With the proliferation of social media, assessing depression risk through user-generated posts has emerged as a critical and prominent research direction. However, existing methods fail to capture clinically relevant linguistic markers due to ungrounded semantic representations in non-expert models. Therefore, this paper proposes a model, ProDoKE, which leverages prompt chaining integrated with domain knowledge to facilitate in-depth extraction and interpretation of clinically significant linguistic and semantic cues from social media posts. ProDoKE comprises two key modules, consisting of a prompt chaining enhancing module and a capsule fusion module. Following the collection of risky posts, the prompt chaining module systematically integrates domain-specific knowledge with Large Language Models (LLMs). This structured reasoning process guides the LLM to identify relevant clinical symptoms and then infer their underlying emotional states, resulting in a richer and more accurate interpretation. Furthermore, the capsule fusion module incorporates contrastive learning to optimize the alignment of emotional and symptom semantics, while using capsule networks to fortify the model's robustness. Evaluations on the eRisk2017 and eRisk2018 datasets show that ProDoKE achieves state-of-the-art performance, significantly outperforming previous baseline models in F1 scores.

The diffusion-based text-to-image model harbors im-mense potential in transferring reference style. However, current encoder-based approaches significantly impair the text controllability of text-to-image models while transfer-ring styles. In this paper, we introduce DEADiff to address this issue using the following two strategies: 1) a mecha-nism to decouple the style and semantics of reference images. The decoupled feature representations are first extracted by Q-Formers which are instructed by different text descriptions. Then they are injected into mutually exclusive subsets of cross-attention layers for better disentanglement. 2) A non-reconstructive learning method. The Q-Formers are trained using paired images rather than the identical target, in which the reference image and the ground-truth image are with the same style or semantics. We show that DEADiff attains the best visual stylization results and optimal balance between the text controllability inherent in the text-to-image model and style similarity to the reference image, as demonstrated both quantitatively and qualitatively. Our project page is https://tianhao-qi.github.io/DEADiff‘/.

We propose Pure and Lightning ID customization (PuLID), a novel tuning-free ID customization method for text-to-image generation. By incorporating a Lightning T2I branch with a standard diffusion one, PuLID introduces both contrastive alignment loss and accurate ID loss, minimizing disruption to the original model and ensuring high ID fidelity. Experiments show that PuLID achieves superior performance in both ID fidelity and editability. Another attractive property of PuLID is that the image elements (\eg, background, lighting, composition, and style) before and after the ID insertion are kept as consistent as possible. Codes and models are available at https://github.com/ToTheBeginning/PuLID

Unsupervised domain adaptation medical image segmentation is aimed to segment unlabeled target domain images with labeled source domain images. However, different medical imaging modalities lead to large domain shift between their images, in which well-trained models from one imaging modality often fail to segment images from anothor imaging modality. In this paper, to mitigate domain shift between source domain and target domain, a style consistency unsupervised domain adaptation image segmentation method is proposed. First, a local phase-enhanced style fusion method is designed to mitigate domain shift and produce locally enhanced organs of interest. Second, a phase consistency discriminator is constructed to distinguish the phase consistency of domain-invariant features between source domain and target domain, so as to enhance the disentanglement of the domain-invariant and style encoders and removal of domain-specific features from the domain-invariant encoder. Third, a style consistency estimation method is proposed to obtain inconsistency maps from intermediate synthesized target domain images with different styles to measure the difficult regions, mitigate domain shift between synthesized target domain images and real target domain images, and improve the integrity of interested organs. Fourth, style consistency entropy is defined for target domain images to further improve the integrity of the interested organ by the concentration on the inconsistent regions. Comprehensive experiments have been performed with an in-house dataset and a publicly available dataset. The experimental results have demonstrated the superiority of our framework over state-of-the-art methods.

In this work, a ten-way power-combined power amplifier is designed using a load modulated balanced amplifier (LMBA)-based architecture. To provide the required magnitude and phase controls between the main and control-signal paths of the LMBA, an unequal power splitter and a phase compensation network are proposed. As proof of concept, the designed power amplifier is implemented in a 45-nm SOI CMOS process. At 40 GHz, it delivers a 25.1 dBm$\text{P}_{\text {sat}}$with a peak power-added efficiency (PAE) of 27.9%. At 6-dB power back-off level, it achieves 1.39 times drain efficiency enhancement over an ideal Class-B power amplifier. Using a 200-MHz single-carrier 64-QAM signal, the designed amplifier delivers an average output power of 16.5 dBm with a PAE of 13.1% at an EVMrmsof −23.9 dB and ACPR of −25.3 dBc. The die size, including all testing pads, is only 1.92 mm2. To the best of the authors’ knowledge, compared with the other recently published silicon-based LMBAs, this design achieves the highest$\text{P}_{\text {sat}}$.

In this work, a wideband millimeter-wave (mm-wave) power amplifier (PA) is reported. To provide excellent input/output impedance matching across broadband, sufficient output power and high power-added efficiency (PAE), a balanced amplifier (BA)-based architecture is used in designing this PA. In particular, an edge-coupled quadrature coupler is designed as its RF-power-division/combination block, and its performance in terms of magnitude/phase balance error is minimized throughout a relatively wide bandwidth. A PA prototype is fabricated in 0.13-$\mu \text{m}$SiGe HBT technology and tested. Under a 1.6-V power supply, the variation of small-signal gain is less than 3 dB within 20-40 GHz, which is equivalent to more than 66% fractional 3-dB bandwidth. Within this frequency range, at least 15.8 dBm saturated output power could be delivered with the peak PAE higher than 16.8%. The designed PA supports a single-carrier 200-MHz 64-quadrature amplitude modulation (QAM) with higher than 12.8-dBm average output power, while still maintaining an error vector magnitude (EVM) below −25 dB at 30 GHz. The size of the designed compact-size PA, including all pads, is only 0.7 mm$\times1.3$mm.

Understanding convergence of SGD-based optimization algorithms can help deal with enormous machine learning problems. To ensure last-iterate convergence of SGD and momentum-based SGD (mSGD), the existing studies usually constrain the step size $\epsilon_{n}$ to decay as $\sum_{n=1}^{+\infty}\epsilon_{n}^{2}<+\infty$, which however is rather conservative and may lead to slow convergence in the early stage of the iteration. In this paper, we relax this requirement by studying an alternate step size for the mSGD. First, we relax the requirement of the decay on step size to $\sum_{n=1}^{+\infty}\epsilon_{n}^{2+\eta_{0}}<+\infty\ (0\le\eta_{0}<1/2)$. This implies that a larger step size, such as $\epsilon_{n}=\frac{1}{\sqrt{n}}$ can be utilized for accelerating the mSGD in the early stage. Under this new step size and some common conditions, we prove that the gradient norm of mSGD for non-convex loss functions asymptotically decays to zero. In addition, we show that this step size can indeed help make the convergence into a neighborhood of the stationary points quicker in the early stage. In addition, we establish the convergence of mSGD under a constant step size $\epsilon_n\equiv\epsilon>0$ by removing the common requirement in the literature on the strong convexity of the loss function. Some experiments are given to illustrate the developed results.

The millimeter-wave (mm-wave) single-pole double-throw (SPDT) switch designed in bulk CMOS technology has limited power-handling capability in terms of 1-dB compression point (P1dB) inherently. This is mainly due to the low threshold voltage of the switching transistors used for shunt-connected configuration. To solve this issue, an innovative approach is presented in this work, which utilizes a unique passive ring structure. It allows a relatively strong RF signal passing through the TX branch, while the switching transistors are turned on. Thus, the fundamental limitation for P1dB due to reduced threshold voltage is overcome. To prove the presented approach is feasible in practice, a 90-GHz asymmetrical SPDT switch is designed in a standard 55-nm bulk CMOS technology. The design has achieved an insertion loss of 3.2 dB and 3.6 dB in TX and RX mode, respectively. Moreover, more than 20 dB isolation is obtained in both modes. Because of using the proposed passive ring structure, a remarkable P1dB is achieved. No gain compression is observed at all, while a 19.5 dBm input power is injected into the TX branch of the designed SPDT switch. The die area of this design is only 0.26 mm2.

Graph neural networks have been developed for various node classification tasks in graph data. However, the noise in graph data would affect the effectiveness of training and prediction of these graph learning models. In this paper, we propose a graph learning approach, named GCN_CL, which introduces a confident learning method into the graph convolutional networks model to classify nodes in graph data with label noise. The proposed approach includes node classification module and confident learning module, where the confident learning module selects clean nodes with high confident labels for node classification module to train an accurate model in the graphs with label noise. Pseudo label method is further applied on unlabeled data to increase samples for confident learning and improve the classification performance of subsequent node classification module. We evaluated classification performance of GCN_CL and compared our approach against other models to classify autism spectrum disorders patients versus health controls in population graph, which is constructed from ABIDE I dataset. The experimental results show GCN_CL achieves achieves the best performance in the population graph with different artificial noise levels.

Image/video stitching is a technology for solving the field of view (FOV) limitation of images/ videos. It stitches multiple overlapping images/videos to generate a wide-FOV image/video, and has been used in various fields such as sports broadcasting, video surveillance, street view, and entertainment. This survey reviews image/video stitching algorithms, with a particular focus on those developed in recent years. Image stitching first calculates the corresponding relationships between multiple overlapping images, deforms and aligns the matched images, and then blends the aligned images to generate a wide-FOV image. A seamless method is always adopted to eliminate such potential flaws as ghosting and blurring caused by parallax or objects moving across the overlapping regions. Video stitching is the further extension of image stitching. It usually stitches selected frames of original videos to generate a stitching template by performing image stitching algorithms, and the subsequent frames can then be stitched according to the template. Video stitching is more complicated with moving objects or violent camera movement, because these factors introduce jitter, shakiness, ghosting, and blurring. Foreground detection technique is usually combined into stitching to eliminate ghosting and blurring, while video stabilization algorithms are adopted to solve the jitter and shakiness. This paper further discusses panoramic stitching as a special-extension of image / video stitching. Panoramic stitching is currently the most widely used application in stitching. This survey reviews the latest image/video stitching methods, and introduces the fundamental principles/advantages/weaknesses of image/video stitching algorithms. Image/video stitching faces long-term challenges such as wide baseline, large parallax, and low-texture problem in the overlapping region. New technologies may present new opportunities to address these issues, such as deep learning-based semantic correspondence, and 3D image stitching. Finally, this survey discusses the challenges of image/video stitching and proposes potential solutions.

BACKGROUND: A typical microarray experiment has many sources of variation which can be attributed to biological and technical causes. Identifying sources of variation and assessing their magnitude, among other factors, are important for optimal experimental design. The objectives of this study were: (1) to estimate relative magnitudes of different sources of variation and (2) to evaluate agreement between biological and technical replicates. RESULTS: We performed a microarray experiment using a total of 24 Affymetrix GeneChip arrays. The study included 4th mammary gland samples from eight 21-day-old Sprague Dawley CD female rats exposed to genistein (soy isoflavone). RNA samples from each rat were split to assess variation arising at labeling and hybridization steps. A general linear model was used to estimate variance components. Pearson correlations were computed to evaluate agreement between technical and biological replicates. CONCLUSION: The greatest source of variation was biological variation, followed by residual error, and finally variation due to labeling when *.cel files were processed with dChip and RMA image processing algorithms. When MAS 5.0 or GCRMA-EB were used, the greatest source of variation was residual error, followed by biology and labeling. Correlations between technical replicates were consistently higher than between biological replicates.