To overcome the performance limitations and computational complexity induced by the unimodular constraint in conventional hybrid precoding designs, this paper proposes a novel optoelectronic base station (OE-BS) architecture, replacing the traditional phase shifter network with an optical network. The proposed OE-BS architecture facilitates simultaneous amplitude and phase control in the analog domain, effectively eliminating the unimodular constraint and significantly enhancing the design flexibility. Based on the proposed OE-BS architecture, a closed-form hybrid precoding solution is first derived for narrowband systems, thus avoiding iterative optimization procedures. For wideband systems, a user-selective hybrid precoding algorithm is developed, where the digital precoder is designed according to the zero-forcing criterion, and the analog precoder and power allocation are jointly optimized using an alternating optimization method. Simulation results demonstrate that the proposed OE-BS schemes outperform conventional methods in both narrowband and wideband scenarios, enhancing overall system performance while reducing computational complexity.

The co-design of neural network architectures, quantization precisions, and hardware accelerators offers a promising approach to achieving an optimal balance between performance and efficiency, particularly for model deployment on resource-constrained edge devices. In this work, we propose the JAQ Framework, which jointly optimizes the three critical dimensions. However, effectively automating the design process across the vast search space of those three dimensions poses significant challenges, especially when pursuing extremely low-bit quantization. Specifical, the primary challenges include: (1) Memory overhead in software-side: Low-precision quantization-aware training can lead to significant memory usage due to storing large intermediate features and latent weights for backpropagation, potentially causing memory exhaustion. (2) Search time-consuming in hardware-side: The discrete nature of hardware parameters and the complex interplay between compiler optimizations and individual operators make the accelerator search time-consuming. To address these issues, JAQ mitigates the memory overhead through a channel-wise sparse quantization (CSQ) scheme, selectively applying quantization to the most sensitive components of the model during optimization. Additionally, JAQ designs BatchTile, which employs a hardware generation network to encode all possible tiling modes, thereby speeding up the search for the optimal compiler mapping strategy. Extensive experiments demonstrate the effectiveness of JAQ, achieving approximately 7% higher Top-1 accuracy on ImageNet compared to previous methods and reducing the hardware search time per iteration to 0.15 seconds.

Precise audio-visual synchronization in speech videos is crucial for content quality and viewer comprehension. Existing methods have made significant strides in addressing this challenge through rule-based approaches and end-to-end learning techniques. However, these methods often rely on limited audio-visual representations and suboptimal learning strategies, potentially constraining their effectiveness in more complex scenarios. To address these limitations, we present UniSync, a novel approach for evaluating audio-visual synchronization using embedding similarities. UniSync offers broad compatibility with various audio representations (e.g., Mel spectrograms, HuBERT) and visual representations (e.g., RGB images, face parsing maps, facial landmarks, 3DMM), effectively handling their significant dimensional differences. We enhance the contrastive learning framework with a margin-based loss component and cross-speaker unsynchronized pairs, improving discriminative capabilities. UniSync outperforms existing methods on standard datasets and demonstrates versatility across diverse audio-visual representations. Its integration into talking face generation frameworks enhances synchronization quality in both natural and AI-generated content.

We present STFTCodec, a novel spectral-based neural audio codec that efficiently compresses audio using Short-Time Fourier Transform (STFT). Unlike waveform-based approaches that require large model capacity and substantial memory consumption, this method leverages STFT for compact spectral representation and introduces unwrapped phase derivatives as auxiliary features. Our architecture employs parallel magnitude and phase processing branches enhanced by advanced feature extraction mechanisms. By relaxing strict phase reconstruction constraints while maintaining phase-aware processing, we achieve superior perceptual quality. Experimental results demonstrate that STFTCodec outperforms both waveform-based and spectral-based approaches across multiple bitrates, while offering unique flexibility in compression ratio adjustment through STFT parameter modification without architectural changes.

The minimum storage rack-aware regenerating (MSRR) code is a variation of regenerating codes that achieves the optimal repair bandwidth for a single node failure within the rack-aware model. We study the access complexity of repairing MSRR codes, that allows collective information processing among nodes within the same rack. A previous study has reported construction of MSRR codes that require accessing the minimum number of symbols to repair a single node. We extend this work by constructing a family of MSRR codes that minimizes the number of symbols accessed to repair sequentially ordered failed nodes in a single rack, and further show that for certain code parameters, another version of MSRR codes can be constructed with reduced sub-packetization while still preserving the optimal access property.

Terahertz (THz) waves possess great properties such as ultra-large bandwidth, good penetration, strong reflectivity to metallic materials, and low photon energy, making them widely applicable in fields such as optical communication, optical networks, and optical imaging. Therefore, THz optical imaging, as a novel non-destructive testing technology, demonstrates significant application value in areas such as body security scanning. However, existing THz imaging systems require that targets be close to the detector in order to avoid issues like visible stripes, increased noise, reduced contrast, and blurred edges, which can significantly reduce the quality of imaging results. To solve this problem, this study proposes an AM-Res2Net model to achieve accurate THz optical image recognition. The proposed model, compared with the Res2Net model, uses multiple small kernel convolutions to replace a single large kernel convolution. Furthermore, the proposed model also replaces the activation function and introduces the attention mechanism in the residual structure. The AM-Res2Net model tested on the THz optical image set constructed in this study achieved a recognition accuracy of 94.3%, which is higher than the ResNet and Res2Net models.

We study the access complexity of repairing MDS array codes in a rack-aware storage model, that allows collective information processing among nodes within the same rack. We construct a family of MDS array codes that only requires accessing the minimum number of symbols to repair consecutive failed nodes in a single rack.