arXiv subjects

Yuhuan Wang

Publications and source records attributed to Yuhuan Wang.

5 recordsLinked to original sources

arXiv2026

Interpretable machine learning for predicting splitting strength of asphalt concrete: insights from SHAP analysis

This paper presents an interpretable machine-learning framework for predicting the splitting strength (ST) of asphalt concrete and supporting data-driven mixture design. A database consisting of 296 samples was established, and 14 input variables related to asphalt properties, aggregate gradation, and fiber characteristics were selected for modeling. Six machine-learning models, namely TabPFN, ANN, SVR, RF, XGBoost, and LightGBM, were developed and compared. Hyperparameter optimization was performed for five models using NSGA-II, while TabPFN was directly applied with its default configuration. The results show that all six models achieved satisfactory predictive capability, whereas TabPFN delivered the best overall performance on the testing set, with the lowest RMSE of 0.28, MAE of 0.21, MAPE of 18.01%, MAD of 0.14, the highest R^2 of 0.88, and the highest composite score of 0.91. SHAP analysis further revealed that nine dominant variables accounted for 92.0% of the total average contribution, among which Ag9.5, FT, Ag4.75, AC, and Du were the most influential. In addition, favorable parameter ranges for improving ST were quantified, such as Ag9.5 < 66.8%, Ag4.75 < 45.0%, AC < 5.4 wt.%, AV < 3.6%, and Du > 134.7 cm. Finally, a GUI platform integrating prediction and SHAP-based explanation was developed to improve the accessibility and practical applicability of the proposed framework.

cs.LG↗

arXiv2025

Balanced Low-Complexity and Flexible Error-Correction List Flip Decoding for Polar Codes

Benefiting from performance advantages under short code lengths, polar codes are well-suited for certain scenarios, such as the future Internet of Things (IoT) applications that require high reliability and low power. Existing list flip decoders can efficiently further enhance the error-correction performance of polar codes with finite code lengths, particularly the dynamic successive cancellation list flip (D-SCLF) decoder with flexible high-order error-correction capability (FHECC). However, to the best of our knowledge, current list flip decoders cannot effectively balance complexity and error-correction efficiency. To address this, we propose a parity-check-aided D-SCLF (PC-DSCLF) decoder. This decoder, based on FHECC and the characteristics of the list flip decoding process, introduces a simplified flip metric and a hybrid check scheme, along with a decoding method that supports the check scheme, enabling it to retain FHECC while achieving low complexity. Simulation results show that the proposed PC-DSCLF decoder achieves up to a 51.1\% average complexity reduction compared to the D-SCLF algorithm with distributed CRC for $PC(512, 256+24)$

cs.IT↗

arXiv2024

Performance Evaluation of PAC Decoding with Deep Neural Networks

By concatenating a polar transform with a convolutional transform, polarization-adjusted convolutional (PAC) codes can reach the dispersion approximation bound in certain rate cases. However, the sequential decoding nature of traditional PAC decoding algorithms results in high decoding latency. Due to the parallel computing capability, deep neural network (DNN) decoders have emerged as a promising solution. In this paper, we propose three types of DNN decoders for PAC codes: multi-layer perceptron (MLP), convolutional neural network (CNN), and recurrent neural network (RNN). The performance of these DNN decoders is evaluated through extensive simulation. Numerical results show that the MLP decoder has the best error-correction performance under a similar model parameter number.

cs.IT↗

arXiv2023

SNR-Independent Joint Source-Channel Coding for wireless image transmission

Significant progress has been made in wireless Joint Source-Channel Coding (JSCC) using deep learning techniques. The latest DL-based image JSCC methods have demonstrated exceptional performance during transmission, while also avoiding cliff effects. However, current channel adaptive JSCC methods rely on channel SNR information, which can lead to performance degradation in practical applications due to channel mismatch effects. This paper proposes a novel approach for image transmission, called SNR Independent Joint Source-Channel Coding (SIJSCC), which utilizes Deep Learning techniques to achieve exceptional performance across various signal-to-noise ratio (SNR) levels without SNR estimating. We have designed an Inverted Residual Attention Bottleneck (IRAB) module for the model, which can effectively reduce the number of parameters while expanding the receptive field. In addition, we have incorporated a convolution and self-attention mixed encoding module to establish long-range dependency relationships between channel symbols. Our experiments have shown that SIJSCC outperforms existing channel adaptive DL-based JSCC methods that rely on SNR information. Furthermore, we found that SNR estimation does not significantly benefit SIJSCC, which provides insights for the future design of DL-based JSCC methods. The reliability of the proposed method is further demonstrated through an analysis of the model bottleneck and its adaptability to different domains, as shown by our experiments.

eess.SP↗

arXiv2021

Adaptive List Decoder with Flip Operations for Polar Codes

Successive cancellation list decoders with flip operations (SCL-Flip) can utilize re-decoding attempts to significantly improve the error-correction performance of polar codes. However, these re-decoding attempts result in extra computation complexity, which thus leads to increased energy consumption and decoding latency to the communication system adopting SCL-Flip decoders. To significantly reduce the computation complexity of current SCL-Flip decoders, we design a new adaptive SCL-Flip (AD-SCLF) decoder, which can be easily implemented based on existing SCL-Flip techniques. Simulation results showed that the AD-SCLF can reduce up to 80.85\% of the computational complexity of a current SCL-Flip decoder at a matched $FER=10^{-3}$. The result implies our decoder can significantly reduce the energy consumption caused by redundant re-decoding attempts from the SCL-Flip decoder.

cs.IT↗