Computer Science - Journal Articles

Permanent URI for this collection


Recent Submissions

Now showing 1 - 5 of 174
  • Item
    RECCE: Deep reinforcement learning for joint routing and scheduling in time-constrained wireless networks
    (Institute of Electrical and Electronics Engineers (IEEE), 2021-09-22) Chilukuri, Shanti; Pesch, Dirk; Horizon 2020; Science Foundation Ireland
    Time Division Multiple Access-based Medium Access Control protocols tend to be the choice for wireless networks that require deterministic delay guarantees, as is the case in many Industrial Internet of Things (IIoT) applications. As the optimal joint scheduling and routing problem for multi-hop wireless networks is NP-hard, heuristics are generally used for building schedules. However, heuristics normally result in sub-optimal schedules, which may result in packets missing their deadlines. In this paper, we present RECCE, a deep REinforcement learning method for joint routing and sCheduling in time-ConstrainEd networks with centralised control. During training, RECCE considers multiple routes and criteria for scheduling in any given time slot and channel in a multi-channel, multi-hop wireless network. This allows RECCE to explore and learn routes and schedules to deliver more packets within the deadline. Simulation results show that RECCE can reduce the number of packets missing the deadline by as much as 55% and increase schedulability by up to 30%, both relative to the best baseline heuristic. RECCE can deal well with dynamic network conditions, performing better than the best baseline heuristic in up to 74% of the scenarios in the training set and in up to 64% of scenarios not in the training set.
  • Item
    Generation and prediction of difficult model counting instances
    (2022-12-06) Escamocher, Guillaume; O'Sullivan, Barry; Science Foundation Ireland; European Regional Development Fund
    We present a way to create small yet difficult model counting instances. Our generator is highly parameterizable: the number of variables of the instances it produces, as well as their number of clauses and the number of literals in each clause, can all be set to any value. Our instances have been tested on state of the art model counters, against other difficult model counting instances, in the Model Counting Competition. The smallest unsolved instances of the competition, both in terms of number of variables and number of clauses, were ours. We also observe a peak of difficulty when fixing the number of variables and varying the number of clauses, in both random instances and instances built by our generator. Using these results, we predict the parameter values for which the hardest to count instances will occur.
  • Item
    Estimating and evaluating the uncertainty of rating predictions and top-n recommendations in recommender systems
    (Association for Computing Machinery, 2023-02-16) Coscrato, Victor; Bridge, Derek; Science Foundation Ireland; European Regional Development Fund
    Uncertainty is a characteristic of every data-driven application, including recommender systems. The quantification of uncertainty can be key to increasing user trust in recommendations or choosing which recommendations should be accompanied by an explanation; and uncertainty estimates can be used to accomplish recommender tasks such as active learning and co-training. Many uncertainty estimators are available but, to date, the literature has lacked a comprehensive survey and a detailed comparison. In this paper, we fulfil these needs. We review the existing methods for uncertainty estimation and metrics for evaluating uncertainty estimates, while also proposing some estimation methods and evaluation metrics of our own. Using two datasets, we compare the methods using the evaluation metrics that we describe, and we discuss their strengths and potential issues. The goal of this work is to provide a foundation to the field of uncertainty estimation in recommender systems, on which further research can be built.
  • Item
    5G NR-V2X: Toward connected and cooperative autonomous driving
    (Institute of Electrical and Electronics Engineers (IEEE), 2021-03-31) Bagheri, Hamidreza; Noor-A-Rahim, Md; Liu, Zilong; Lee, Haeyoung; Pesch, Dirk; Moessner, Klaus; Xiao, Pei; Horizon 2020; Science Foundation Ireland; European Regional Development Fund
    5G New Radio (NR) is touted as a pivotal enabling technology for the genuine realization of connected and cooperative autonomous driving. Despite numerous research efforts in recent years, a systematic overview on the role of 5G NR in future connected autonomous communication networks is missing. To fill this gap and to spark more future research, this article introduces the technology components of 5G NR and discusses the evolution from existing cellular vehicle-to-everything (V2X) technology toward NR-V2X. We primarily focus on the key features and functionalities of the physical layer, sidelink communication and its resource allocation, architecture flexibility, security and privacy mechanisms, and precise positioning techniques. Moreover, we envisage and highlight the potential of machine learning for further performance enhancement in NR-V2X services. Lastly, we show how 5G NR can be configured to support advanced V2X use cases.
  • Item
    Low-dimensional space modeling-based differential evolution for large scale global optimization problems
    (Institute of Electrical and Electronics Engineers (IEEE), 2022-12-07) Fonseca, Thiago Henrique Lemos; Nassar, Silvia Modesto; de Oliveira, Alexandre César Muniz; Agard, Bruno; Coordenação de Aperfeiçoamento de Pessoal de Nível Superior; Science Foundation Ireland; European Regional Development Fund
    Large-Scale Global Optimization (LSGO) has been an active research field. Part of this interest is supported by its application to cutting-edge research such as Deep Learning, Big Data, and complex real-world problems such as image encryption, real-time traffic management, and more. However, the high dimensionality makes solving LSGO a significant challenge. Some recent research deal with the high dimensionality by mapping the optimization process to a reduced alternative space. Nonetheless, these works suffer from the changes in the search space topology and the loss of information caused by the dimensionality reduction. This paper proposes a hybrid metaheuristic, so-called LSMDE (Low-dimensional Space Modeling-based Differential Evolution), that uses the Singular Value Decomposition to build a low-dimensional search space from the features of candidate solutions generated by a new SHADE-based algorithm (GM-SHADE). GM-SHADE combines a Gaussian Mixture Model (GMM) and two specialized local algorithms: MTS-LS1 and L-BFGS-B, to promote a better exploration of the reduced search space. GMM mitigates the loss of information in mapping high-dimensional individuals to low-dimensional individuals. Furthermore, the proposal does not require prior knowledge of the search space topology, which makes it more flexible and adaptable to different LSGO problems. The results indicate that LSMDE is the most efficient method to deal with partially separable functions compared to other state-of-the-art algorithms and has the best overall performance in two of the three proposed experiments. Experimental results also show that the new approach achieves competitive results for non-separable and overlapping functions on the most recent test suite for LSGO problems.