Zihan Wei

Ziqian Bi, Lu Chen, Junhao Song et al.

This study presents the first comprehensive evaluation of thinking budget mechanisms in medical reasoning tasks, revealing fundamental scaling laws between computational resources and reasoning quality. We systematically evaluated two major model families, Qwen3 (1.7B to 235B parameters) and DeepSeek-R1 (1.5B to 70B parameters), across 15 medical datasets spanning diverse specialties and difficulty levels. Through controlled experiments with thinking budgets ranging from zero to unlimited tokens, we establish logarithmic scaling relationships where accuracy improvements follow a predictable pattern with both thinking budget and model size. Our findings identify three distinct efficiency regimes: high-efficiency (0 to 256 tokens) suitable for real-time applications, balanced (256 to 512 tokens) offering optimal cost-performance tradeoffs for routine clinical support, and high-accuracy (above 512 tokens) justified only for critical diagnostic tasks. Notably, smaller models demonstrate disproportionately larger benefits from extended thinking, with 15 to 20% improvements compared to 5 to 10% for larger models, suggesting a complementary relationship where thinking budget provides greater relative benefits for capacity-constrained models. Domain-specific patterns emerge clearly, with neurology and gastroenterology requiring significantly deeper reasoning processes than cardiovascular or respiratory medicine. The consistency between Qwen3 native thinking budget API and our proposed truncation method for DeepSeek-R1 validates the generalizability of thinking budget concepts across architectures. These results establish thinking budget control as a critical mechanism for optimizing medical AI systems, enabling dynamic resource allocation aligned with clinical needs while maintaining the transparency essential for healthcare deployment.

Zihan Wei, Tapabrata Chakraborti · oxford

Recent works have shown that deep learning based skin lesion image classification models trained on unbalanced dataset can exhibit bias toward protected demographic attributes such as race, age,and gender. Current bias mitigation methods usually either achieve high level of fairness with the degradation of accuracy, or only improve the model fairness on a single attribute. Additionally usually most bias mitigation strategies are either pre hoc through data processing or post hoc through fairness evaluation, instead of being integrated into the model learning itself. To solve these existing drawbacks, we propose a new Individual Typology Angle (ITA) Loss-based skin tone normalization and data augmentation method that directly feeds into an adaptable meta learning-based joint channel pruning framework. In skin tone normalization, ITA is used to estimate skin tone type and adjust automatically to target tones for dataset balancing. In the joint channel pruning framework, two nested optimization loops are used to find critical channels.The inner optimization loop finds and prunes the local critical channels by weighted soft nearest neighbor loss, and the outer optimization loop updates the weight of each attribute using group wise variance loss on meta-set. Experiments conducted in the ISIC2019 dataset validate the effectiveness of our method in simultaneously improving the fairness of the model on multiple sensitive attributes without significant degradation of accuracy. Finally, although the pruning mechanism adds some computational cost during training phase, usually training is done off line. More importantly,

Han Wang, Ruoyun He, Guoguang Lao et al.

Early identification of high-risk ICU patients is crucial for directing limited medical resources. We introduce ALFIA (Adaptive Layer Fusion with Intelligent Attention), a modular, attention-based architecture that jointly trains LoRA (Low-Rank Adaptation) adapters and an adaptive layer-weighting mechanism to fuse multi-layer semantic features from a BERT backbone. Trained on our rigorous cw-24 (CriticalWindow-24) benchmark, ALFIA surpasses state-of-the-art tabular classifiers in AUPRC while preserving a balanced precision-recall profile. The embeddings produced by ALFIA's fusion module, capturing both fine-grained clinical cues and high-level concepts, enable seamless pairing with GBDTs (CatBoost/LightGBM) as ALFIA-boost, and deep neuro networks as ALFIA-nn, yielding additional performance gains. Our experiments confirm ALFIA's superior early-warning performance, by operating directly on routine clinical text, it furnishes clinicians with a convenient yet robust tool for risk stratification and timely intervention in critical-care settings.

Baozhu Huang, Cheng Chen, Xuanhe Hou et al.

The intensive care unit (ICU) manages critically ill patients, many of whom face a high risk of mortality. Early and accurate prediction of in-hospital mortality within the first 24 hours of ICU admission is crucial for timely clinical interventions, resource optimization, and improved patient outcomes. Traditional scoring systems, while useful, often have limitations in predictive accuracy and adaptability. Objective: This review aims to systematically evaluate and benchmark innovative methodologies that leverage data available within the first day of ICU admission for predicting in-hospital mortality. We focus on advancements in machine learning, novel biomarker applications, and the integration of diverse data types.

Lei Cao, Huijun Zhang, Ling Feng et al.

Despite detection of suicidal ideation on social media has made great progress in recent years, people's implicitly and anti-real contrarily expressed posts still remain as an obstacle, constraining the detectors to acquire higher satisfactory performance. Enlightened by the hidden "tree holes" phenomenon on microblog, where people at suicide risk tend to disclose their inner real feelings and thoughts to the microblog space whose authors have committed suicide, we explore the use of tree holes to enhance microblog-based suicide risk detection from the following two perspectives. (1) We build suicide-oriented word embeddings based on tree hole contents to strength the sensibility of suicide-related lexicons and context based on tree hole contents. (2) A two-layered attention mechanism is deployed to grasp intermittently changing points from individual's open blog streams, revealing one's inner emotional world more or less. Our experimental results show that with suicide-oriented word embeddings and attention, microblog-based suicide risk detection can achieve over 91\% accuracy. A large-scale well-labelled suicide data set is also reported in the paper.