CryptoL: Towards Scale Dominance and Physics Constraints Mitigation in Financial Multivariate Time Series Forecasting
本文提出CryptoL框架，通过上下文归一化、适应性数值稳定及软可行性损失等方法解决加密货币预测中的规模异质性与物理约束问题。
0 citationsRead paper
Selected work
Representative Papers
Hierarchical Fourier Approximation for Variational Quantum Distribution Learning
本文通过布尔立方体上的Walsh-Fourier层次逼近方法研究变分量子分布学习问题，利用量子电路Born机进行学习，并分析了这种方法的学习保证和误差。
0 citationsRead paper
Denoising Diffusion Generative Models Secretly Calculate Attentions
研究揭示了去噪扩散模型隐含使用类似Transformer的注意力机制，提出基于此简化图像生成算法，减少训练时间和计算资源。
0 citationsRead paper
XHotpotQA: A Benchmark for Cross-Lingual Knowledge Composition in Multi-Hop Question Answering
为解决多语言证据整合问题，本文引入XHotpotQA基准，通过构建跨语言证据依赖图来评估和改进多跳问答系统中的知识组合能力。
0 citationsRead paper
Bit-Level Triangular Content-Aware Permutation for Fragile Image Watermarking: Zero False Positive Rate, Single-Bit Sensitivity, and Arbitrary Dimension Support
This study addresses the vulnerability of existing fragile watermarking schemes to vector quantization (VQ) and collage attacks, as well as their inherent size constraints, by proposing a dimension-independent fragile watermarking algorithm. The method employs bit-level triangular content-aware permutation to replace conventional hash functions, combined with a vertical sandwich transform and residue merging strategy, enabling content-dependent watermark generation and secure tamper localization for arbitrary-sized images. Experimental results demonstrate zero false positives and false negatives across 17 attack types, achieving a PSNR of 99.33 dB. The proposed approach effectively resists VQ, collage, and geometric attacks while maintaining single-bit sensitivity and eliminating zero-padding limitations, rendering it highly suitable for digital forensics and medical image authentication applications.
0 citationsRead paper
Recent publications
Latest Papers
CryptoL: Towards Scale Dominance and Physics Constraints Mitigation in Financial Multivariate Time Series Forecasting
本文提出CryptoL框架，通过上下文归一化、适应性数值稳定及软可行性损失等方法解决加密货币预测中的规模异质性与物理约束问题。
0 citationsRead paper
Hierarchical Fourier Approximation for Variational Quantum Distribution Learning
本文通过布尔立方体上的Walsh-Fourier层次逼近方法研究变分量子分布学习问题，利用量子电路Born机进行学习，并分析了这种方法的学习保证和误差。
0 citationsRead paper
Denoising Diffusion Generative Models Secretly Calculate Attentions
研究揭示了去噪扩散模型隐含使用类似Transformer的注意力机制，提出基于此简化图像生成算法，减少训练时间和计算资源。
0 citationsRead paper
XHotpotQA: A Benchmark for Cross-Lingual Knowledge Composition in Multi-Hop Question Answering
为解决多语言证据整合问题，本文引入XHotpotQA基准，通过构建跨语言证据依赖图来评估和改进多跳问答系统中的知识组合能力。
0 citationsRead paper
Bit-Level Triangular Content-Aware Permutation for Fragile Image Watermarking: Zero False Positive Rate, Single-Bit Sensitivity, and Arbitrary Dimension Support
This study addresses the vulnerability of existing fragile watermarking schemes to vector quantization (VQ) and collage attacks, as well as their inherent size constraints, by proposing a dimension-independent fragile watermarking algorithm. The method employs bit-level triangular content-aware permutation to replace conventional hash functions, combined with a vertical sandwich transform and residue merging strategy, enabling content-dependent watermark generation and secure tamper localization for arbitrary-sized images. Experimental results demonstrate zero false positives and false negatives across 17 attack types, achieving a PSNR of 99.33 dB. The proposed approach effectively resists VQ, collage, and geometric attacks while maintaining single-bit sensitivity and eliminating zero-padding limitations, rendering it highly suitable for digital forensics and medical image authentication applications.
0 citationsRead paper