Monitoring in-service aerodynamic performance of airliners is critical for operational efficiency and safety, yet presents significant challenges when using operational Quick Access Recorder (QAR) data due to sensor noise, low excitation, and the absence of key model parameters like moments of inertia. These constraints render conventional state-propagation and recursive estimation methods unsuitable. To address these challenges, this paper proposes and validates the Constant-Gain Equation-Error Method (CG-EEM), a robust framework tailored for QAR data analysis. The CG-EEM employs a custom constant-gain estimator that avoids both the infeasibility of state-propagation filters and the premature convergence or instability issues of standard recursive algorithms. Extensive validation on a multi-fleet dataset of over 200 flights demonstrates that the framework produces highly consistent and physically meaningful aerodynamic parameters. Crucially, follow-up work has verified that this approach successfully resolves the fundamental thrust-drag ambiguity problem, ensuring the estimates are not just plausible, but physically unique and correct. This demonstrates that CG-EEM is a scalable and computationally efficient tool for reliable fleet-wide performance monitoring and early detection of airframe degradation.

Deep learning trends to use low precision numeral formats to cope with the ever-growing model sizes. For example, the large language model LLaMA2 has been widely deployed in 4-bit precision. With larger models and fewer unique values caused by low precision, an increasing proportion of arithmetic in matrix multiplication is repeating. Although discussed in prior works, such value redundancy has not been fully exploited, and the cost to leverage the value redundancy often offsets any advantages. In this paper, we propose to primitivize the matrix multiplication, that is decomposing it down to the 1-ary successor function (a.k.a. counting) to merge repeating arithmetic. We revisited various techniques to propose Cambricon-C SA, a 4-bit primitive matrix multiplication unit that doubles the energy efficiency over conventional systolic arrays. Experimental results show that Cambricon-C SA can achieve$\mathbf{1}.\mathbf{95}\times$energy efficiency improvement compared with MAC-based systolic array.

The firefly algorithm is a stochastic meta-heuristic algorithm that incorporates randomness into a search process. In essence, the randomness is useful when determining the next point in the search space and therefore has a crucial impact when exploring the new solution. Simultaneously, randomized mechanism plays an important role in balance the exploration and exploitation during the process. In this paper, an extensive comparison is made between 8 different probability distributions that can be used for randomizing the firefly algorithm's attractive mechanism, e.g., Uniform distribution, Gaussian distribution, Exponential distribution, Cauchy distribution, and so on. In our experiments, variously randomized firefly algorithms are developed and extensive experiments are conducted on 13-benchmark functions. The results of these experiments show that these randomized mechanisms can improve the convergence rate and the robustness of the firefly algorithm significantly.