Search

With the ever-evolving concerns on privacy protection, vertical federated learning (FL), where participants own non-overlapping features for the same set of instances, is becoming a heated topic since it enables multiple enterprises to strengthen the machine learning models collaboratively with privacy guarantees. Nevertheless, to achieve privacy preservation, vertical FL algorithms involve complicated training routines and time-consuming cryptography operations, leading to slow training speed. This paper explores the efficiency of the gradient boosting decision tree (GBDT) algorithm under the vertical FL setting. Specifically, we introduce VF^2Boost, a novel and efficient vertical federated GBDT system. Significant solutions are developed to tackle the major bottlenecks. First, to handle the deficiency caused by frequent mutual-waiting in federated training, we propose a concurrent training protocol to reduce the idle periods. Second, to speed up the cryptography operations, we analyze the characteristics of the algorithm and propose customized operations. Empirical results show that our system can be 12.8-18.9 times faster than the existing vertical federated implementations and support much larger datasets.

Feature selection is an important problem in machine learning and data mining. In reality, the wrapper methods are broadly used in feature selection. It treats feature selection as a search problem using a predictor as a black-box. However, most wrapper methods are time-consuming due to the large search space. In this paper, we propose a novel wrapper method, called FeatureBand, for feature selection. We use the early stopping strategy to terminate bad candidate feature subsets and avoid wasting of training time. Further, we use a genetic local search to generate new subsets based on previous ones. These two techniques are combined under an iterative framework in which we gradually allocate more resources for more promising candidate feature subsets. The experimental result shows that FeatureBand achieves a better trade-off between search time and search accuracy. It is 1.45\(\times \) to 17.6\(\times \) faster than the state-of-the-art wrapper-based methods without accuracy loss.

Manufacturing processes may be viewed as operational units in the overall manufacturing/production system. Changes in technology, production patterns, and process settings are typically made at the process (unit) level. Very often, environmental problems are apparent when the manufacturing/production system is viewed as a whole (many units joined together). However, with the units aggregated to form a system, it is often difficult to identify the source of an environmental problem or judge the singular effect of changes to a process unit; the changes become lost in the complexity of the system as a whole. Recent efforts have employed input–output modeling to describe the flow of materials and the environmental consequences associated with manufacturing processes. A method is introduced for aggregating process-level material input–output models to form a combined material input–output model for a manufacturing system. The resulting model serves as a bridge between unit-level changes and broader system behaviors. The model form permits identification of opportunities for reducing environmental impacts at the process level (e.g., reduction of emissions, waste generation, and material use) and driving the system toward zero emissions based on an examination of the aggregated manufacturing system level model. Case studies are used to illustrate the application of the aggregated material input–output model to minimize waste and resource consumptions.

This paper proposes a position tracking control approach of AGV's platform using neural network The control problem of AGV's active suspension is first analysed. Then a neural network control scheme for controlling active suspension is presented. The controller adopts a multilayer back-propagation neural network To obtain the fastest possible speed of convergence and to meet the need of real-time control, a prediction-correction method for adjusting learning parameters is developed. Finally, various manoeuvres of the AGV are simulated to evaluate the performance of the controller. The simulation results indicate that the neural network control scheme is feasible. The neuro-controller can accommodate the variation of operational conditions and environment.