Automatic parameter tuning methods for planning algorithms, which integrate pipeline approaches with learning-based techniques, are regarded as promising due to their stability and capability to handle highly constrained environments. While existing parameter tuning methods have demonstrated considerable success, further performance improvements require a more structured approach. In this paper, we propose a hierarchical architecture for reinforcement learning-based parameter tuning. The architecture introduces a hierarchical structure with low-frequency parameter tuning, mid-frequency planning, and high-frequency control, enabling concurrent enhancement of both upper-layer parameter tuning and lower-layer control through iterative training. Experimental evaluations in both simulated and real-world environments show that our method surpasses existing parameter tuning approaches. Furthermore, our approach achieves first place in the Benchmark for Autonomous Robot Navigation (BARN) Challenge.

Smart contracts are a core component of blockchain ecosystems, but their transparency and immutability make them vulnerable to attacks, leading to significant financial losses. Thus, repairing vulnerabilities in smart contracts is crucial for establishing a trustworthy blockchain environment. Existing smart contract vulnerability repair methods suffer from a critical "one-for-all" design limitation, where a single model is tasked with fixing diverse vulnerability types, leading to suboptimal performance due to insufficient specialization. To address this, we propose MoEFix, a novel framework leveraging a Mixture-of-Experts (MoE) architecture tailored for smart contract characteristics. MoEFix partitions vulnerabilities into subspaces, trains specialized experts for each type (e.g., reentrancy, integer overflow), and employs a vulnerability-aware router to dynamically allocate repairs. We further redesign the repair workflow to align with large language models, enabling end-to-end secure contract generation instead of partial patches, and to achieve this, we curated a dataset of 1,391 contracts covering five critical vulnerability types.To validate our approach, we extend the benchmark PVD test suite. Experiments demonstrate that MoEFix outperforms state-of-the-art methods by 21.64% in overall accuracy, achieving improvements of 26.19% (reentrancy) and 23.08% (delegatecall) for specific vulnerabilities.

The EEG recorded in fMRI scanner are totally submerged in gradient, ballistocardiogram and ocular artifacts. Traditional ways to reduce these artifacts have many limitations. AAS has a high requirement on recording and calculating and ICA denoising can be hardly done objectively without extra references. In this paper, we try to simplify and improve the AAS to lower its requirements on calculating. PCA dimension reduction is also introduced to overcome the problem of ICA denoising. A various artifacts removal method is developed based on improved AAS called Re-seg AAS and ICA&PCA. Comparisons before and after each processing step in both time and frequency domain shows that proposed method reduce gradient, ballistocardiogram and ocular artifacts effectively.

In order to reduce the inter-carrier interference (ICI) produced by frequency offset in OFDM systems, we set up an optimization problem to find the transmission pulse which maximizes the signal-to-average-ICI power ratio. Furthermore, by solving a constrained convex optimization problem, our pulses can satisfy various decay requirements, for instance with respect to a prescribed spectral mask. The latter is an important feature in order to comply with industry standard specifications. Furthermore we discuss issues concerning the numerical treatment of the constrained convex optimization problem. Simulation results show that our pulses outperform many currently known transmission pulses with respect to the ICI criterion.

We present a digital background technique for correcting the time and gain mismatches in a time-interleaved analog-to-digital converter (ADC) system. The proposed blind calibration is applicable to any number of time-interleaved ADCs and requires only modest oversampling. Simulation results show fast convergence and desirable detection accuracy. After the mismatch errors detection, the resulting signal to noise ratio (SNR) of the output signal is shown to be higher than the SNR of the input signal in a 16-ADC system.