Federated learning (FL) trains machine learning models over heterogeneous and decentralized datasets. Communication between client and server can be a major bottleneck for FL, especially in cases of large models. Moreover, real-world FL problems often involve data heterogeneity issues such as class imbalance. We propose an approach to address these issues of class balance and communication efficiency in Federated Learning. Our strategy is based on two key elements: 1. a novel adaptive voting-based federated dropout on client models addressing communication bottlenecks and data heterogeneity, and 2. a heterogeneous quantization method that can adjust to clients' bandwidth requirements. We conduct experiments across several datasets and models demonstrating that these two components work together to balance the trade-off between communication costs and model performance with clients having heterogeneous communication bandwidth. Importantly, our approach improves performance on imbalanced datasets like CIFAR-10-LT and CIFAR-100-LT, which is critical for addressing class imbalance in federated learning. On CIFAR-10 We get approximately a seven-factor reduction in communication cost without degrading the quality of the model.

Processing big data poses a significant challenge when transitioning from sequential to distributed code, primarily due to the extensive scale at which data is handled. This complexity arises from both syntax and semantic differences between the codes. Unfortunately, current methods are inefficient and require more effective automated solutions. To address this problem, we utilized the Transformer-based BERT model because of its exceptional capability to understand and capture deep contextual relationships in large datasets. Our method involved creating two comprehensive datasets containing 10k and 100k sequential code snippets paired with their corresponding PySpark API calls. We optimized the BERT model by fine-tuning it to predict distributed API calls for previously unseen sequential snippets. For the 10k dataset, the model demonstrated robustness, achieving a training accuracy of 99.36%, a test accuracy of 99.7%, a Balanced Accuracy of 99.78%, and an F1 Score of 0.997. In contrast, the 100k dataset metrics were equally impressive, with a training accuracy of 89.48%, a test accuracy of 99.99%, a Balanced Accuracy of 99.98%, and an F1 Score of 0.999. Our work demonstrates the feasibility of automating the sequential-to-distributed code transition with notable precision.

Online learning updates models incrementally with new data, avoiding large storage requirements and costly model recalculations. In this paper, we introduce “OLR-WA; OnLine Regression with Weighted Average”, a novel and versatile multivariate online linear regression model. We also investigate scenarios involving drift, where the underlying patterns in the data evolve over time, conduct convergence analysis, and compare our approach with existing online regression models. The results of OLR-WA demonstrate its ability to achieve performance comparable to the batch regression, while also showcasing comparable or superior performance when compared with other state-of-the-art online models, thus establishing its effectiveness. Moreover, OLR-WA exhibits exceptional performance in terms of rapid convergence, surpassing other online models with consistently achieving high r2values as a performance measure from the first iteration to the last iteration, even when initialized with minimal amount of data points, as little as 1% to 10% of the total data points. In addition to its ability to handle time-based (temporal drift) scenarios, remarkably, OLR-WA stands out as the only model capable of effectively managing confidence-based challenging scenarios. It achieves this by adopting a conservative approach in its updates, giving priority to older data points with higher confidence levels. In summary, OLR-WA’s performance further solidifies its versatility and utility across different contexts, making it a valuable solution for online linear regression tasks.

In the typical database system, an execution is correct if it is equivalent to some serial execution. This criterion, called serializability, is unacceptable for new database applications which require long-duration transactions. We present a new transaction model which allows correctness criteria more suitable for these applications. This model combines three enhancements to the standard model: nested transactions, explicit predicates, and multiple versions. These features yield the name of the new model, nested transactions with predicates and versions, or NT/PV. The modular nature of the NT/PV model allows a straightforward representation of simple systems. It also provides a formal framework for describing complex interactions. The most complex interactions the model allows can be captured by a protocol which exploits all of the semantics available to the NT/PV model. An example of these interactions is shown in a CASE application. The example shows how a system based on the NT/PV model is superior to both standard database techniques and unrestricted systems in both correctness and performance.

An example of a software development application is considered, and the formal model of H. Korth and G. Speegle (1988) is applied to show how this example could be represented as a set of database transactions. It is shown that, although the standard notion of correctness (serializability) is too strict, the notion of correctness in the Korth and Speegle model allows sufficient concurrency with acceptable overhead. An extrapolation is made from this example to draw some conclusions regarding the potential usefulness of a formal approach to the management of long-duration design transactions.>

In the classical approach to transaction processing, a concurrent execution is considered to be correct if it is equivalent to a non-concurrent schedule. This notion of correctness is called serializability. Serializability has proven to be a highly useful concept for transaction systems for data-processing style applications. Recent interest in applying database concepts to applications in computer-aided design, office information systems, etc. has resulted in transactions of relatively long duration. For such transactions, there are serious consequences to requiring serializability as the notion of correctness. Specifically, such systems either impose long-duration waits or require the abortion of long transactions. In this paper, we define a transaction model that allows for several alternative notions of correctness without the requirement of serializability. After introducing the model, we investigate classes of schedules for transactions. We show that these classes are richer than analogous classes under the classical model. Finally, we show the potential practicality of our model by describing protocols that permit a transaction manager to allow correct non-serializable executions