Entropy-Driven Stochastic Federated Learning in Non-IID 6G Edge-RAN

Scalable and sustainable AI-driven analytics are necessary to enable large-scale and heterogeneous service deployment in sixth-generation (6G) ultra-dense networks. This implies that the exchange of raw monitoring data should be minimized across the network by bringing the analysis functions closer to the data collection points. While federated learning (FL) is an efficient tool to implement such a decentralized strategy, real networks are generally characterized by time- and space-varying traffic patterns and channel conditions, making thereby the data collected in different points non independent and identically distributed (non-IID), which is challenging for FL. To sidestep this issue, we first introduce a new a priori metric that we call dataset entropy, whose role is to capture the distribution, the quantity of information, the unbalanced structure and the “non-IIDness” of a dataset independently of the models. This a priori entropy is calculated using a multi-dimensional spectral clustering scheme over both the features and the supervised output spaces, and is suitable for classification as well as regression tasks. The FL aggregation operations support system (OSS) server then uses the reported dataset entropies to devise 1) an entropy-based federated averaging scheme, and 2) a stochastic participant selection policy to significantly stabilize the training, minimize the convergence time, and reduce the corresponding computation cost. Numerical results are provided to show the superiority of these novel approaches.

Multicriteria Decision-Making Method and Application in the Setting of Trapezoidal Neutrosophic Z-Numbers

The information expression and modeling of decision-making are critical problems in the fuzzy decision theory and method. However, existing trapezoidal neutrosophic numbers (TrNNs) and neutrosophic Z-numbers (NZNs) and their multicriteria decision-making (MDM) methods reveal their insufficiencies, such as without considering the reliability measures in TrNN and continuous Z-numbers in NZN. To overcome the insufficiencies, it is necessary that one needs to propose trapezoidal neutrosophic Z-numbers (TrNZNs), their aggregation operations, and an MDM method for solving MDM problems with TrNZN information. Hence, this study first proposes a TrNZN set, some basic operations of TrNZNs, and the score and accuracy functions of TrNZN and their ranking laws. Then, the TrNZN weighted arithmetic averaging (TrNZNWAA) and TrNZN weighted geometric averaging (TrNZNWGA) operators are presented based on the operations of TrNZNs. Next, an MDM approach using the proposed aggregation operators and score and accuracy functions is established to carry out MDM problems under the environment of TrNZNs. In the end, the established MDM approach is applied to an MDM example of software selection for revealing its rationality and efficiency in the setting of TrNZNs. The main advantage of this study is that the established approach not only makes assessment information continuous and reliable but also strengthens the decision rationality and efficiency in the setting of TrNZNs.

hep_tables: Heterogeneous Array Programming for HEP

Array operations are one of the most concise ways of expressing common filtering and simple aggregation operations that are the hallmark of a particle physics analysis: selection, filtering, basic vector operations, and filling histograms. The High Luminosity run of the Large Hadron Collider (HL-LHC), scheduled to start in 2026, will require physicists to regularly skim datasets that are over a PB in size, and repeatedly run over datasets that are 100’s of TB’s – too big to fit in memory. Declarative programming techniques are a way of separating the intent of the physicist from the mechanics of finding the data and using distributed computing to process and make histograms. This paper describes a library that implements a declarative distributed framework based on array programming. This prototype library provides a framework for different sub-systems to cooperate in producing plots via plug-in’s. This prototype has a ServiceX data-delivery sub-system and an awkward array sub-system cooperating to generate requested data or plots. The ServiceX system runs against ATLAS xAOD data and flat ROOT TTree’s and awkward on the columnar data produced by ServiceX.

Empirical Evaluation of NoSQL and Relational Database Systems

Aim: To evaluate the performance of Relational and NoSQL databases in terms of execution time and memory consumption during operations involving structured data. Objective: To outline the criteria that decision makers should consider while making a choice of the database most suited to an application. Methods: Extensive experiments were performed on MySQL, MongoDB, Cassandra, Redis using the data for a IMDB movies schema prorated into 4 datasets of 1000, 10000, 25000 and 50000 records. The experiments involved typical database operations of insertion, deletion, update read of records with and without indexing as well as aggregation operations. Databases’ performance has been evaluated by measuring the time taken for operations and computing memory usage. Results: * Redis provides the best performance for write, update and delete operations in terms of time elapsed and memory usage whereas MongoDB gives the worst performance when the size of data increases, due to its locking mechanism. * For the read operations, Redis provides better performance in terms of latency than Cassandra and MongoDB. MySQL shows worst performance due to its relational architecture. On the other hand, MongoDB shows the best performance among all databases in terms of efficient memory usage. * Indexing improves the performance of any database only for covered queries. * Redis and MongoDB give good performance for range based queries and for fetching complete data in terms of elapsed time whereas MySQL gives the worst performance. * MySQL provides better performance for aggregate functions. NoSQL is not suitable for complex queries and aggregate functions. Conclusion: It has been found from the extensive empirical analysis that NoSQL outperforms SQL based systems in terms of basic read and write operations. However, SQL based systems are better if queries on the dataset mainly involves aggregation operations.