Multimodal sentiment analysis (MSA) seeks to decode human emotions by integrating heterogeneous modalities. However, real-world scenarios often involve missing or misaligned data due to sensor failures or transmission errors, leading to disrupted temporal dynamics and degraded cross-modal correlations. To address these challenges, we propose RECAP (REcovery of Coherent Affective Patterns), a robust two-stage framework to restore temporal and structural emotional integrity under modality incompleteness. The first stage employs a causality-aware adversarial generator for multi-granularity temporal reconstruction, complemented by a contrastive mutual information factorization module that disentangles shared and modality-specific semantics. The second stage introduces a mutual information-guided attention fusion mechanism with a ranking-based objective, enabling adaptive integration of complementary signals for refined prediction. Extensive experiments on MOSI, MOSEI, and SIMS under various missing-modality conditions demonstrate that RECAP consistently outperforms state-of-the-art methods. Notably, it improves ACC-7 on MOSI by 2.71 percentage points and F1 on SIMS by 6.38 percentage points. These results verify the performance of RECAP in terms of capturing fine-grained emotional cues and robustness.

Text-guided image editors can now manipulate authentic medical scans with high fidelity, enabling lesion implantation/removal that threatens clinical trust and safety. Existing defenses are inadequate for healthcare. Medical detectors are largely black-box, while MLLM-based explainers are typically post-hoc, lack medical expertise, and may hallucinate evidence on ambiguous cases. We present MedForge, a data-and-method solution for pre-hoc, evidence-grounded medical forgery detection. We introduce MedForge-90K, a large-scale benchmark of realistic lesion edits across 19 pathologies with expert-guided reasoning supervision via doctor inspection guidelines and gold edit locations. Building on it, MedForge-Reasoner performs localize-then-analyze reasoning, predicting suspicious regions before producing a verdict, and is further aligned with Forgery-aware GSPO to strengthen grounding and reduce hallucinations. Experiments demonstrate state-of-the-art detection accuracy and trustworthy, expert-aligned explanations.

Timely and personalized treatment decisions are essential across a wide range of healthcare settings where patient responses can vary significantly and evolve over time. Clinical data used to support these treatment decisions are often irregularly sampled, where missing data frequencies may implicitly convey information about the patient's condition. Existing Reinforcement Learning (RL) based clinical decision support systems often ignore the missing patterns and distort them with coarse discretization and simple imputation. They are also predominantly model-free and largely depend on retrospective data, which could lead to insufficient exploration and bias by historical behaviors. To address these limitations, we propose medDreamer, a novel model-based reinforcement learning framework for personalized treatment recommendation. medDreamer contains a world model with an Adaptive Feature Integration module that simulates latent patient states from irregular data and a two-phase policy trained on a hybrid of real and imagined trajectories. This enables learning optimal policies that go beyond the sub-optimality of historical clinical decisions, while remaining close to real clinical data. We evaluate medDreamer on both sepsis and mechanical ventilation treatment tasks using two large-scale Electronic Health Records (EHRs) datasets. Comprehensive evaluations show that medDreamer significantly outperforms model-free and model-based baselines in both clinical outcomes and off-policy metrics.

OBJECTIVES: This study introduces Smart Imitator (SI), a 2-phase reinforcement learning (RL) solution enhancing personalized treatment policies in healthcare, addressing challenges from imperfect clinician data and complex environments. MATERIALS AND METHODS: Smart Imitator's first phase uses adversarial cooperative imitation learning with a novel sample selection schema to categorize clinician policies from optimal to nonoptimal. The second phase creates a parameterized reward function to guide the learning of superior treatment policies through RL. Smart Imitator's effectiveness was validated on 2 datasets: a sepsis dataset with 19 711 patient trajectories and a diabetes dataset with 7234 trajectories. RESULTS: Extensive quantitative and qualitative experiments showed that SI significantly outperformed state-of-the-art baselines in both datasets. For sepsis, SI reduced estimated mortality rates by 19.6% compared to the best baseline. For diabetes, SI reduced HbA1c-High rates by 12.2%. The learned policies aligned closely with successful clinical decisions and deviated strategically when necessary. These deviations aligned with recent clinical findings, suggesting improved outcomes. DISCUSSION: Smart Imitator advances RL applications by addressing challenges such as imperfect data and environmental complexities, demonstrating effectiveness within the tested conditions of sepsis and diabetes. Further validation across diverse conditions and exploration of additional RL algorithms are needed to enhance precision and generalizability. CONCLUSION: This study shows potential in advancing personalized healthcare learning from clinician behaviors to improve treatment outcomes. Its methodology offers a robust approach for adaptive, personalized strategies in various complex and uncertain environments.

Early disease diagnosis is critical for timely clinical intervention and treatment. Intelligent models have shown significant potential in addressing the challenges of misdiagnosis, especially given the shortage of experienced experts. However, there exists complexity of ultrasound image information, and subtle differences between positive and negative samples, combined with limited disease data, pose challenges in feature extraction, class imbalance, and the lack of representative positive class prototypes. To this end, we propose a large vision model distillation framework with fuzzy perception to improve rare disease diagnosis. Specifically, we first fine-tune the pre-trained Medical Segment Anything Model (MedSAM) to adapt it to the target domain. Through image augmentation and latent feature relationship distillation, we enhance feature extraction robustness and reduce inter-class ambiguity, which helps mitigate the impact of class imbalance on the lightweight student model. Second, as imaging style differences in ultrasound images are often more pronounced than subtle variations between positive and negative samples, we construct image style subsets and introduce a fuzzy style matching strategy to perceive these differences. Finally, we combine features from both the differential perception and semantic enhancement branches to strengthen disease classification. Extensive experiments on an internal fetal spina bifida dataset and three widely used imbalanced benign-malignant medical datasets demonstrate the effectiveness of the proposed method.

Retrieval-Augmented Generation (RAG) offers a promising solution to the limitations of static knowledge and hallucinations in Large Language Models (LLMs). While prior research has introduced numerous enhancements to RAG systems, a significant challenge remains under-explored: the potential conflict between external retrievals and LLMs' internal priors, which can undermine the quality of generated outputs. To tackle this issue, we present theEpitome-AugmentedGeneration (EAG) framework, which strategically aligns queries, external retrievals, and internal priors to produce high-quality LLM generations by selecting fuzzy inputs. EAG employs two novel lightweight modules, Criticism and Distillation, allowing traditional RAGs to be upgraded to EAGs without the need for specialized training data. Extensive experiments on five datasets across general and medical domains, including both open-ended and closed-ended tasks, validate the effectiveness of EAG. Our framework achieves substantial F1 score improvements: 7.03%, 23.35%, and 21.58% over baseline RAGs in medical QA tasks, 11.80% in law domain, 7.95% in finance domain, and 4.13% and 5.16% in general domain. Beyond performance gains, our study delves into the interplay between LLMs' internal priors and external retrievals, uncovering key principles that govern generation quality and providing valuable insights for future retrieval-augmented frameworks.

Multimodal survival analysis aims to combine heterogeneous data sources to improve the prediction quality of survival outcomes. However, this task is particularly challenging due to high heterogeneity and noise across data sources. Additionally, the exact survival time is often censored (partially known) due to incomplete event observation. To address the above challenges, we propose a novel interpretable evidential multimodal survival fusion model, EsurvFusion. This model is designed to combine multimodal data at the decision level using Epistemic Random Fuzzy Sets that jointly handle both data and model uncertainty while incorporating modality-level reliability. Specifically, EsurvFusion first models unimodal data with newly introduced Gaussian random fuzzy numbers, producing possible unimodal survival predictions along with corresponding aleatory and epistemic uncertainty. It then estimates modality-level reliability through a reliability discounting layer to correct the misleading impact of noisy data modalities. Finally, a multimodal evidence fusion layer is introduced to combine the discounted predictions, revealing modality-level influence based on the learned reliability coefficients. Extensive experiments on four multimodal cancer survival datasets demonstrate the effectiveness of our model in handling highly heterogeneous data, establishing a new state-of-the-art performance on several benchmarks.

Accurate identification of neurological disorders such as Alzheimer's disease (AD), Parkinson's disease (PD), and Autism Spectrum Disorder (ASD) is challenging due to subtle early-stage symptoms and heterogeneous brain dynamics. Resting-state functional MRI (rs-fMRI) enables the construction of functional brain networks, where Graph Neural Networks (GNNs) have shown promise for disease classification. However, existing GNN-based methods face three key limitations: correlation-based graph construction introduces noise and negative edges; domain knowledge about brain regions is ignored; and demographic or clinical metadata are fused through simplistic encodings. To overcome these limitations, we propose BrainPrompt+, a knowledge-guided framework that integrates Large Language Models (LLMs) with multi-level natural language prompts. Five types of prompts are introduced: spectral (frequency-domain BOLD features), spatial (inter-ROI connectivity), ROI (anatomical and functional knowledge), disease (progression stages), and subject (demographic context). These prompts are encoded by a frozen LLM and incorporated into a GNN pipeline, unifying imaging, clinical, and external knowledge in a semantically enriched and interpretable manner. Experiments on three rs-fMRI datasets show that BrainPrompt+ consistently outperforms state-of-the-art baselines, achieving accuracy gains of up to 8.93%. Biomarker analysis further demonstrates that the highlighted ROIs align with established neuroscience findings, confirming the interpretability of the model. BrainPrompt+ thus establishes a flexible and generalizable paradigm for knowledge-guided brain network analysis. The source code is available at https://github.com/AngusMonroe/BrainPromptPlus.

Kai He, Yucheng Huang, Wenqing Wang, Delong Ran, Dongming Sheng, Junxuan Huang, Qika Lin, Jiaxing Xu, Wenqiang Liu, Mengling Feng. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

Qika Lin, Tianzhe Zhao, Kai He, Zhen Peng, Fangzhi Xu, Ling Huang, Jingying Ma, Mengling Feng. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

Detecting LLM-generated text in specialized and high-stakes domains like medicine and law is crucial for combating misinformation and ensuring authenticity.However, current zeroshot detectors, while effective on general text, often fail when applied to specialized content due to domain shift.We provide a theoretical analysis showing this failure is fundamentally linked to the KL divergence between human, detector, and source text distributions.To address this, we propose DivScore, a zero-shot detection framework using normalized entropybased scoring and domain knowledge distillation to robustly identify LLM-generated text in specialized domains.We also release a domainspecific benchmark for LLM-generated text detection in the medical and legal domains.Experiments on our benchmark show that Di-vScore consistently outperforms state-of-theart detectors, with 14.4% higher AUROC and 64.0% higher recall (0.1% false positive rate threshold).In adversarial settings, DivScore demonstrates superior robustness to other baselines, achieving on average 22.8% advantage in AUROC and 29.5% in recall.Code and data are publicly available 1 .

Sleep staging is critical to assess sleep quality and diagnose disorders. Despite advancements in artificial intelligence enabling automated sleep staging, significant challenges remain: (1) Simultaneously extracting prominent temporal and spatial sleep features from multi-channel raw signals, including characteristic sleep waveforms and salient spatial brain networks. (2) Capturing the spatial-temporal coupling patterns essential for accurate sleep staging. To address these challenges, we propose a novel framework named ST-USleepNet, comprising a spatial-temporal graph construction module (ST) and a U-shaped sleep network (USleepNet). The ST module converts raw signals into a spatial-temporal graph based on signal similarity, temporal, and spatial relationships to model spatial-temporal coupling patterns. The USleepNet employs a U-shaped structure for both the temporal and spatial streams, mirroring its original use in image segmentation to isolate significant targets. Applied to raw sleep signals and graph data from the ST module, USleepNet effectively segments these inputs, simultaneously extracting prominent temporal and spatial sleep features. Testing on three datasets demonstrates that ST-USleepNet outperforms existing baselines, and model visualizations confirm its efficacy in extracting prominent sleep features and temporal-spatial coupling patterns across various sleep stages. The code is available at https://github.com/Majy-Yuji/ST-USleepNet.

While recent multimodal large language models (MLLMs) have advanced automated ECG interpretation, they still face two key limitations: (1) insufficient multimodal synergy between ECG time series and ECG images, and (2) limited explainability in linking diagnoses to granular waveform evidence. We introduce GEM, the first MLLM unifying ECG time series, 12-lead ECG images and text for grounded and clinician-aligned ECG interpretation. GEM enables feature-grounded analysis, evidence-driven reasoning, and a clinician-like diagnostic process through three core innovations: a dual-encoder framework extracting complementary time series and image features, cross-modal alignment for effective multimodal understanding, and knowledge-guided instruction data generation for generating high-granularity grounding data (ECG-Grounding) linking diagnoses to measurable parameters ($e.g.$, QRS/PR Intervals). Additionally, we propose the Grounded ECG Understanding task, a clinically motivated benchmark designed to comprehensively assess the MLLM's capability in grounded ECG understanding. Experimental results on both existing and our proposed benchmarks show GEM significantly improves predictive performance (CSN $7.4\%$ $\uparrow$), explainability ($22.7\%$ $\uparrow$), and grounding ($25.3\%$ $\uparrow$), making it a promising approach for real-world clinical applications. Codes, model, and data are available at https://github.com/lanxiang1017/GEM.

Time-to-event analysis provides insights into clinical prognosis and treatment recommendations. However, this task is more challenging than standard regression problems due to the presence of censored observations. Additionally, the lack of confidence assessment, model robustness, and prediction calibration raises concerns about the reliability of predictions. To address these challenges, we propose an evidential regression model specifically designed for time-to-event prediction. Our approach computes a degree of belief for the event time occurring within a time interval, without any strict distribution assumption. Meanwhile, the proposed model quantifies both epistemic and aleatory uncertainties using Gaussian Random Fuzzy Numbers and belief functions, providing clinicians with uncertainty-aware survival time predictions. Experimental evaluations using simulated and real-world survival datasets highlight the potential of our approach for enhancing clinical decision-making in survival analysis. • Calculate survival time without strict distribution assumptions. • Quantifies epistemic and aleatory uncertainties via GRFN and belief functions. • Provides uncertainty-aware survival predictions for clinicians. • Validated on simulated & real-world survival datasets.

Multimodal medical applications have garnered considerable attention due to their potential to offer comprehensive and robust support for medical assistance. Specifically, within this domain, difference-aware medical Visual Question Answering (VQA) has emerged as a topic of increasing interest that enables the recognition of changes in physical conditions over time when compared to previous states and provides customized suggestions accordingly. However, it is challenging because samples usually exhibit characteristics of complexity, diversity, and inherent noise. Besides, there is a need for multimodal knowledge understanding of the medical domain. The difference-aware setting requiring image comparison further intensifies these situations. To this end, we propose a cross-Modal knowlEdge diffusioN-baseD gEneration netwoRk (MENDER), where the diffusion mechanism with multi-step denoising and knowledge injection from global to local level are employed to tackle the aforementioned challenges, respectively. The diffusion process is to gradually generate answers with the sequence input of questions, random noises for the answer masks and virtual vision prompts of images. The strategy of answer nosing and knowledge cascading is specifically tailored for this task and is implemented during forward and reverse diffusion processes. Moreover, the visual and structure knowledge injection are proposed to learn virtual vision prompts to guide the diffusion process, where the former is realized using a pre-trained medical image-text network and the latter is modeled with spatial and semantic graph structures processed by the heterogeneous graph Transformer models. Experiment results demonstrate the effectiveness of MENDER for difference-aware medical VQA. Furthermore, it also exhibits notable performance in the low-resource setting and conventional medical VQA tasks.

Studying frailty is crucial for enhancing the health and quality of life among older adults, refining healthcare delivery methods, and tackling the obstacles linked to an aging demographic. Approaches to frailty modeling often utilise simple analytic techniques rather than available advanced machine learning methods, which may be sub-optimal. There is no large-scale systematic review on applications of machine learning methods on frailty modeling. In this study we explore the use of machine learning methods to predict or classify frailty in older persons in routinely collected data. We reviewed 181 research articles, and categorised analytic methods into three categories: generalised linear models, survival models, and non-linear models. These methods have a moderate agreement with existing frailty scores and predictive validity for adverse outcomes. Limited evidence suggests that non-linear methods outperform generalised linear methods. The top-three predictor/input variables are specific diagnosis or groups of diagnoses, functional performance (e.g., ADLs), and impaired cognition. Mortality, hospital admissions and prolonged hospital stay are the mainly predicted outcomes. Most studies utilise classical machine learning methods with cross-sectional data. Longitudinal data collected by wearable sensors have been used for frailty modeling. We also discuss the opportunities to use more advanced machine learning methods with high dimensional longitudinal data for more personalised and accessible frailty tools.

*Not all positive pairs are beneficial to time series contrastive learning*. In this paper, we study two types of bad positive pairs that can impair the quality of time series representation learned through contrastive learning: the noisy positive pair and the faulty positive pair. We observe that, with the presence of noisy positive pairs, the model tends to simply learn the pattern of noise (Noisy Alignment). Meanwhile, when faulty positive pairs arise, the model wastes considerable amount of effort aligning non-representative patterns (Faulty Alignment). To address this problem, we propose a Dynamic Bad Pair Mining (DBPM) algorithm, which reliably identifies and suppresses bad positive pairs in time series contrastive learning. Specifically, DBPM utilizes a memory module to dynamically track the training behavior of each positive pair along training process. This allows us to identify potential bad positive pairs at each epoch based on their historical training behaviors. The identified bad pairs are subsequently down-weighted through a transformation module, thereby mitigating their negative impact on the representation learning process. DBPM is a simple algorithm designed as a lightweight **plug-in** without learnable parameters to enhance the performance of existing state-of-the-art methods. Through extensive experiments conducted on four large-scale, real-world time series datasets, we demonstrate DBPM's efficacy in mitigating the adverse effects of bad positive pairs.

KDD AIDSH 2024 aims to foster discussions and developments that push the boundaries of Artificial Intelligence (AI) and Data Science (DS) in healthcare, enhance diagnostic accuracy and promote human-centric approaches to healthcare, thus stimulating future interdisciplinary collaborations. This year's symposium will focus on expanding the application of AI/DS in healthcare/medicine and bridging existing gaps. The workshop invites submissions of full papers as well as work-in-progress on the application of AI/DS in healthcare. The workshop will feature three invited talks from eminent speakers, spanning academia, industry, and clinical researchers. In addition, selected papers will be invited to publish in Health Data Science, a Science Partner Journal. This summary provides a brief description of the half-day workshop to be held on August 26th, 2024. The webpage for the workshop can be found at https://aimel.ai/kdd2024aidsh.

The comprehensive integration of machine learning healthcare models within clinical practice remains suboptimal, notwithstanding the proliferation of high-performing solutions reported in the literature. A predominant factor hindering widespread adoption pertains to an insufficiency of evidence affirming the reliability of the aforementioned models. Recently, uncertainty quantification methods have been proposed as a potential solution to quantify the reliability of machine learning models and thus increase the interpretability and acceptability of the results. In this review, we offer a comprehensive overview of the prevailing methods proposed to quantify the uncertainty inherent in machine learning models developed for various medical image tasks. Contrary to earlier reviews that exclusively focused on probabilistic methods, this review also explores non-probabilistic approaches, thereby furnishing a more holistic survey of research pertaining to uncertainty quantification for machine learning models. Analysis of medical images with the summary and discussion on medical applications and the corresponding uncertainty evaluation protocols are presented, which focus on the specific challenges of uncertainty in medical image analysis. We also highlight some potential future research work at the end. Generally, this review aims to allow researchers from both clinical and technical backgrounds to gain a quick and yet in-depth understanding of the research in uncertainty quantification for medical image analysis machine learning models.

Self-supervised pre-trained speech models were shown effective for various downstream speech processing tasks. Since they are mainly pre-trained to map input speech to pseudo-labels, the resulting representations are only effective for the type of pre-train data used, either clean or mixture speech. With the idea of selective auditory attention, we propose a novel pre-training solution called Selective-HuBERT, or SHuBERT, which learns the selective extraction of target speech representations from either clean or mixture speech. Specifically, SHuBERT is trained to predict pseudo labels of a target speaker, conditioned on an enrolment speech from the target speaker. By doing so, SHuBERT is expected to selectively attend to the target speaker in a complex acoustic environment, thus benefiting various downstream tasks. We further introduce a dual-path training strategy and use the cross-correlation constraint between the two branches to encourage the model to generate noise-invariant representation. Experiments on SUPERB benchmark and LibriMix dataset demonstrate the universality and noise-robustness of SHuBERT. Furthermore, we find that our high-quality representation can be easily integrated with conventional supervised learning methods to achieve significant performance, even under extremely low-resource labeled data.

The encoder-decoder model is a commonly used Deep Neural Network (DNN) model for medical image segmentation. Conventional encoder-decoder models make pixel-wise predictions focusing heavily on local patterns around the pixel. This makes it challenging to give segmentation that preserves the object's shape and topology, which often requires an understanding of the global context. In this work, we propose a Fourier Coefficient Segmentation Network (FCSN)—a novel global context-aware DNN model that segments an object by learning the complex Fourier coefficients of the object's masks. The Fourier coefficients are calculated by integrating over the whole contour. Therefore, for our model to make a precise estimation of the coefficients, the model is motivated to incorporate the global context of the object, leading to a more accurate segmentation of the object's shape. This global context awareness also makes our model robust to unseen local perturbations during inference, such as additive noise or motion blur that are prevalent in medical images. We compare FCSN with other state-of-the-art global context-aware models (UNet++, DeepLabV3+, UNETR) on 5 medical image segmentation tasks, of which 3 are camera imaging datasets (ISIC_2018, RIM_CUP, RIM_DISC) and 2 are medical imaging datasets (PROSTATE, FETAL). When FCSN is compared with UNETR, FCSN attains significantly lower Hausdorff scores with 19.14 (6%), 17.42 (6%), 9.16 (14%), 11.18 (22%), and 5.98 (6%) for ISIC_2018, RIM_CUP, RIM_DISC, PROSTATE, and FETAL tasks respectively. Moreover, FCSN is lightweight by discarding the decoder module, which incurs significant computational overhead. FCSN only requires 29.7 M parameters which are 75.6 M and 9.9 M fewer parameters than UNETR and DeepLabV3+, respectively. FCSN attains inference and training speeds of 1.6 ms/img and 6.3 ms/img, which is 8× and 3× faster than UNet and UNETR. The code for FCSN is made publicly available athttps://github.com/nus-mornin-lab/FCSN.

Reinforcement Learning (RL) techniques are widely adopted in healthcare research to learn effective policies for complex sequential decision-making tasks, such as treatment recommendation to achieve better clinical outcomes. However, existing RL based solutions suffer from two major limitations. (1) Burdens a single agent to make complex treatment recommendations: existing RL algorithms do not explicitly support the common hierarchical decision-making processes used in practise. That is, they consider treatment recommendation as a singular decision and the treatment space as a flat search space, and expect a single RL agent to select the best treatment outright. (2) Sub optimal patient state representation techniques: accurate representation of patient states and their dynamics is essential to make effective clinical decisions. However, there are three main limitations in current state representation techniques in the clinical environment: (i) lack of clear structure and high variability of input features; (ii) lack of a temporal context, which makes it difficult to fully capture patient conditions and their dynamics at a given time; (iii) and the use of imputation techniques, which undermines personalization, introduce biases and errors that can cause harm to patients. To address the complexities of treatment recommendation and generate superior treatment policies, we propose a consolidated RL solution that incorporates a hierarchical multi-agent architecture and advanced patient state representation technique. In our proposed architecture, the complex task is decomposed into a hierarchy of subtasks, each handled by a specialized agent that coordinates skills learned by lower level agents to determine a high-level treatment policy. The patient state representation technique is designed to effectively encode the changing states of patients throughout their hospital stay, addressing the three primary limitations of current approaches. Our experiments using the sepsis treatment task shows that the proposed solution consistently outperforms state-of-the-art baselines and improves patient survival.

The morphology and distribution of microcalcifications are the most important descriptors for radiologists to diagnose breast cancer based on mammograms. However, it is very challenging and time-consuming for radiologists to characterize these descriptors manually, and there also lacks of effective and automatic solutions for this problem. We observed that the distribution and morphology descriptors are determined by the radiologists based on the spatial and visual relationships among calcifications. Thus, we hypothesize that this information can be effectively modelled by learning a relationship-aware representation using graph convolutional networks (GCNs). In this study, we propose a multi-task deep GCN method for automatic characterization of both the morphology and distribution of microcalcifications in mammograms. Our proposed method transforms morphology and distribution characterization into node and graph classification problem and learns the representations concurrently. We trained and validated the proposed method in an in-house dataset and public DDSM dataset with 195 and 583 cases,respectively. The proposed method reaches good and stable results with distribution AUC at 0.812 ± 0.043 and 0.873 ± 0.019, morphology AUC at 0.663 ± 0.016 and 0.700 ± 0.044 for both in-house and public datasets. In both datasets, our proposed method demonstrates statistically significant improvements compared to the baseline models. The performance improvements brought by our proposed multi-task mechanism can be attributed to the association between the distribution and morphology of calcifications in mammograms, which is interpretable using graphical visualizations and consistent with the definitions of descriptors in the standard BI-RADS guideline. In short, we explore, for the first time, the application of GCNs in microcalcification characterization that suggests the potential of using graph learning for more robust understanding of medical images.

Learning information-rich and generalizable representations effectively from unlabeled multivariate cardiac signals to identify abnormal heart rhythms (cardiac arrhythmias) is valuable in real-world clinical settings but often challenging due to its complex temporal dynamics. Cardiac arrhythmias can vary significantly in temporal patterns even for the same patient (i.e., intra subject difference). Meanwhile, the same type of cardiac arrhythmia can show different temporal patterns among different patients due to different cardiac structures (i.e., inter subject difference). In this paper, we address the challenges by proposing an Intra-Inter Subject Self-Supervised Learning (ISL) model that is customized for multivariate cardiac signals. Our proposed ISL model integrates medical knowledge into self-supervision to effectively learn from intra-inter subject differences. In intra subject self-supervision, ISL model first extracts heartbeat-level features from each subject using a channel-wise attentional CNN-RNN encoder. Then a stationarity test module is employed to capture the temporal dependencies between heartbeats. In inter subject self-supervision, we design a set of data augmentations according to the clinical characteristics of cardiac signals and perform contrastive learning among subjects to learn distinctive representations for various types of patients. Extensive experiments on three real-world datasets were conducted. In a semi-supervised transfer learning scenario, our pre-trained ISL model leads about 10% improvement over supervised training when only 1% labeled data is available, suggesting strong generalizability and robustness of the model.

In data-driven medical research, multi-center studies have long been preferred over single-center ones due to a single institute sometimes not having enough data to obtain sufficient statistical power for certain hypothesis testings as well as predictive and subgroup studies. The wide adoption of electronic health records (EHRs) has made multi-institutional collaboration much more feasible. However, concerns over infrastructures, regulations, privacy, and data standardization present a challenge to data sharing across healthcare institutions. Federated Learning (FL), which allows multiple sites to collaboratively train a global model without directly sharing data, has become a promising paradigm to break the data isolation. In this study, we surveyed existing works on FL applications in EHRs and evaluated the performance of current state-of-the-art FL algorithms on two EHR machine learning tasks of significant clinical importance on a real world multi-center EHR dataset.

Driving cycle is a crucial topic for the auto industry. It is developed to provide a quantitative measure on the fuel consumption and emission of a vehicle. In recent years, massive amount of driving data has been collected but has not yet been commonly used for the evaluation of driving cycle. We believe the collection of such data and the advancement in analytics models may provide a fresh perspective for the construction of driving cycle. Therefore, we propose a novel clustering-based optimization method for the construction of driving cycles. We employ the principal component analysis and spectral clustering algorithms to eliminate redundant features and analyze data structure. We further develop an adaptive optimization algorithm to select the appropriate kinematic segments to form a representative driving cycle. To demonstrate the effectiveness of our method, we compare our performance against the baselines including the New European Driving Cycle (NEDC), Federal Test Procedure (FTP), and Markov chain-based methods. The model performance is evaluated with real driving data from two cities in Fujian, China. Our proposed method is shown to be superior to all baselines. In addition, based on our optimized driving cycle, we can also estimate the fuel consumption to evaluate its energy economy. To sum up, this study offers a novel methodology to establish the driving cycle based on real and localized traffic data, where the constructed driving cycle can further be used for the development of energy economy and emission control.

Disk fault is known to be the key cause of data loss in the modern large-scale data center, which affects the reliability and stability of the server and even the whole IT infrastructure, resulting in high financial cost. Recent works on disk fault prediction demonstrate the ability of machine learning techniques in the early prediction of disk failure. However, two limitations hinder the real-world application of current methods. First, they ignore the data heterogeneity in the data center, where distribution shifts commonly exist across different disk model types that decrease the model's performance. Second, the number of disks of different disk model types varies greatly in the data center, and current methods failed to deliver an acceptable performance over disk model types with few training samples. To address these limitations, we propose adversarial domain adaptation with correlation-based association networks (ADA-CBAN) to both mitigate the distribution shift problem and also to boost the performance on small-scaled data. Extensive experiments prove that our proposed model is effective and achieves new state-of-the-art. In addition, our post model analysis can also reveal important feature interactions and highlight the crucial period before disk faults, where both are useful information in real-world applications.

We propose an interpretable and lightweight 3D deep neural network model that diagnoses anterior cruciate ligament (ACL) tears from a knee MRI exam. Previous works focused primarily on achieving better diagnostic accuracy but paid less attention to practical aspects such as explainability and model size. They mainly relied on ImageNet pre-trained 2D deep neural network backbones, such as AlexNet or ResNet, which are computationally expensive. Some of them tried to interpret the models using post-inference visualization tools, such as CAM or Grad-CAM, which lack in generating accurate heatmaps. Our work addresses the two limitations by understanding the characteristics of ACL tear diagnosis. We argue that the semantic features required for classifying ACL tears are locally confined and highly homogeneous. We harness the unique characteristics of the task by incorporating: 1) attention modules and Gaussian positional encoding to reinforce the seeking of local features; 2) squeeze modules and fewer convolutional filters to reflect the homogeneity of the features. As a result, our model is interpretable: our attention modules can precisely highlight the ACL region without any location information given to them. Our model is extremely lightweight: consisting of only 43 K trainable parameters and 7.1 G of Floating-point operations per second (FLOPs), that is 225 times smaller and 91 times lesser than the previous state-of-the-art, respectively. Our model is accurate: our model outperforms the previous state-of-the-art with the average ROC-AUC of 0.983 and 0.980 on the Chiba and Stanford knee datasets, respectively.

BACKGROUND: The widespread adoption of electronic health records allows us to ask evidence-based questions about the need for and benefits of specific clinical interventions in critical-care settings across large populations. OBJECTIVE: We investigated the prediction of vasopressor administration and weaning in the intensive care unit. Vasopressors are commonly used to control hypotension, and changes in timing and dosage can have a large impact on patient outcomes. MATERIALS AND METHODS: We considered a cohort of 15 695 intensive care unit patients without orders for reduced care who were alive 30 days post-discharge. A switching-state autoregressive model (SSAM) was trained to predict the multidimensional physiological time series of patients before, during, and after vasopressor administration. The latent states from the SSAM were used as predictors of vasopressor administration and weaning. RESULTS: The unsupervised SSAM features were able to predict patient vasopressor administration and successful patient weaning. Features derived from the SSAM achieved areas under the receiver operating curve of 0.92, 0.88, and 0.71 for predicting ungapped vasopressor administration, gapped vasopressor administration, and vasopressor weaning, respectively. We also demonstrated many cases where our model predicted weaning well in advance of a successful wean. CONCLUSION: Models that used SSAM features increased performance on both predictive tasks. These improvements may reflect an underlying, and ultimately predictive, latent state detectable from the physiological time series.

Acute hypotension is a significant risk factor for in-hospital mortality at intensive care units. Prolonged hypotension can cause tissue hypoperfusion, leading to cellular dysfunction and severe injuries to multiple organs. Prompt medical interventions are thus extremely important for dealing with acute hypotensive episodes (AHE). Population level prognostic scoring systems for risk stratification of patients are suboptimal in such scenarios. However, the design of an efficient risk prediction system can significantly help in the identification of critical care patients, who are at risk of developing an AHE within a future time span. Toward this objective, a pattern mining algorithm is employed to extract informative sequential contrast patterns from hemodynamic data, for the prediction of hypotensive episodes. The hypotensive and normotensive patient groups are extracted from the MIMIC-II critical care research database, following an appropriate clinical inclusion criteria. The proposed method consists of a data preprocessing step to convert the blood pressure time series into symbolic sequences, using a symbolic aggregate approximation algorithm. Then, distinguishing subsequences are identified using the sequential contrast mining algorithm. These subsequences are used to predict the occurrence of an AHE in a future time window separated by a user-defined gap interval. Results indicate that the method performs well in terms of the prediction performance as well as in the generation of sequential patterns of clinical significance. Hence, the novelty of sequential patterns is in their usefulness as potential physiological biomarkers for building optimal patient risk stratification systems and for further clinical investigation of interesting patterns in critical care patients.

The ability to determine patient acuity (or severity of illness) has immediate practical use for clinicians. We evaluate the use of multivariate timeseries modeling with the multi-task Gaussian process (GP) models using noisy, incomplete, sparse, heterogeneous and unevenly-sampled clinical data, including both physiological signals and clinical notes. The learned multi-task GP (MTGP) hyperparameters are then used to assess and forecast patient acuity. Experiments were conducted with two real clinical data sets acquired from ICU patients: firstly, estimating cerebrovascular pressure reactivity, an important indicator of secondary damage for traumatic brain injury patients, by learning the interactions between intracranial pressure and mean arterial blood pressure signals, and secondly, mortality prediction using clinical progress notes. In both cases, MTGPs provided improved results: an MTGP model provided better results than single-task GP models for signal interpolation and forecasting (0.91 vs 0.69 RMSE), and the use of MTGP hyperparameters obtained improved results when used as additional classification features (0.812 vs 0.788 AUC).

Conventional hypothesis testing is carried out in a hypothesis-driven manner. A scientist must first formulate a hypothesis based on what he or she sees and then devise a variety of experiments to test it. Given the rapid growth of data, it has become virtually impossible for a person to manually inspect all data to find all of the interesting hypotheses for testing. In this article, we propose and develop a data-driven framework for automatic hypothesis testing and analysis. We define a hypothesis as a comparison between two or more subpopulations. We find subpopulations for comparison using frequent pattern mining techniques and then pair them up for statistical hypothesis testing. We also generate additional information for further analysis of the hypotheses that are deemed significant. The number of hypotheses generated can be very large, and many of them are very similar. We develop algorithms to remove redundant hypotheses and present a succinct set of significant hypotheses to users. We conducted a set of experiments to show the efficiency and effectiveness of the proposed algorithms. The results show that our system can help users (1) identify significant hypotheses efficiently, (2) isolate the reasons behind significant hypotheses efficiently, and (3) find confounding factors that form Simpson’s paradoxes with discovered significant hypotheses.

Analyzing Biomedical Big Data (BBD) is computationally expensive due to high dimensionality and large data volume. Performance and scalability issues of traditional database management systems (DBMS) often limit the usage of more sophisticated and complex data queries and analytic models. Moreover, in the conventional setting, data management and analysis use separate software platforms. Exporting and importing large amounts of data across platforms require a significant amount of computational and I/O resources, as well as potentially putting sensitive data at a security risk. In this tutorial, the participants will learn the difference between in-memory DBMS and traditional DBMS through hands-on exercises using SAP's cloud-based HANA in-memory DBMS in conjunction with the Multi-parameter Intelligent Monitoring in Intensive Care (MIMIC) dataset. MIMIC is an open-access critical care EHR archive (over 4TB in size) and consists of structured, unstructured and waveform data. Furthermore, this tutorial will seek to educate the participants on how a combination of dynamic querying, and in-memory DBMS may enhance the management and analysis of complex clinical data.

Intracranial pressure (ICP) is an important physiological signal for patients with traumatic brain injuries. Accurate ICP forecasting enables active and early interventions for more effective control of ICP levels. To achieve high accuracy, most existing methods require a high sampling rate (100 Hz), which is infeasible for online medical applications. Therefore, we propose an online ICP forecasting method requiring only low rate signal sampling (0.1 Hz). Our ARIMA based forecasting method applies empirical mode decomposition (EMD) to remove non-stationarities from the ICP signal, and robust estimation to mitigate the influence of motion induced artifacts. Experimental performance assessment with simulated and clinically collected data demonstrate that the proposed method is more accurate compared to previously proposed and standard methods.

We present a system called AssocExplorer to support exploratory data analysis via association rule visualization and exploration. AssocExplorer is designed by following the visual information-seeking mantra: overview first, zoom and filter, then details on demand. It effectively uses coloring to deliver information so that users can easily detect things that are interesting to them. If users find a rule interesting, they can explore related rules for further analysis, which allows users to find interesting phenomenon that are difficult to detect when rules are examined separately. Our system also allows users to compare rules and inspect rules with similar item composition but different statistics so that the key factors that contribute to the difference can be isolated.

Abstract Contact: [email protected]

Hypothesis testing is a well-established tool for scientific discovery. Conventional hypothesis testing is carried out in a hypothesis-driven manner. A scientist must first formulate a hypothesis based on his/her knowledge and experience, and then devise a variety of experiments to test it. Given the rapid growth of data, it has become virtually impossible for a person to manually inspect all the data to find all the interesting hypotheses for testing. In this paper, we propose and develop a data-driven system for automatic hypothesis testing and analysis. We define a hypothesis as a comparison between two or more sub-populations. We find sub-populations for comparison using frequent pattern mining techniques and then pair them up for statistical testing. We also generate additional information for further analysis of the hypotheses that are deemed significant. We conducted a set of experiments to show the efficiency of the proposed algorithms, and the usefulness of the generated hypotheses. The results show that our system can help users (1) identify significant hypotheses; (2) isolate the reasons behind significant hypotheses; and (3) find confounding factors that form Simpson's Paradoxes with discovered significant hypotheses.

MOTIVATION: Many new methods have recently been proposed for detecting epistatic interactions in GWAS data. There is, however, no in-depth independent comparison of these methods yet. RESULTS: Five recent methods-TEAM, BOOST, SNPHarvester, SNPRuler and Screen and Clean (SC)-are evaluated here in terms of power, type-1 error rate, scalability and completeness. In terms of power, TEAM performs best on data with main effect and BOOST performs best on data without main effect. In terms of type-1 error rate, TEAM and BOOST have higher type-1 error rates than SNPRuler and SNPHarvester. SC does not control type-1 error rate well. In terms of scalability, we tested the five methods using a dataset with 100 000 SNPs on a 64 bit Ubuntu system, with Intel (R) Xeon(R) CPU 2.66 GHz, 16 GB memory. TEAM takes ~36 days to finish and SNPRuler reports heap allocation problems. BOOST scales up to 100 000 SNPs and the cost is much lower than that of TEAM. SC and SNPHarvester are the most scalable. In terms of completeness, we study how frequently the pruning techniques employed by these methods incorrectly prune away the most significant epistatic interactions. We find that, on average, 20% of datasets without main effect and 60% of datasets with main effect are pruned incorrectly by BOOST, SNPRuler and SNPHarvester. AVAILABILITY: The software for the five methods tested are available from the URLs below. TEAM: http://csbio.unc.edu/epistasis/download.php BOOST: http://ihome.ust.hk/~eeyang/papers.html. SNPHarvester: http://bioinformatics.ust.hk/SNPHarvester.html. SNPRuler: http://bioinformatics.ust.hk/SNPRuler.zip. Screen and Clean: http://wpicr.wpic.pitt.edu/WPICCompGen/. CONTACT: [email protected].

This article addresses the incremental and decremental maintenance of the frequent pattern space. We conduct an in‐depth investigation on how the frequent pattern space evolves under both incremental and decremental updates. Based on the evolution analysis, a new data structure, Generator‐Enumeration Tree (GE‐tree), is developed to facilitate the maintenance of the frequent pattern space. With the concept of GE‐tree, we propose two novel algorithms, Pattern Space Maintainer+ (PSM+) and Pattern Space Maintainer− (PSM−), for the incremental and decremental maintenance of frequent patterns. Experimental results demonstrate that the proposed algorithms, on average, outperform the representative state‐of‐the‐art methods by an order of magnitude.

We are often interested to test whether a given cause has a given effect. If we cannot specify the nature of the factors involved, such tests are called model-free studies. There are two major strategies to demonstrate associations between risk factors (ie. patterns) and outcome phenotypes (ie. class labels). The first is that of prospective study designs, and the analysis is based on the concept of "relative risk": What fraction of the exposed (ie. has the pattern) or unexposed (ie. lacks the pattern) individuals have the phenotype (ie. the class label)? The second is that of retrospective designs, and the analysis is based on the concept of "odds ratio": The odds that a case has been exposed to a risk factor is compared to the odds for a case that has not been exposed. The efficient extraction of patterns that have good relative risk and/or odds ratio has not been previously studied in the data mining context. In this paper, we investigate such patterns. We show that this pattern space can be systematically stratified into plateaus of convex spaces based on their support levels. Exploiting convexity, we formulate a number of sound and complete algorithms to extract the most general and the most specific of such patterns at each support level. We compare these algorithms. We further demonstrate that the most efficient among these algorithms is able to mine these sophisticated patterns at a speed comparable to that of mining frequent closed patterns, which are patterns that satisfy considerably simpler conditions.

This paper proposes an adaptive binarization method, based on the criterion of maximizing local contrast, for document image analysis. The proposed method has overcome, to a large extent, the general problems of poor quality document images, such as non-uniform illumination, undesirable shadows and random noise. It was tested against a variety of challenging images, and the experimental results are presented to show the effectiveness and superiority of the proposed method.