real robots, as well as safety issues for both the robot and human bystanders.While machine learning is general by nature, overcoming the difficulties with real-world robots outlined above remains a challenge. In this thesis we look for a middle ground on robot learning, leveraging the strengths of both data-driven machine learning, as well as engineering techniques from robotics and control. This includes combing data-driven world models with fast techniques for planning motions under safety constraints, using machine learning to generalize such techniques to problems with high uncertainty, as well as using machine learning to find computationally efficient approximations for use on small embedded systems.We demonstrate such behavior synthesis techniques with real robots, solving a class of difficult dynamic collision avoidance problems under uncertainty, such as induced by the presence of humans without prior coordination. Initially using online planning offloaded to a desktop CPU, and ultimately as a deep neural network policy embedded on board a 7 quadcopter.

With the rise of deep reinforcement learning, there has also been a string of successes on continuous control problems using physics simulators. This has lead to some optimism regarding use in autonomous robots and vehicles. However, to successful apply such techniques to the real world requires a firm grasp of their limitations. As recent work has raised questions of how diverse these simulation benchmarks really are, we here instead analyze a popular deep RL approach on toy examples from robot obstacle avoidance. We find that these converge very slowly, if at all, to safe policies. We identify convergence issues on stochastic environments and local minima as problems that warrant more attention for safety-critical control applications.

A research arena (WARA-PS) for sensing, data fusion, user interaction, planning and control of collaborative autonomous aerial and surface vehicles in public safety applications is presented. The objective is to demonstrate scientific discoveries and to generate new directions for future research on autonomous systems for societal challenges. The enabler is a computational infrastructure with a core system architecture for industrial and academic collaboration. This includes a control and command system together with a framework for planning and executing tasks for unmanned surface vehicles and aerial vehicles. The motivating application for the demonstration is marine search and rescue operations. A state-of-art delegation framework for the mission planning together with three specific applications is also presented. The first one concerns model predictive control for cooperative rendezvous of autonomous unmanned aerial and surface vehicles. The second project is about learning to make safe real-time decisions under uncertainty for autonomous vehicles, and the third one is on robust terrain-aided navigation through sensor fusion and virtual reality tele-operation to support a GPS-free positioning system in marine environments. The research results have been experimentally evaluated and demonstrated to industry and public sector audiences at a marine test facility. It would be most difficult to do experiments on this large scale without the WARA-PS research arena. Furthermore, these demonstrator activities have resulted in effective research dissemination with high public visibility, business impact and new research collaborations between academia and industry.

Reinforcement learning for robot control tasks in continuous environments is a challenging problem due to the dimensionality of the state and action spaces, time and resource costs for learning with a real robot as well as constraints imposed for its safe operation. In this paper we propose a model-based reinforcement learning approach for continuous environments with constraints. The approach combines model-based reinforcement learning with recent advances in approximate optimal control. This results in a bounded-rationality agent that makes decisions in real-time by efficiently solving a sequence of constrained optimization problems on learned sparse Gaussian process models. Such a combination has several advantages. No high-dimensional policy needs to be computed or stored while the learning problem often reduces to a set of lower-dimensional models of the dynamics. In addition, hard constraints can easily be included and objectives can also be changed in real-time to allow for multiple or dynamic tasks. The efficacy of the approach is demonstrated on both an extended cart pole domain and a challenging quadcopter navigation task using real data.

A key requirement of autonomous vehicles is the capability to safely navigate in their environment. However, outside of controlled environments, safe navigation is a very difficult problem. In particular, the real-world often contains both complex 3D structure, and dynamic obstacles such as people or other vehicles. Dynamic obstacles are particularly challenging, as a principled solution requires planning trajectories with regard to both vehicle dynamics, and the motion of the obstacles. Additionally, the real-time requirements imposed by obstacle motion, coupled with real-world computational limitations, make classical optimality and completeness guarantees difficult to satisfy. We present a unified optimization-based motion planning and control solution, that can navigate in the presence of both static and dynamic obstacles. By combining optimal and receding-horizon control, with temporal multi-resolution lattices, we can precompute optimal motion primitives, and allow real-time planning of physically-feasible trajectories in complex environments with dynamic obstacles. We demonstrate the framework by solving difficult indoor 3D quadcopter navigation scenarios, where it is necessary to plan in time. Including waiting on, and taking detours around, the motions of other people and quadcopters.

Aerial robots hold great potential for aiding Search and Rescue (SAR) efforts over large areas, such as during natural disasters. Traditional approaches typically search an area exhaustively, thereby ignoring that the density of victims varies based on predictable factors, such as the terrain, population density and the type of disaster. We present a probabilistic model to automate SAR planning, with explicit minimization of the expected time to discovery. The proposed model is a spatial point process with three interacting spatial fields for i) the point patterns of persons in the area, ii) the probability of detecting persons and iii) the probability of injury. This structure allows inclusion of informative priors from e.g. geographic or cell phone traffic data, while falling back to latent Gaussian processes when priors are missing or inaccurate. To solve this problem in real-time, we propose a combination of fast approximate inference using Integrated Nested Laplace Approximation (INLA), and a novel Monte Carlo tree search tailored to the problem. Experiments using data simulated from real world Geographic Information System (GIS) maps show that the framework outperforms competing approaches, finding many more injured in the crucial first hours.

Modern optimization-based approaches to control increasingly allow automatic generation of complex behavior from only a model and an objective. Recent years has seen growing interest in fast solvers to also allow real-time operation on robots, but the computational cost of such trajectory optimization remains prohibitive for many applications. In this paper we examine a novel deep neural network approximation and validate it on a safe navigation problem with a real nano-quadcopter. As the risk of costly failures is a major concern with real robots, we propose a risk-aware resampling technique. Contrary to prior work this active learning approach is easy to use with existing solvers for trajectory optimization, as well as deep learning. We demonstrate the efficacy of the approach on a difficult collision avoidance problem with non-cooperative moving obstacles. Our findings indicate that the resulting neural network approximations are least 50 times faster than the trajectory optimizer while still satisfying the safety requirements. We demonstrate the potential of the approach by implementing a synthesized deep neural network policy on the nano-quadcopter microcontroller.

Robots are increasingly expected to move out of the controlled environment of research labs and into populated streets and workplaces. Collision avoidance in such cluttered and dynamic environments is of increasing importance as robots gain more autonomy. However, efficient avoidance is fundamentally difficult since computing safe trajectories may require considering both dynamics and uncertainty. While heuristics are often used in practice, we take a holistic stochastic trajectory optimization perspective that merges both collision avoidance and control. We examine dynamic obstacles moving without prior coordination, like pedestrians or vehicles. We find that common stochastic simplifications lead to poor approximations when obstacle behavior is difficult to predict. We instead compute efficient approximations by drawing upon techniques from machine learning. We propose to combine policy search with model-predictive control. This allows us to use recent fast constrained model-predictive control solvers, while gaining the stochastic properties of policy-based methods. We exploit recent advances in Bayesian optimization to efficiently solve the resulting probabilistically-constrained policy optimization problems. Finally, we present a real-time implementation of an obstacle avoiding controller for a quadcopter. We demonstrate the results in simulation as well as with real flight experiments.

Society today brings a high pace development and demand of Artificial intelligence systems as well as robotics. To further expand and to take one step closer to have Unmanned Aerial Vehicles (UAVs) working in the cities, the European Union Aviation Safety Agency launched a project that introduces U-space airspace, an airspace where UAVs, for instance, are allowed to operate for commercial services.The problems defined for U-space airspace resemble problems defined in the area of multi-agent path finding, such as scaling and traffic etc., resulting an interest to research whether MAPF-solutions can be applied to U-space scenarios. The following thesis extends the state-of-the-art MAPF-algorithm Continuous-time Conflict based search (CCBS) to handle simplified U-space scenarios, as well as extend other A*-based algorithms, such as a version of the Receding Horizon Lattice-based Motion Planning named Extended Multi-agent A* algorithm with Wait-Time (EMAWT) and an extended A* named Extended Multi-agent A* algorithm (EMA) to handle them. Comparisons of the three algorithms resulted in the EMAWT being the most reliable and stable solution throughout all tests, whilst for fewer agents, the CCBS being the clear best solution.

This study conducts an experimental evaluation of two hypotheses about the contributions of formal and semantic features to the grammatical gender assignment of nouns. One of the hypotheses (Corbett and Fraser 2000) claims that semantic features dominate formal ones. The other hypothesis, formulated within the optimal gender assignment theory (Rice 2006), states that form and semantics contribute equally. Both hypotheses claim that the combination of formal and semantic features yields the most accurate gender identification. In this paper, we operationalize and test these hypotheses by trying to predict grammatical gender using only character-based embeddings (that capture only formal features), only context-based embeddings (that capture only semantic features) and the combination of both. We performed the experiment using data from three languages with different gender systems (French, German and Russian). Formal features are a significantly better predictor of gender than semantic ones, and the difference in prediction accuracy is very large. Overall, formal features are also significantly better than the combination of form and semantics, but the difference is very small and the results for this comparison are not entirely consistent across languages.

Since its introduction, partial satisfaction planning (PSP), including both oversubscription (OSP) and net-benefit, has received significant attention in the classical planning community. However, hierarchical aspects have been mostly ignored in this context, although several problem domains that form the main motivation for PSP, such as the rover domain, have an inherent hierarchical structure. In this paper, we are taking the necessary steps for facilitating this research direction. First, we formally define hierarchical partial satisfaction planning problems and discuss the usefulness and necessity of this formalism. Second, we present a carefully structured set of benchmarks consisting of OSP and net-benefit problems with hierarchical structure. We describe and analyze the different domains of the benchmark set and the desiderata that are met to provide an interesting and challenging starting point for upcoming research. Third, we introduce various planning techniques that can solve hierarchical OSP problems and investigate their empirical behaviour on our proposed benchmark.

Air combat is a complex situation, training for it and analysis of possible tactics are time consuming and expensive. In order to circumvent those problems, mathematical models of air combat can be used. This thesis presents air combat as a one-on-one influence diagram game where the influence diagram allows the dynamics of the aircraft, the preferences of the pilots and the uncertainty of decision making in a structural and transparent way to be taken into account. To obtain the players’ game optimal control sequence with respect to their preferences, the influence diagram has to be solved. This is done by truncating the diagram with a moving horizon technique and determining and implementing the optimal controls for a dynamic game which only lasts a few time steps.

The result is a working air combat model, where a player estimates the probability that it resides in any of four possible states. The pilot’s preferences are modeled by utility functions, one for each possible state. In each time step, the players are maximizing the cumulative sum of the utilities for each state which each possible action gives. These are weighted with the corresponding probabilities. The model is demonstrated and evaluated in a few interesting aspects. The presented model offers a way of analyzing air combat tactics and maneuvering as well as a way of making autonomous decisions in for example air combat simulators. 

Man overboard is one of the most common and dangerous accidents that can occur whentraveling on a boat. Available research on man overboard systems with cameras have focusedon man overboard taking place from larger ships, which involves a fall from a height.Recreational boat manufacturers often use cord-based kill switches that turns of the engineif the wearer falls overboard. The aim of this thesis is to create a man overboard warningsystem based on state-of-the-art object detection models that can detect man overboard situationthrough inputs from a camera. Awell performing warning system would allow boatmanufactures to comply with safety regulations and expand the kill-switch coverage to allpassengers on the boat. Furthermore, the aim is also to create two new datasets: one dedicatedto human detection and one with man overboard fall sequences. YOLOv5 achievedthe highest performance on a new human detection dataset, with an average precision of97%. A Mobilenet-SSD-v1 network based on weights from training on the PASCAL VOCdataset and additional training on the new man overboard dataset is used as the detectionmodel in final warning system. The man overboard warning system achieves an accuracyof 50% at best, with a precision of 58% and recall of 78%.

Characters recognition in natural images is a challenging problem, asit involves segmenting characters of various colours on various background. Inthis article, we present a method for segmenting images that use a colour percep-tion graph. Our algorithm is inspired by graph cut segmentation techniques andit use an edge detection technique for filtering the graph before the graph-cut aswell as merging segments as a final step. We also present both qualitative andquantitative results, which show that our algorithm perform at slightly better andfaster to a state of the art algorithm.

Skeletonisation is a key process in character recognition in natural images. Under the assumption that a character is made of a stroke of uniform colour, with small variation in thickness, the process of recognising characters can be decomposed in the three steps. First the image is segmented, then each segment is transformed into a set of connected strokes (skeletonisation), which are then abstracted in a descriptor that can be used to recognise the character. The main issue with skeletonisation is the sensitivity with noise, and especially, the presence of holes in the masks. In this article, a new method for the extraction of strokes is presented, which address the problem of holes in the mask and does not use any parameters.

We present a general framework to estimate the parameters of both a robot and landmarks in 3D. It relies on the use of a stochastic gradient descent method for the optimisation of the nodes in a graph of weak constraints where the landmarks and robot poses are the nodes. Then a belief propagation method combined with covariance intersection is used to estimate the uncertainties of the nodes. The first part of the article describes what is needed to define a constraint and a node models, how those models are used to update the parameters and the uncertainties of the nodes. The second part present the models used for robot poses and interest points, as well as simulation results.

In the context of collaborative robotics, distributed situation awareness is essential for supporting collective intelligence in teams of robots and human agents where it can be used for both individual and collective decision support. This is particularly important in applications pertaining to emergency rescue and crisis management. During operational missions, data and knowledge is gathered incrementally and in different ways by heterogeneous robots and humans. The purpose of this paper is to describe an RDF Graph Synchronization System called RGS⊕. It is assumed that a dynamic set of agents provide or retrieve knowledge stored in their local RDF Graphs which are continuously synchronized between agents. The RGS⊕ System was designed to handle unreliable communication and does not rely on a static centralized infrastructure. It is capable of synchronizing knowledge as timely as possible and allows agents to access knowledge while it is incrementally acquired. A deeper empirical analysis of the RGS⊕ System is provided that shows both its efficiency and efficacy.

This paper presents a model-driven approach to detect image line segments. The approach incrementally detects segments on the gradient image using a linear Kalman filter that estimates the supporting line parameters and their associated variances. The algorithm is fast and robust with respect to image noise and illumination variations, it allows the detection of longer line segments than data-driven approaches, and does not require any tedious parameters tuning. An extension of the algorithm that exploits a pyramidal approach to enhance the quality of results is proposed. Results with varying scene illumination and comparisons to classic existing approaches are presented.

In this work we are investigating reactive avoidance techniques which can be used on board of a small quadcopter and which do not require absolute localisation. We propose a local map representation which can be updated with proprioceptive sensors. The local map is centred around the robot and uses spherical coordinates to represent a point cloud. The local map is updated using a depth sensor, the Inertial Measurement Unit and a registration algorithm. We propose an extension of the Dynamic Window Approach to compute a velocity vector based on the current local map. We propose to use an OctoMap structure to compute a 2-pass A* which provide a path which is converted to a velocity vector. Both approaches are reactive as they only make use of local information. The algorithms were evaluated in a simulator which offers a realistic environment, both in terms of control and sensors. The results obtained were also validated by running the algorithms on a real platform.

In this paper we consider a problem of scanningan outdoor area with a team of heterogeneous Unmanned AirVehicles (UAVs) equipped with different sensors (e.g. LIDARs).Depending on the availability of the UAV platforms and themission requirements there is a need to either minimise thetotal mission time or to maximise certain properties of thescan output, such as the point cloud density. The key challengeis to divide the scanning task among UAVs while taking intoaccount the differences in capabilities between platforms andsensors. Additionally, the system should be able to ensure thatconstraints such as limit on the flight time are not violated.We present an approach that uses an optimisation techniqueto find a solution by dividing the area between platforms,generating efficient scan trajectories and selecting flight andscanning parameters, such as velocity and flight altitude. Thismethod has been extensively tested on a large set of randomlygenerated scanning missions covering a wide range of realisticscenarios as well as in real flights.

Soft skills are recognized as crucial for engineers as technical work is becoming more and more collaborative and interdisciplinary. Today many engineering educations fail to give appropriate training in soft skills. Linköping University has therefore developed a completely new course “Professionalism for Engineers” for two of its 5-year engineering programs in the area of computer science. The course stretches over the first 3 years with students from the three years taking it together. The purpose of the course is to give engineering students training in soft skills that are of importance during the engineering education as well as during their professional career. The examination is based on the Dialogue Seminar Method developed for learning from experience and through reflection. The organization of the course is innovative in many ways.

This thesis describes a method to generate new trajectories based on historic positiondata for a given geographical area. The thesis uses AIS-data from fishing boats to first describe a method that uses DBSCAN and OPTICS algorithms to cluster the data into clustersbased on routes where the boats travel and areas where the boats fish.Here bayesian optimization has been utilized to search for parameters for the clusteringalgorithms. In this given scenario it was shown DBSCAN is better in all fields, but it hasmany points where OPTICS has the potential to become better if it was modified a bit.This is followed by a method describing how to take the clusters and build a nodenetwork that then can be traversed using a path finding algorithm combined with internalrules to generate new routes that can be used in simulations to give a realistic enoughsituation picture. Finally a method to evaluate these generated routes are described andused to compare the routes to each other

Validation and verification in machine learning is an open problem which becomes increasingly important as its applications becomes more critical. Amongst the applications are autonomous vehicles and medical diagnostics. These systems all needs to be validated before being put into use or else the consequences might be fatal.

This master’s thesis focuses on improving both learning and validating machine learning models in cases where data can either be generated or collected based on a chosen position. This can for example be taking and labeling photos at the position or running some simulation which generates data from the chosen positions.

The approach is twofold. The first part concerns modeling the relationship between any fixed-size set of positions and some real valued performance measure. The second part involves calculating such a performance measure by estimating the performance over a region of positions.

The result is two different algorithms, both variations of Bayesian optimization. The first algorithm models the relationship between a set of points and some performance measure while also optimizing the function and thus finding the set of points which yields the highest performance. The second algorithm uses Bayesian optimization to approximate the integral of performance over the region of interest. The resulting algorithms are validated in two different simulated environments.

The resulting algorithms are applicable not only to machine learning but can also be used to optimize any function which takes a set of positions and returns a value, but are more suitable when the function is expensive to evaluate.

Validation and verification of supervised machine learning models is becoming increasingly important as their complexity and range of applications grows. This paper describes an extension to Bayesian optimization which allows for selecting both training and validation data, in cases where data can be generated or calculated as a function of a spatial location.

Radiologists' workload has been steadily increasing for decades. As digital technology matures it improves the workflow for radiology departments and decreases the time necessary to examine patients. Computer systems are widely used in health care and are for example used to view radiology images. To simplify this, display protocols based on examination data are used to automatically create a layout and hang images for the user. To cover a wide variety of examinations hundreds of protocols must be created, which is a time-consuming task and the system can still fail to hang series if strict requirements on the protocols are not met. To remove the need for this manual step we propose to use machine learning based on past manually corrected presentations. The classifiers are trained on the metadata in the examination and how the radiologist preferred to hang the series. The chosen approach was to create classifiers for different layout rules and then use these predictions in an algorithm for assigning series types to individual image slots according to categories based on metadata, similar to how display protocol works. The resulting presentations shows that the system is able to learn, but must increase its prediction accuracy if it is to be used commercially. Analyses of the different parts show that increased accuracy in early steps should improve overall success.

The propositional model counting problem #SAT asks to compute the number of satisfying as- signments for a given propositional formula. Recently, three #SAT proof systems kcps (knowledge compilation proof system), MICE (model counting induction by claim extension), and CPOG (certified partitioned-operation graphs) have been introduced with the aim to model #SAT solving and enable proof logging for solvers.

Prior to this paper, the relations between these proof systems have been unclear and very few proof complexity results are known. We completely determine the simulation order of the three systems, establishing that CPOG simulates both MICE and kcps, while MICE and kcps are exponentially incomparable. This implies that CPOG is strictly stronger than the other two systems.

n/a

Contemporary systems situated in real-world open environments frequently have to cope with incomplete and inconsistent information that typically increases complexity of reasoning and decision processes. Realistic modeling of such informationally complex environments calls for nuanced tools. In particular, incomplete and inconsistent information should neither trivialize nor stop both reasoning or planning. The paper introduces ACTLOG, a rule-based four-valued language designed to specify actions in a paraconsistent and paracomplete manner. ACTLOG is an extension of 4QL(Bel), a language for reasoning with paraconsistent belief bases. Each belief base stores multiple world representations. In this context, ACTLOGs action may be seen as a belief bases transformer. In contrast to other approaches, ACTLOG actions can be executed even when the underlying belief base contents is inconsistent and/or partial. ACTLOG provides a nuanced action specification tools, allowing for subtle interplay among various forms of nonmonotonic, paraconsistent, paracomplete and doxastic reasoning methods applicable in informationally complex environments. Despite its rich modeling possibilities, it remains tractable. ACTLOG permits for composite actions by using sequential and parallel compositions as well as conditional specifications. The framework is illustrated on a decontamination case study known from the literature.

This paper introduces 4QL(Bel), a four-valued rule language designed for reasoning with paraconsistent and paracomplete belief bases as well as belief structures. Belief bases consist of finite sets of ground literals providing (partial and possibly inconsistent) complementary or alternative views of the world. As introduced earlier, belief structures consist of constituents, epistemic profiles and consequents. Constituents and consequents are belief bases playing different roles. Agents perceive the world forming their constituents, which are further transformed into consequents via the agents or groups epistemic profile. In order to construct 4QL(Bel), we extend 4QL, a four-valued rule language permitting for many forms of reasoning, including doxastic reasoning. Despite the expressiveness of 4QL(Bel), we show that its tractability is retained.

The paper introduces ActLog, a rule-based language capable of specifying actions paraconsistently. ActLog is an extension of 4QL Bel " role="presentation"> Bel , a rule-based language for reasoning with paraconsistent and paracomplete belief bases and belief structures. Actions considered in the paper act on belief bases rather than states represented as sets of ground literals. Each belief base stores multiple world representations which can be though of as a representation of possible states. In this context ActLog’s action may be then seen as a method of transforming one belief base into another. In contrast to other approaches, ActLog permits to execute actions even if the underlying belief base state is partial or inconsistent. Finally, the framework introduced in this paper is tractable.

Dealing with heterogeneous information sources and reasoning techniques allowing for incomplete and inconsistent information is one of current challenges in the area of knowledge representation and reasoning. We advocate for 4QL, a rule-based query language, as a proper tool allowing one to address these challenges. To justify this point of view we discuss a rescue robotics scenario for which a simulator has been developed and tested. In particular, we present a planner using 4QL and, therefore, capable to deal with lack of knowledge and inconsistencies. Through the case study we show that our approach allows one to use lightweight knowledge representation tools: due to the use of 4QL tractability of modeling and reasoning is guaranteed and high usability is achieved.

We propose a visualization system for incident commanders in urban search~\&~rescue scenarios that supports access path planning for post-disaster structures. Utilizing point cloud data acquired from unmanned robots, we provide methods for assessment of automatically generated paths. As data uncertainty and a priori unknown information make fully automated systems impractical, we present a set of viable access paths, based on varying risk factors, in a 3D environment combined with the visual analysis tools enabling informed decisions and trade-offs. Based on these decisions, a responder is guided along the path by the incident commander, who can interactively annotate and reevaluate the acquired point cloud to react to the dynamics of the situation. We describe design considerations for our system, technical realizations, and discuss the results of an expert evaluation.

The ACL Anthology is a prime resource for research papers within computational linguistics and natural language processing, while continuing to be an open-source and community-driven project. Since Gildea et al. (2018) reported on its state and planned directions, the Anthology has seen major technical changes. We discuss what led to these changes and how they impact long-term maintainability and community engagement, describe which open-source data and software tools the Anthology currently provides, and provide a survey of literature that has used the Anthology as a main data source.

Recently,a simple but powerful language for expressing and learning general policies and problem decompositions (sketches) has been introduced in terms of rules defined over a set of Boolean and numerical features. In this work, we consider three extensions of this language aimed at making policies and sketches more flexible and reusable: internal memory states, as in finite state controllers; indexical features, whose values are a function of the state and a number of internal registers that can be loaded with objects; and modules that wrap up policies and sketches and allow them to call each other by passing parameters. In addition, unlike general policies that select state transitions rather than ground actions, the new language allows for the selection of such actions. The expressive power of the resulting language for policies and sketches is illustrated through a number of examples.

It has been observed that many classical planning domains with atomic goals can be solved by means of a simple polynomial exploration procedure, called IW, that runs in time exponential in the problem width, which in these cases is bounded and small. Yet, while the notion of width has become part of state-of-the-art planning algorithms such as BFWS, there is no good explanation for why so many benchmark domains have bounded width when atomic goals are considered. In this work, we address this question by relating bounded width with the existence of general optimal policies that in each planning instance are represented by tuples of atoms of bounded size. We also define the notions of (explicit) serializations and serialized width that have a broader scope, as many domains have a bounded serialized width but no bounded width. Such problems are solved nonoptimally in polynomial time by a variant of the Serialized IW algorithm. Finally, the language of general policies and the semantics of serializations are combined to yield a simple, meaningful, and expressive language for specifying serializations in compact form in the form of sketches, which can be used for encoding domain control knowledge by hand or for learning it from examples. Sketches express general problem decompositions in terms of subgoals, and terminating sketches of bounded width express problem decompositions that can be solved in polynomial time.

In the last decade, there has been a growing interest in applying AI technologies to implement complex data analytics over data streams. To this end, researchers in various fields have been organising a yearly event called the "Stream Reasoning Workshop" to share perspectives, challenges, and experiences around this topic.

In this paper, the previous organisers of the workshops and other community members provide a summary of the main research results that have been discussed during the first six editions of the event. These results can be categorised into four main research areas: The first is concerned with the technological challenges related to handling large data streams. The second area aims at adapting and extending existing semantic technologies to data streams. The third and fourth areas focus on how to implement reasoning techniques, either considering deductive or inductive techniques, to extract new and valuable knowledge from the data in the stream.

This summary is written not only to provide a crystallisation of the field, but also to point out distinctive traits of the stream reasoning community. Moreover, it also provides a foundation for future research by enumerating a list of use cases and open challenges, to stimulate others to join this exciting research area.

This thesis investigates how pre-trained contextualized language models can be adapted for multi-label text classification of Swedish news articles. Various classifiers are built on pre-trained BERT and ELECTRA models, exploring global and local classifier approaches. Furthermore, the effects of domain specialization, using additional metadata features and model compression are investigated. Several hundred thousand news articles are gathered to create unlabeled and labeled datasets for pre-training and fine-tuning, respectively. The findings show that a local classifier approach is superior to a global classifier approach and that BERT outperforms ELECTRA significantly. Notably, a baseline classifier built on SVMs yields competitive performance. The effect of further in-domain pre-training varies; ELECTRA’s performance improves while BERT’s is largely unaffected. It is found that utilizing metadata features in combination with text representations improves performance. Both BERT and ELECTRA exhibit robustness to quantization and pruning, allowing model sizes to be cut in half without any performance loss.

This article offers an exploratory quantitative analysis of the conceptual career of climate in US English over the period 1800-2010. Our aim is to qualify two, closely related arguments circulating in Environmental Humanities scholarship regarding the concepts history, namely that we only started to think of climate as a global entity after the introduction of general circulation models during the final quarter of the twentieth century, and, second, that climatic change only became an issue of environmental concern once scientists began to approach climate as a global model. While we do not dispute that the computer revolution resulted in a significantly new understanding of climate, our analysis points to a longer process of singularization and growing abstraction starting in the early nineteenth century that might help to nuance and deepen insights developed in environmental history.

Cavity filters are a necessary component in base stations used for telecommunication. Without these filters it would not be possible for base stations to send and receive signals at the same time. Today these cavity filters require fine tuning by humans before they can be deployed. This thesis have designed and implemented a neural network that can tune cavity filters. Different types of design parameters have been evaluated, such as neural network architecture, data presentation and data preprocessing. While the results was not comparable to human fine tuning, it was shown that there was a relationship between error and number of weights in the neural network. The thesis also presents some rules of thumb for future designs of neural network used for filter tuning.

The self-rationalising capabilities of LLMs are appealing because the generated explanations can give insights into the plausibility of the predictions. However, how faithful the explanations are to the predictions is questionable, raising the need to explore the patterns behind them further. To this end, we propose a hypothesis-driven statistical framework. We use a Bayesian network to implement a hypothesis about how a task (in our example, natural language inference) is solved, and its internal states are translated into natural language with templates. Those explanations are then compared to LLM-generated free-text explanations using automatic and human evaluations. This allows us to judge how similar the LLM’s and the Bayesian network’s decision processes are. We demonstrate the usage of our framework with an example hypothesis and two realisations in Bayesian networks. The resulting models do not exhibit a strong similarity to GPT-3.5. We discuss the implications of this as well as the framework’s potential to approximate LLM decisions better in future work.

Spatio-temporal reasoning is the area of automated reasoning about space and time and is important in the field of robotics. It is desirable for an autonomous robot to have the ability to reason about both time and space. ST0 is a logic that allows for such reasoning by, among other things, defining a formalism used to describe the relationship between spatial regions and a calculus that allows for deducing further information regarding such spatial relations. An extension of ST0 is ST1 that can be used to describe the relationship between spatial entities across time-points (inter-temporal relations) while ST0 is constrained to doing so within a single time-point. This allows for a better ability of expressing how spatial entities change over time. A major obstacle in using ST1 in practise however, is the fact that any observations made regarding spatial relations between regions is constrained to the time-point in which the observation was made, so we are unable to observe inter-temporal relations. Further complicating things is the fact that deducing such inter-temporal relations is not possible without a frame of reference. This thesis examines one method of overcoming these problems by considering the concept of rigid regions which are assumed to always be unchanging and using them as the frame of reference, or as landmarks. The effectiveness of this method is studied by conducting experiments where a comparison is made between various landmark ratios with respect to the total number of regions under consideration. Results show that when a high degree of intra-temporal relations are fully or partially known, increasing the number of landmark regions will reduce the percentage of inter-temporal relations to be completely unknown. Despite this, very few inter-temporal relations can be fully determined even with a high ratio of landmark regions.

We consider the directed Steiner tree problem (DSTP) with a constraint on the total number of arcs (hops) in the tree. This problem is known to be NP-hard, and therefore, only heuristics can be applied in the case of its large-scale instances.   For the hop-constrained DSTP, we propose local search strategies aimed at improving any heuristically produced initial Steiner tree. They are based on solving a sequence of hop-constrained shortest path problems for which we have recently developed ecient label correcting algorithms.   The presented approach is applied to nding suitable 3D locations where unmanned aerial vehicles (UAVs) can be placed to relay information gathered in multi-target monitoring and surveillance. The eciency of our algorithms is illustrated by results of numerical experiments involving problem instances with up to 40 000 nodes and up to 20 million arcs.

We consider the directed Steiner tree problem (DSTP) with a constraint on the total number of arcs (hops) in the tree. This problem is known to be NP-hard, and therefore, only heuristics can be applied in the case of its large-scale instances.For the hop-constrained DSTP, we propose local search strategies aimed at improving any heuristically produced initial Steiner tree. They are based on solving a sequence of hop-constrained shortest path problems for which we have recently developed efficient label correcting algorithms.The presented approach is applied to finding suitable 3D locations where unmanned aerial vehicles (UAVs) can be placed to relay information gathered in multi-target monitoring and surveillance. The efficiency of our algorithms is illustrated by results of numerical experiments involving problem instances with up to 40 000 nodes and up to 20 million arcs.

Guarding the perimeter of an area in order to detect potential intruders is an important task in a variety of security-related applications. This task can in many circumstances be performed by a set of camera-equipped unmanned aerial vehicles (UAVs). Such UAVs will occasionally require refueling or recharging, in which case they must temporarily be replaced by other UAVs in order to maintain complete surveillance of the perimeter. In this paper we consider the problem of scheduling such replacements. We present optimal replacement strategies and justify their optimality.

Guarding the perimeter of an area in order to detect potential intruders is an important task in a variety of security-related applications. This task can in many circumstances be performed by a set of camera-equipped unmanned aerial vehicles (UAVs). Such UAVs will occasionally require refueling or recharging, in which case they must temporarily be replaced by other UAVs in order to maintain complete surveillance of the perimeter. In this paper we consider the problem of scheduling such replacements. We present optimal replacement strategies and justify their optimality.