Large Language Models (LLMs) and causal learning each hold strong potential for clinical decision making (CDM). However, their synergy remains poorly understood, largely due to the lack of systematic benchmarks evaluating their integration in clinical risk prediction. In real-world healthcare, identifying features with causal influence on outcomes is crucial for actionable and trustworthy predictions. While recent work highlights LLMs' emerging causal reasoning abilities, there lacks comprehensive benchmarks to assess their causal learning and performance informed by causal features in clinical risk prediction. To address this, we introduce REACT-LLM, a benchmark designed to evaluate whether combining LLMs with causal features can enhance clinical prognostic performance and potentially outperform traditional machine learning (ML) methods. Unlike existing LLM-clinical benchmarks that often focus on a limited set of outcomes, REACT-LLM evaluates 7 clinical outcomes across 2 real-world datasets, comparing 15 prominent LLMs, 6 traditional ML models, and 3 causal discovery (CD) algorithms. Our findings indicate that while LLMs perform reasonably in clinical prognostics, they have not yet outperformed traditional ML models. Integrating causal features derived from CD algorithms into LLMs offers limited performance gains, primarily due to the strict assumptions of many CD methods, which are often violated in complex clinical data. While the direct integration yields limited improvement, our benchmark reveals a more promising synergy: LLMs serve effectively as knowledge-rich collaborators for identifying and optimizing causal features. Additionally, in-context learning improves LLM predictions when prompts are tailored to the task and model. Different LLMs show varying sensitivity to structured data encoding formats, for example, open-source models perform better with JSON, while smaller models benefit from narrative serialization. These findings highlight the need to match prompts and data formats to model architecture and pretraining.

Aging progresses unevenly across cells, making chronological age an incomplete measure of biological decline. Aging clocks, machine learning models trained on molecular features, offer a promising approach by capturing predictable molecular patterns. However, identifying cell-type-specific aging signals from high-dimensional and heterogeneous single-cell data remains challenging. Temporal causal learning provides a potential solution by uncovering causal relationships among gene features, enabling more stable and reliable feature selection. Despite its potential, its application to aging clocks is largely limited by the scarcity of longitudinal genetic data. Reconstructing continuous aging trajectories from sparse, unevenly distributed cross-sectional data poses a barrier to introduce causal learning to age clocks. To address these challenges, this study proposes LCANet, a novel causal knowledge distillation framework for lung cell aging prediction. LCANet constructs pseudo-temporal lung cell trajectories, performs causal discovery across adjacent pseudo-time slices, and distills the identified causal knowledge into a student model. Experiments on a popular lung cell dataset demonstrate that LCANet effectively identifies reliable causal genes and integrates them into the distillation process, enabling the lightweight student model to predict cellular age with high efficiency and accuracy.

Group activity recognition in weakly supervised settings can be approached through either detector-based or detector-free methods. Detector-based methods typically rely on predefined graph structures to capture group relationships, which limits their ability to model the dynamic, multi-scale, and fine-grained changes in group interactions. In contrast, detector-free methods face challenges in effectively filtering irrelevant background and spectator information. In this paper, we propose a novel multi-scale hypergraph neural network framework, MHGAR, for weakly supervised group activity recognition. Unlike traditional methods with fixed graph structures, MHGAR dynamically constructs hyperedges to capture complex team formations and sub-group interactions at multiple scales. We initialize player representations by combining visual appearance and spatial positions, and iteratively refine these representations through dynamic hyperedge construction and bi-directional message passing. A cross-attentive activity classifier integrates the refined multi-scale features with ball position information to predict group activities. Our method effectively captures both fine-grained sub-group interactions and high-level team formations through its adaptive multi-scale modeling. Experimental results demonstrate state-of-the-art performance on the widely used Volleyball and NBA datasets.

Accurate prediction of unplanned readmissions allows healthcare systems to adopt preventive measures, reducing these occurrences. Creating a model that accurately predicts readmissions while simultaneously providing insights that are clinically interpretable is a complex and demanding task. This interpretability is vital, as it informs clinicians about the underlying reasons behind a model’s decisions, thus aiding in more informed clinical decision-making. In light of this, we introduce CARE-30 (Causal Analysis based Readmission Estimator within a 30-day time frame), a novel causal inference-based model, designed with clear cause-and-effect relationships, enhancing its potential acceptance and understanding among clinicians.In our approach, we generate a directed acyclic graph (DAG) to elucidate the latent causal relationships among pivotal clinical variables. Leveraging this causal graph, we can effectively apply the front-door criterion to construct the model and eliminate bias. By utilizing transformers, we integrate clinical texts, time-series data and categorical data into patient representation data as a robust input to the CARE-30 model. Our work signifies the effective application of knowledge from the field of causal inference in the domain of clinical medicine, and it also reflects the originality and impact of our research. CARE-30 unites causality and clinical prediction, unlocking new vistas in healthcare forecasting. Experiments demonstrated that our model outperforms all other methods in Accuracy and F1 scores.

Images taken under low-light conditions tend to suffer from poor visibility, which can decrease image quality and even reduce the performance of the downstream tasks. It is hard for a CNN-based method to learn generalized features that can recover normal images from the ones under various unknow low-light conditions. In this paper, we propose to incorporate the contrastive learning into an illumination correction network to learn abstract representations to distinguish various low-light conditions in the representation space, with the purpose of enhancing the generalizability of the network. Considering that light conditions can change the frequency components of the images, the representations are learned and compared in both spatial and frequency domains to make full advantage of the contrastive learning. The proposed method is evaluated on LOL and LOL-V2 datasets, the results show that the proposed method achieves better qualitative and quantitative results compared with other state-of-the-arts.

Abstract Single‐Image Super‐Resolution (SISR) has always been an important topic in the field of image processing, which attempts to improve the image resolution and is of great significance in practice. Recently, SISR has made substantial progress aided by deep learning (DL), which has demonstrated impressive potential in many low‐level tasks. In the current DL‐based SISR approaches, most of them are based on supervised learning. However, in the real world, only low‐resolution (LR) images with unknown degradation are provided, which limit the application of current supervised models. To mitigate this problem, in this paper, a two‐stage semi‐supervised SISR method called SRAttentionGAN, is proposed. First, an upsampling network SRResNet, which is pre‐trained in a supervised manner, is employed to scale the LR image to the desired size. Then, the upsampled results are fed into our improved unsupervised CycleGAN framework, which does not need paired samples, to obtain sharper and more realistic super‐resolution (SR) images. Specifically, in the improved CycleGAN part, an attention‐guided generator is proposed to perceive the discriminative semantic parts between the source and target images, to avoid the impact from low‐level information. It also prevents the overall color tone from being changed. A multi‐scale discriminator is also adopted to further rich texture details. The effectiveness of the proposed SRAttentionGAN experiments is validated using four benchmarks (Set5, Set14, Urban100, and BSDS100) in both quantitative and qualitative aspects. Compared with the state‐of‐the‐arts, the results are visually promising and show competitive performance in perceptual metrics, Natural Image Quality Evaluator (NIQE) and Perception Index (PI), which have better agreement with the human visual perception.

Unmanned aerial vehicle (UAV)-based tracking is attracting increasing attention and developing rapidly in applications such as agriculture, aviation, navigation, transportation and public security. Recently, discriminative correlation filters (DCF)-based trackers have stood out in UAV tracking community for their high efficiency and appealing robustness on a single CPU. However, due to limited onboard computation resources and other challenges the efficiency and accuracy of existing DCF-based approaches is still not satisfying. In this paper, inspired by residue representation, we exploit the residue nature inherent to videos and propose residue-aware correlation filters that show better convergence properties in filter learning. Moreover, we explore using segmentation by the GrabCut to improve the wildly adopted discriminative scale estimation in DCF-based trackers, which, as a mater of fact, greatly impacts the precision and accuracy of the trackers since accumulated scale error degrades the appearance model as online updating goes on. Extensive experiments are conducted on four UAV benchmarks, namely, UAV123@10fps, DTB70, UAVDT and Vistrone2018 (VisDrone2018-test-dev). The results show that our method achieves state-of-the-art performance.

Discriminative correlation filters (DCF) are efficient in visual tracking and have advanced the field significantly. However, the symmetry of correlation (or convolution) operator results in computational problems and does harm to the generalized translation equivariance. The former problem has been approached in many ways, whereas the latter one has not been well recognized. In this paper, we analyze the problems with the symmetry of circular convolution and propose an asymmetric one, which as a generalization of the former has a weak generalized translation equivariance property. With this operator, we propose a tracker called the asymmetric discriminative correlation filter (ADCF), which is more sensitive to translations of targets. Its asymmetry allows the filter and the samples to have different sizes. This flexibility makes the computational complexity of ADCF more controllable in the sense that the number of filter parameters will not grow with the sample size. Moreover, the normal matrix of ADCF is a block matrix with each block being a two-level block Toeplitz matrix. With this well-structured normal matrix, we design an algorithm for multiplying an N × N two-level block Toeplitz matrix by a vector with time complexity O ( N log N ) and space complexity O ( N ), instead of O ( N 2 ). Unlike DCF-based trackers, introducing spatial or temporal regularization does not increase the essential computational complexity of ADCF. Comparative experiments are performed on a synthetic dataset and four benchmarks, including OTB-2013, OTB-2015, VOT-2016, and Temple-Color, and the results show that our method achieves state-of-the-art visual tracking performance.