Accurate Throughput Prediction (TP) represents a real challenge for reliable adaptive streaming in challenging mediums, such as cellular networks. State-of-the-art solutions adopt Deep Learning (DL) models to improve TP accuracy for various multimedia systems. This article illustrates that designing black-box TP engines that depend solely on the model’s capacity and power of learning does not achieve consistent accuracy across all throughput ranges. Additionally, we propose MATURE, a novel multistage DL-based TP model designed to capture network operating context to improve prediction accuracy. MATURE’s prediction involves characterizing the operating context before estimating the network throughput. We show that MATURE delivers consistent, accurate prediction for all throughput ranges in both 4G and 5G networks. We also show that light-weight MATURE models that use quantized parameters maintain their accuracy while featuring up to 100× faster inference, thus making them suitable for mobile implementation. Our real video streaming experiments further show that MATURE improves the average user Quality of Experience by up to 20% when compared to other TP methods.

Fully-automated optimal controller design for engineering systems is a challenging task. While, optimization-based, automated control parameter tuning techniques have been widely discussed in the literature, most works do not discuss expert knowledge requirements for system design, which result in significant human intervention. In this work, we discuss a multistage controller tuning framework for decentralized control that highlights expert knowledge requirements in automated controller design. We propose a methodology to automate the input-output pairing and stage definition steps in the framework using Large Language Models (LLMs) for a family of multi-tank benchmarks. We achieve this by proposing a mathematical language to describe the system and design an algorithm to bind this mathematical representation to the input prompt space of an LLM. We demonstrate that our methodology can produce consistent expert knowledge outputs from the LLM with over 97% accuracy for the multi-tank benchmarks. We also empirically show that, correct stage definition by the LLM can improve tuned controller performance by up to 52%.

Fully-automated optimal design of Cyber-Physical Systems is a challenging task that involves multiple components. In this paper, we are interested in the automated design of one of these components: the system controller. State-of-the-art con-trol parameter automated tuning techniques are optimization-based. However, for high-dimension control parameter spaces, computational costs can be high. We present a multistage controller tuning framework that decomposes controller tuning into sub-tasks, each with a reduced-dimension search space. We show formally that this framework reduces the sample complexity of the control-tuning task. We empirically validate this result by applying a Bayesian optimization approach to tuning multiple PID controllers in an unmanned underwater vehicle benchmark system. We demonstrate an 86 % decrease in computational time and a 36 % decrease in sample complex-ity. Furthermore, the proposed framework highlights existing challenges in fully automated control parameter tuning.

Accurate Throughput Prediction (TP) represents a cornerstone for reliable adaptive streaming in challenging mediums, such as cellular networks. Challenged by the highly dynamic wireless medium, recent state-of-the-art solutions adopt Deep Learning (DL) models to improve TP accuracy. However, these models perform poorly in critical, rare network conditions, leading to degraded user Quality of Experience (QoE). Such performance results from depending solely on the model's capacity and power of learning, without integrating system knowledge into the design. In this paper, we propose MATURE, a novel multi-stage DL-based TP model designed to capture network operating context to improve prediction accuracy and user experience. MATURE's operation involves characterising the operating context before estimating the network throughput. Our performance evaluation shows that MATURE improves the average user QoE by 4% - 90% in critical network conditions when compared to state-of-the-art.

The Boolean Satisfiability Problem (SAT) can be framed as a binary classification task. Recently, numerous machine and deep learning techniques have been successfully deployed to predict whether a CNF has a solution. However, these approaches do not provide a variables assignment when the instance is satisfiable and have not been used as part of SAT solvers. In this work, we investigate the possibility of using a machine-learning SAT/UNSAT classifier to assign a truth value to a variable. A heuristic solver can be created by iteratively assigning one variable to the value that leads to higher predicted satisfiability. We test our approach with and without probing features and compare it to a heuristic assignment based on the variable's purity. We consider as objective the maximisation of the number of literals fixed before making the CNF unsatisfiable. The preliminary results show that this iterative procedure can consistently fix variables without compromising the formula's satisfiability, finding a complete assignment in almost all test instances.

Cyberattacks against Industrial Control Systems (ICS) can have harmful physical impacts. Investigating such attacks can be difficult, as evidence could be lost to physical damage. This is especially true with stealthy attacks ; i.e., attacks that can evade detection. In this paper, we aim to engineer Forensic Readiness (FR) in safety-critical, geographically distributed ICS, by proactively collecting potential evidence of stealthy attacks. The collection of all data generated by an ICS at all times is infeasible due to the large volume of such data. Hence, our approach only triggers data collection when there is the possibility for a potential stealthy attack to cause damage. We determine the conditions for such an event by performing predictive, model-based, safety checks. Furthermore, we use the geographical layout of the ICS and the safety predictions to identify data that is at risk of being lost due to damage, i.e., relevant data. Finally, to reduce the control performance overhead resulting from real-time data collection, we select a subset of relevant data to collect by performing a trade-off between expected impact of the attack and the estimated cost of collection. We demonstrate these ideas using simulations of the widely-used Tennessee–Eastman Process (TEP) benchmark. We show that the proposed approach does not miss relevant data and results in a reduced control performance overhead compared to the case when all data generated by the ICS is collected. We also showcase the applicability of our approach in improving the efficiency of existing ICS forensic log analysis tools.

Stealthy attackson Industrial Control Systems can cause significant damage while evading detection. In this article, instead of focusing on the detection of stealthy attacks, we aim to provide early warnings to operators, in order to avoid physical damage and preserve in advance data that may serve as an evidence during an investigation. We propose a framework to providegrounds for suspicion, i.e., preliminary indicators reflecting the likelihood of success of a stealthy attack. We propose two grounds for suspicion based on the behaviour of the physical process: (i)feasibilityof a stealthy attack, and (ii)proximityto unsafe operating regions. We propose a metric to measure grounds for suspicion in real-time and provide soundness principles to ensure that such a metric is consistent with the grounds for suspicion. We apply our framework to Linear Time-Invariant (LTI) systems and formulate the suspicion metric computation as a real-time reachability problem. We validate our framework on a case study involving the benchmark Tennessee-Eastman process. We show through numerical simulation that we can provide early warnings well before a potential stealthy attack can cause damage, while incurring minimal load on the network. Finally, we apply our framework on a use case to illustrate its usefulness in supporting early evidence collection.

The control of complex autonomous systems has significantly improved in recent years and unmanned aerial vehicles (UAVs) have become popular in the research community. Although the use of UAVs is increasing, much work remains to guarantee fault- tolerant control (FTC) properties of these vehicles. Model-based controllers are the standard way to control UAVs, however obtaining models of the system and environment for every possible operating condition a UAV can experience in a real-world scenario is not feasible. Reinforcement Learning has shown promise in controlling complex systems but requires training in a simulator (requiring a model) of the system. Further, stability guarantees do not exist for learning-based controllers, which limits their large scale application in the real-world. We propose a novel hybrid FTC approach that uses a learned supervisory controller (together with low-level PID controllers) with key stability guarantees. We use a robust reinforcement learning approach to learn the supervisory control parameters and prove stability. We empirically validate our framework using trajectory-following experiments (in simulation) for a quadcopter subject to rotor faults, wind disturbances, and severe position and attitude noise.

A broad range of real-world systems can be defined using discrete-time hybrid systems, e.g., chemical process plants and manufacturing systems. We characterize this application domain using a class of discrete-event systems, max-plus linear discrete-event systems, which captures synchronization without concurrency or selection. The model framework of these hybrid systems is non-linear in a conventional algebra, but linear in the max-plus algebra, thereby enabling linear-time inference. We use an observer-based framework for monitoring and diagnosing max-plus diagnostics models, and further improve computational efficiency by searching over only the most-likely space of behaviours. We illustrate our approach using a chemical process-control example.

We conduct a comparative study between two approaches for combining signals from several MPCs designed for different fault scenarios. The first is MPC switching where a switch dictates which of the MPC controllers is currently active. The second is MPC mixing where all MPCs are running concurrently and their outputs are blended in proportion to the current estimate of fault state. We demonstrate results using a gravity drained multi-tank system. Our empirical results show that the mixing approach responds more quickly to faults than the switching approach. Further, we show that the speed and accuracy of fault isolation has a critical impact on fault tolerance.

In epidemic modeling, state filtering is an excellent tool for enhancing the performance of traditional epidemic models. We introduce a novel state filter algorithm to further improve the performance of state-of-the-art approaches based on Susceptible-Infected-Recovered (SIR) models. The proposed algorithm merges two techniques, which are typically used separately: linear correction, as seen in the Ensemble Kalman Filter (EnKF), and resampling, as used in the Particle Filter (PF). We compare the inferential accuracy of our approach against the EnKF and the Ensemble Adjustment Kalman Filter (EAKF), using algorithms employing both an uncentered covariance matrix (UCM) and the standard column-centered covariance matrix (CCM). Our algorithm requires O(DN) more time than EnKF does, where D is the ensemble dimension and N denotes the ensemble size. We demonstrate empirically that our algorithm with UCM achieves the lowest root-mean-square-error (RMSE) and the highest correlation coefficient (CORR) amongst the selected methods, in 11 out of 14 major real-world scenarios. We show that the EnKF with UCM outperforms the EnKF with CCM, while the EAKF gains better accuracy with CCM in most scenarios.

The Model-Based Diagnosis (MBD) framework developed by Reiter has been a strong theoretical foundation for MBD, yet is limited to models that are described in terms of logical sentences. We propose a more general framework that covers a wide range of modelling languages, ranging from AI-based languages (e.g., logic and Bayesian networks) to FDI-based languages (e.g., linear Gaussian models). We show that a graph-theoretic basis for decomposable system models can be augmented with several languages and corresponding inference algorithms based on valuation algebras.

Fault diagnosis of analogue linear systems poses many challenges, such as the size of the search space that must be explored and the possibility of simulation instabilities introduced by particular fault classes. We study a novel algorithm that addresses both problems. This algorithm dynamically modifies the simulation model during diagnosis by pruning parametrized components that cause discontinuity in the model. We provide a theoretical framework for predicting the speedups, which depends on the topology of the model. We empirically validate the theoretical predictions through extensive experimentation on a benchmark of circuits.

A model-based diagnosis problem occurs when an observation is inconsistent with the assumption that the diagnosed system is not faulty. The task of a diagnosis engine is to compute diagnoses, which are assumptions on the health of components in the diagnosed system that explain the observation. In this paper, we extend Reiter's well-known theory of diagnosis by exploiting the duality of the relation between conflicts and diagnoses. This duality means that a diagnosis is a hitting set of conflicts, but a conflict is also a hitting set of diagnoses. We use this property to interleave the search for diagnoses and conflicts: a set of conflicts can guide the search for diagnosis, and the computed diagnoses can guide the search for more conflicts. We provide the formal basis for this dual conflict-diagnosis relation, and propose a novel diagnosis algorithm that exploits this duality. Experimental results show that the new algorithm is able to find a minimal cardinality diagnosis faster than the well-known Conflict-Directed A*.

Data sets acquired from complex scientific simulation, high precision engineering experiment and high-speed computer network have been exponentially increased, and visualization and analysis of such large-scale of data sets have been identified as a significant challenge to the visualization community. Over the past years many scientists have made attempt to address this problem by proposing various data reduction techniques. Consequently the size of data can be reduced and issues associated to the visualization can be improved (e.g. real-time interaction and visual overload). However, during the process of data reduction, the information of original data sets was approximated and potential errors were introduced. It leads to a new problem with regard to the integrity of the data and might mislead users for incorrect decision making. Therefore in this paper we aim to solve the problem by introducing three novel uncertainty visualization methods, which depict both the multi-resolution (MR) approximations of the original data set and the errors associated with each of its low resolution representations. As a result we faithfully represent the MR data sets and allow users to make suitable decisions from the visual output. We applied our techniques on a data set from medical domain to demonstrate their effectiveness and usability.

MOTIVATION: Conserved patterns across a multiple sequence alignment can be visualized by generating sequence logos. Sequence logos show each column in the alignment as stacks of symbol(s) where the height of a stack is proportional to its informational content, whereas the height of each symbol within the stack is proportional to its frequency in the column. Sequence logos use symbols of either nucleotide or amino acid alphabets. However, certain regulatory signals in messenger RNA (mRNA) act as combinations of codons. Yet no tool is available for visualization of conserved codon patterns. RESULTS: We present the first application which allows visualization of conserved regions in a multiple sequence alignment in the context of codons. CodonLogo is based on WebLogo3 and uses the same heuristics but treats codons as inseparable units of a 64-letter alphabet. CodonLogo can discriminate patterns of codon conservation from patterns of nucleotide conservation that appear indistinguishable in standard sequence logos. AVAILABILITY: The CodonLogo source code and its implementation (in a local version of the Galaxy Browser) are available at http://recode.ucc.ie/CodonLogo and through the Galaxy Tool Shed at http://toolshed.g2.bx.psu.edu/.

The aim of a modern Building Automation System (BAS) is to enhance interactive control strategies for energy efficiency and user comfort. In this context, we develop a novel control algorithm that uses a stochastic building occupancy model to improve mean energy efficiency while minimizing expected discomfort. We compare by simulation our Stochastic Model Predictive Control (SMPC) strategy to the standard heating control method to empirically demonstrate a 4.3% reduction in energy use and 38.3% reduction in expected discomfort.

We propose a StochAstic Fault diagnosis AlgoRIthm, called SAFARI, which trades off guarantees of computing minimal diagnoses for computational efficiency. We empirically demonstrate, using the 74XXX and ISCAS-85 suites of benchmark combinatorial circuits, that SAFARI achieves several orders-of-magnitude speedup over two well-known deterministic algorithms, CDA* and HA*, for multiple-fault diagnoses; further, SAFARI can compute a range of multiple-fault diagnoses that CDA* and HA* cannot. We also prove that SAFARI is optimal for a range of propositional fault models, such as the widely-used weak-fault models (models with ignorance of abnormal behavior). We discuss the optimality of SAFARI in a class of strong-fault circuit models with stuck-at failure modes. By modeling the algorithm itself as a Markov chain, we provide exact bounds on the minimality of the diagnosis computed. SAFARI also displays strong anytime behavior, and will return a diagnosis after any non-trivial inference time.

Model-based diagnostic reasoning often leads to a large number of diagnostic hypotheses. The set of diagnoses can be reduced by taking into account extra observations (passive monitoring), measuring additional variables (probing) or executing additional tests (sequential diagnosis/test sequencing). In this paper we combine the above approaches with techniques from Automated Test Pattern Generation (ATPG) and Model-Based Diagnosis (MBD) into a framework called FRACTAL (FRamework for ACtive Testing ALgorithms). Apart from the inputs and outputs that connect a system to its environment, in active testing we consider additional input variables to which a sequence of test vectors can be supplied. We address the computationally hard problem of computing optimal control assignments (as defined in FRACTAL) in terms of a greedy approximation algorithm called FRACTAL-G. We compare the decrease in the number of remaining minimal cardinality diagnoses of FRACTAL-G to that of two more FRACTAL algorithms: FRACTAL-ATPG and FRACTAL-P. FRACTAL-ATPG is based on ATPG and sequential diagnosis while FRACTAL-P is based on probing and, although not an active testing algorithm, provides a baseline for comparing the lower bound on the number of reachable diagnoses for the FRACTAL algorithms. We empirically evaluate the trade-offs of the three FRACTAL algorithms by performing extensive experimentation on the ISCAS85/74XXX benchmark of combinational circuits.

The six papers in this special issue cover different approaches and different applications to model-based diagnostics.

It is critical to use automated generators for synthetic models and data given the sparsity of benchmark models for empirical analysis and the cost of generating models by hand. We describe an automated generator for benchmark models that is based on using a compositional modeling framework and employs graphical models for the system topology. We propose a three-step process for synthetic model generation: 1) domain analysis; 2) topology generation; and 3) system-level behavioral model generation. To demonstrate our approach on two highly different domains, we generate models using this process for circuits drawn from the International Symposium on Circuits and Systems benchmark suite and a process-control system. We then analyze the synthetic models according to two criteria: 1) topological fidelity and 2) diagnostic efficiency. Based on this comparison, we identify parameters necessary for the autogenerated models to generate benchmark diagnosis circuit and process-control models with realistic properties.

We propose a methodology for overcoming the current approach of writing diagnostics code for industrial automation applications after the system is designed, which results in significant extra effort/cost, and potential discrepancies between design and diagnostics output. We show how we can automatically generate diagnostics from a more complex simulation model. We show how a model-transformation framework can transform a hybrid-systems simulation model into a propositional-logic diagnostics model with appropriate transformation rules. We illustrate our approach with an example from the domain of control for building lighting systems.

In embedded software development, model-driven design is well recognized. In this article we describe a compositional model-driven approach for auto-generating embedded code for improving the energy efficiency of building automation applications. We show how we can use a component-based hybrid-systems modelling framework to generate models for simulation and verification. Then, we auto-generate embeddable code from these models. We empirically demonstrate this approach using the hybrid-systems tool Charon, for the domain of energy-efficient lighting control for smart buildings. The paper provides a detailed description of the code generation steps and outlines some results obtained through a simple lighting example.

This article formalises the dual problem to model-based diagnosis (MBD), i.e., generating tests to isolate multiple simultaneous faults. Using a standard propositional MBD framework, we first define a test of minimal size that can isolate multiple simultaneous faults of an arbitrary nature. Second, we prove complexity results for multiplefault tests of minimal size in propositional system models, showing such problems have complexity similar to those of MBD problems, i.e., complexity at the second level of the polynomial hierarchy.

Two approaches have been used to perform approximate inference in Bayesian networks for which exact inference is infeasible: employing an approximation algorithm, or approximating the structure. In this article we compare two structure-approximation techniques, edge-deletion and approximate structure learning based on sub-sampling, in terms of relative accuracy and computational efficiency. Our empirical results indicate that edge-deletion techniques dominate the subsampling/induction strategy, in both accuracy and performance of generating the approximate network. We show, for several large Bayesian networks, how edge-deletion can create approximate networks with order-of-magnitude inference speedups and relatively little loss of accuracy.

The weighted constraint satisfaction problem (WCSP) is a popular formalism for encoding instances of hard optimization problems. The common approach to solving WCSP is branch-and-bound (BB), whose efficiency strongly depends on the method of computing a lower bound (LB) associated with the current node of the search tree. Two of the most important approaches for computing LB include (1) using local inconsistency counts, such as maintaining directed arc-consistency (MDAC), and (2) Russian Doll search (RDS). In this paper we present two BB-based algorithms. The first algorithm extends RDS. The second algorithm combines RDS and MDAC in an adaptive manner. We empirically demonstrate that the WCSP solver combining the above two algorithms outperforms both RDS and MDAC, over all the problem domains and instances we studied. To the best of our knowledge this is the first attempt to combine these two methodologies of computing LB for a BB-based algorithm.

Modern computing technologies-hardware, software, and algorithmic-have enabled the deployment of more exacting diagnostic reasoning (DR) systems than has heretofore been possible. Compromises in algorithm and modeling paradigm complexity, due to computational throughput and state-space explosion constraints, have historically dominated practical applications of such systems. This paper describes approaches that have been shown to be applicable in a wide set of domains. The algorithms used are highly scaleable and support a symbolic modeling formalism for analyzing the properties of the complex, dynamic systems. Moreover, analysis of simultaneous failures occurs as a natural byproduct of this formalism.

We describe a new paradigm for implementing inference in belief networks, which consists of two steps: (1) compiling a belief network into an arithmetic expression called a Query DAG (Q-DAG); and (2) answering queries using a simple evaluation algorithm. Each node of a Q-DAG represents a numeric operation, a number, or a symbol for evidence. Each leaf node of a Q-DAG represents the answer to a network query, that is, the probability of some event of interest. It appears that Q-DAGs can be generated using any of the standard algorithms for exact inference in belief networks (we show how they can be generated using clustering and conditioning algorithms). The time and space complexity of a Q-DAG generation algorithm is no worse than the time complexity of the inference algorithm on which it is based. The complexity of a Q-DAG evaluation algorithm is linear in the size of the Q-DAG, and such inference amounts to a standard evaluation of the arithmetic expression it represents. The intended value of Q-DAGs is in reducing the software and hardware resources required to utilize belief networks in on-line, real-world applications. The proposed framework also facilitates the development of on-line inference on different software and hardware platforms due to the simplicity of the Q-DAG evaluation algorithm. Interestingly enough, Q-DAGs were found to serve other purposes: simple techniques for reducing Q-DAGs tend to subsume relatively complex optimization techniques for belief-network inference, such as network-pruning and computation-caching.

Bayesian belief networks are being increasingly used as a knowledge representation for reasoning under uncertainty. Some researchers have questioned the practicality of obtaining the numerical probabilities with sufficient precision to create belief networks for large-scale applications. In this work, we investigate how precise the probabilities need to be by measuring how imprecision in the probabilities affects diagnostic performance. We conducted a series of experiments on a set of real-world belief networks for medical diagnosis in liver and bile disease. We examined the effects on diagnostic performance of (1) varying the mappings from qualitative frequency weights into numerical probabilities, (2) adding random noise to the numerical probabilities, (3) simplifying from quaternary domains for diseases and findings—absent, mild, moderate, and severe—to binary domains—absent and present, and (4) using test cases that contain diseases outside the network. We found that even extreme differences in the probability mappings and large amounts of noise lead to only modest reductions in diagnostic performance. We found no significant effect of the simplification from quaternary to binary representation. We also found that outside diseases degraded performance modestly. Overall, these findings indicate that even highly imprecise input probabilities may not impair diagnostic performance significantly, and that simple binary representations may often be adequate. These findings of robustness suggest that belief networks are a practical representation without requiring undue precision.

In many domains, the ability to use a knowledge base to automatically construct alternative probabilistic network models and then compare them is desirable. This paper makes two novel contributions towards achieving that goal: first, it analyzes a parameterized class of (a) static, and (b) temporal influence diagram, models which differ in the time-series process describing the temporal evolution of the system being modeled. Second, it applies general scoring metrics for comparing these models with respect to predictive accuracy and computational efficiency. The network rankings facilitate comparing the accuracy/efficiency tradeoffs entailed in using TIDs which differ in (1) the accuracy of capturing the temporal evolution of a dynamic system and (2) data and computational requirements. The scoring metrics are used to compare networks in which all variables evolve according to a Markov process with two novel domain-dependent network approximations. These approximations model the evolution of a parsimonious subset of variables rather than all variables.>

Computing diagnoses in domains with continuously changing data is difficult but essential aspect of solving many problems. To address this task, a dynamic influence diagram (ID) construction and updating system (DYNASTY) and its application to constructing a decision-theoretic model to diagnose acute abdominal pain, which is a domain in which the findings evolve during the diagnostic process, are described. For a system that evolves over time, DYNASTY constructs a parsimonious ID and then dynamically updates the ID, rather than constructing a new network from scratch for every time interval. In addition, DYNASTY contains algorithms that test the sensitivity of the constructed network's system parameters. The main contributions are: (1) presenting an efficient temporal influence diagram technique based on parsimonious model construction; and (2) formalizing the principles underlying a diagnostic tool for acute abdominal pain that explicitly models time-varying findings.>

This reply to papers by Pearl and Shafer focuses on two issues underlying the debate on the validity of using Dempster-Shafer theory, namely the requirement of a process-independent semantics and the a priori need for multiple uncertainty calculi. Pearl shows deficiencies of Dempster-Shafer theory in dealing with several instances of commonsense reasoning in a process-independent manner. Although this argument is correct under the assumptions stated, it is weakened somewhat by introducing questions of whether a process-independent semantics is always necessary or desirable. Another issue underlying both papers, whether multiple uncertainty representations are necessary, is also discussed. Shafer claims that multiple uncertainty representations are necessary. He presents a goal of developing all uncertainty representations in parallel and defining domains in which each representation is best suited. In contrast, Pearl implicitly claims that probability theory alone is necessary, unless the use of another representation (such as Dempster-Shafer theory) is shown to be clearly advantageous. These two perspectives lead to different approaches to defining the form of uncertainty best modeled by Dempster-Shafer theory or any other uncertainty calculus.