Continual Learning (CL) aims to enable models to sequentially learn multiple tasks without forgetting previous knowledge. Recent studies have shown that optimizing towards flatter loss minima can improve model generalization. However, existing sharpness-aware methods for CL suffer from two key limitations: (1) they treat sharpness regularization as a unified signal without distinguishing the contributions of its components. and (2) they introduce substantial computational overhead that impedes practical deployment. To address these challenges, we propose FLAD, a novel optimization framework that decomposes sharpness-aware perturbations into gradient-aligned and stochastic-noise components, and show that retaining only the noise component promotes generalization. We further introduce a lightweight scheduling scheme that enables FLAD to maintain significant performance gains even under constrained training time. FLAD can be seamlessly integrated into various CL paradigms and consistently outperforms standard and sharpness-aware optimizers in diverse experimental settings, demonstrating its effectiveness and practicality in CL.

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.

Monitoring Earth's evolving land covers requires methods capable of detecting changes across a wide range of categories and contexts. Existing change detection methods are hindered by their dependency on predefined classes, reducing their effectiveness in open-world applications. To address this issue, we introduce open-vocabulary change detection (OVCD), a novel task that bridges vision and language to detect changes across any category. Considering the lack of high-quality data and annotation, we propose two training-free frameworks, M-C-I and I-M-C, which leverage and integrate off-the-shelf foundation models for the OVCD task. The insight behind the M-C-I~framework is to discover all potential changes and then classify these changes, while the insight of I-M-C~framework is to identify all targets of interest and then determine whether their states have changed. Based on these two frameworks, we instantiate to obtain several methods, e.g., SAM-DINOv2-SegEarth-OV, Grounding-DINO-SAM2-DINO, etc. Extensive evaluations on 4 benchmark datasets demonstrate the superior generalization and robustness of our OVCD methods over existing supervised and unsupervised methods. To support continued exploration, we release DynamicEarth, a dedicated codebase designed to advance research and application of OVCD.

Interactive segmentation offers greater clinical potential in computational pathology compared to traditional automatic segmentation. By incorporating interactive input, it addresses the limitations of fully automatic segmentation models, which often fail to meet pathologists' requirements and rely heavily on large-scale, pixel-level annotated datasets. However, current interactive segmentation methods struggle to balance interaction cost and segmentation performance, and they fail to adapt effectively to slide-level segmentation, a task that is even more crucial in routine pathology analysis. In this study, we propose a Prototype-Guided Interactive Segmentation (ProGIS) framework for pathological image segmentation, designed to deliver precise segmentation results efficiently with minimal interaction signals. ProGIS identifies all same-type tissue connected components in a single interaction and supports multi-class segmentation without predefined categories during inference. Moreover, ProGIS can be easily adapted for slide-level interactive segmentation. Specifically, ProGIS consists of three modules: Prototype Initialization, Prototype Navigation, and Local Refinement. First, the Prototype Initialization module identifies categorical prototypes, which are then utilized in the Prototype Navigation module to identify all tissue connected components belonging to the same type. The local refinement module further refines the segmentation results using detailed correction signals to ensure the accuracy of challenging-to-distinguish regions. We evaluate our framework on two regions of interest level and two slide-level pathological segmentation datasets, achieving new state-of-the-art performance with fewer interactions than existing methods. Our code is available at https://github.com/JSGe-AI/ProGIS.

Text-to-image diffusion models significantly enhance the efficiency of artistic creation with high-fidelity image generation. However, in typical application scenarios like comic book production, they can neither place each subject into its expected spot nor maintain the consistent appearance of each subject across images. For these issues, we pioneer a novel task, Layout-to-Consistent-Image (L2CI) generation, which produces consistent and compositional images in accordance with the given layout conditions and text prompts. To accomplish this challenging task, we present a new formalization of dual energy guidance with optimization in a dual semantic-latent space and thus propose a training-free pipeline, SpotActor, which features a layout-conditioned optimizing stage and a consistent sampling stage. In the optimizing stage, we innovate a nuanced layout energy function to mimic the attention activations with a sigmoid-like objective. While in the sampling stage, we design Regional Interconnection Self-Attention (RISA) and Semantic Fusion Cross-Attention (SFCA) mechanisms that allow mutual interactions across images. To evaluate the performance, we present ActorBench, a specified benchmark with hundreds of reasonable prompt-box pairs stemming from object detection datasets. Comprehensive experiments are conducted to demonstrate the effectiveness of our method. The results prove that SpotActor fulfills the expectations of this task and showcases the potential for practical applications with superior layout alignment, subject consistency, prompt conformity and background diversity.

Classifier-Free Guidance (CFG), which combines the conditional and unconditional score functions with two coefficients summing to one, serves as a practical technique for diffusion model sampling. Theoretically, however, denoising with CFG cannot be expressed as a reciprocal diffusion process, which may consequently leave some hidden risks during use. In this work, we revisit the theory behind CFG and rigorously confirm that the improper configuration of the combination coefficients (i.e., the widely used summing-to-one version) brings about expectation shift of the generative distribution. To rectify this issue, we propose ReCFG1with a relaxation on the guidance coefficients such that denoising with ReCFG strictly aligns with the diffusion theory. We further show that our approach enjoys a closed-form solution given the guidance strength. That way, the rectified coefficients can be readily pre-computed via traversing the observed data, leaving the sampling speed barely affected. Empirical evidence on real-world data demonstrate the compatibility of our post-hoc design with existing state-of-the-art diffusion models, including both class-conditioned ones (e.g., EDM2 on ImageNet) and text-conditioned ones (e.g., SD3 on CC12M), without any retraining. Code is available at https://github.com/thuxmf/recfg.

Adversarially robust generalization of Graph Convolutional Networks (GCNs) has garnered significant attention in various security-sensitive application areas, driven by intrinsic adversarial vulnerability. Albeit remarkable empirical advancement, theoretical understanding of the generalization behavior of GCNs subjected to adversarial attacks remains elusive. To make progress on the mystery, we establish unified high-probability generalization bounds for GCNs in the context of node classification, by leveraging adversarial Transductive Rademacher Complexity (TRC) and developing a novel contraction technique on graph convolution. Our bounds capture the interaction between generalization error and adversarial perturbations, revealing the importance of key quantities in mitigating the negative effects of perturbations, such as low-dimensional feature projection, perturbation-dependent norm regularization, normalized graph matrix, proper number of network layers, etc. Furthermore, we provide TRC-based bounds of popular GCNs with $\ell_r$-norm-additive perturbations for arbitrary $r\geq 1$. A comparison of theoretical results demonstrates that specific network architectures (e.g., residual connection) can help alleviate the cumulative effect of perturbations during the forward propagation of deep GCNs. Experimental results on benchmark datasets validate our theoretical findings.

Information-theoretic bounds, while achieving significant success in analyzing the generalization of randomized learning algorithms, have been criticized for their slow convergence rates and overestimation. This paper presents novel bounds that bridge the expected empirical and population risks through a binarized variant of the Jensen-Shannon divergence. Leveraging our foundational lemma that characterizes the interaction between an arbitrary and a binary variable, we derive hypothesis-based bounds that enhance existing conditional mutual information bounds by reducing the number of conditioned samples from $2$ to $1$. We additionally establish prediction-based bounds that surpass prior bounds based on evaluated loss mutual information measures. Thereafter, through a new binarization technique for the evaluated loss variables, we obtain exactly tight generalization bounds broadly applicable to general randomized learning algorithms for any bounded loss functions. Our results effectively address key limitations of previous results in analyzing certain stochastic convex optimization problems, without requiring additional stability or compressibility assumptions about the learning algorithm.

The Segment Anything Model (SAM), a vision foundation model, exhibits impressive zero-shot capabilities in general tasks but struggles in specialized domains. Parameter-efficient fine-tuning (PEFT) is a promising approach to unleash the potential of SAM in novel scenarios. However, existing PEFT methods for SAM neglect the domain-invariant relations encoded in the pre-trained model. To bridge this gap, we propose InfoSAM, an information-theoretic approach that enhances SAM fine-tuning by distilling and preserving its pre-trained segmentation knowledge. Specifically, we formulate the knowledge transfer process as two novel mutual information-based objectives: (i) to compress the domain-invariant relation extracted from pre-trained SAM, excluding pseudo-invariant information as possible, and (ii) to maximize mutual information between the relational knowledge learned by the teacher (pre-trained SAM) and the student (fine-tuned model). The proposed InfoSAM establishes a robust distillation framework for PEFT of SAM. Extensive experiments across diverse benchmarks validate InfoSAM’s effectiveness in improving SAM family’s performance on real-world tasks, demonstrating its adaptability and superiority in handling specialized scenarios. The code and models are available at https://muyaoyuan.github.io/InfoSAM_Page.

Recently, the mathematical tool from fractal geometry (i.e., fractal dimension) has been employed to investigate optimization trajectory-dependent generalization ability for some pointwise learning models with independent and identically distributed (i.i.d.) observations. This paper goes beyond the limitations of pointwise learning and i.i.d. samples, and establishes generalization bounds for pairwise learning with uniformly strong mixing samples. The derived theoretical results fill the gap of trajectory-dependent generalization analysis for pairwise learning, and can be applied to wide learning paradigms, e.g., metric learning, ranking and gradient learning. Technically, our framework brings concentration estimation with Rademacher complexity and trajectory-dependent fractal dimension together in a coherent way for felicitous learning theory analysis. In addition, the efficient computation of fractal dimension can be guaranteed for random algorithms (e.g., stochastic gradient descent algorithm for deep neural networks) by bridging topological data analysis tools and the trajectory-dependent fractal dimension.

Domain generalization aims to learn invariance across multiple source domains, thereby enhancing generalization against out-of-distribution data. While gradient or representation matching algorithms have achieved remarkable success in domain generalization, these methods generally lack generalization guarantees or depend on strong assumptions, leaving a gap in understanding the underlying mechanism of distribution matching. In this work, we formulate domain generalization from a novel probabilistic perspective, ensuring robustness while avoiding overly conservative solutions. Through comprehensive information-theoretic analysis, we provide key insights into the roles of gradient and representation matching in promoting generalization. Our results reveal the complementary relationship between these two components, indicating that existing works focusing solely on either gradient or representation alignment are insufficient to solve the domain generalization problem. In light of these theoretical findings, we introduce IDM to simultaneously align the inter-domain gradients and representations. Integrated with the proposed PDM method for complex distribution matching, IDM achieves superior performance over various baseline methods.

The proliferation of video data and advancements in Deep Neural Networks (DNNs) have greatly boosted live video analytics, driven by the growing video capture capabilities of mobile devices. However, resource limitations necessitate the transmission of endpoint-collected videos to servers for inference. To meet real-time requirements and ensure accurate inference, it is essential to adjust video configurations at the endpoint. Traditional methods rely on deterministic strategies, posing difficulties in adapting to dynamic networks and video content. Meanwhile, emerging learning-based schemes suffer from trial-and-error exploration mechanisms, resulting in a concerning long-tail effect on upload latency. In this paper, we propose a novel lightweight and robust configuration adaptation policy (LCA), which fuses heuristic and RL-based agents using multi-teacher knowledge distillation (MKD) theory. Firstly, we propose a content-sensitive and bandwidth-adaptive RL agent and introduce a Lyapunov-based optimization agent for ensuring latency robustness. To leverage both agents' strengths, we design a feature-guided multi-teacher distillation network to transfer their advantages to the student. The experimental results across two vision tasks (pose estimation and semantic segmentation) demonstrate that LCA significantly reduces transmission latency compared to prior work (average reduction of 47.11%-89.55%, 95-percentile reduction of 27.63%-88.78%) and computational overhead while maintaining comparable inference accuracy.

Multiple instance learning (MIL) has emerged as a prominent paradigm for processing the whole slide image with pyramid structure and giga-pixel size in digital pathology. However, existing attention-based MIL methods are primarily trained on the image modality and a pre-defined label set, leading to limited generalization and interpretability. Recently, vision language models (VLM) have achieved promising performance and transferability, offering potential solutions to the limitations of MIL-based methods. Pathological diagnosis is an intricate process that requires pathologists to examine the WSI step-by-step. In the field of natural language process, the chain-of-thought (CoT) prompting method is widely utilized to imitate the human reasoning process. Inspired by the CoT prompt and pathologists' clinic knowledge, we propose a chain-of-diagnosis prompting multiple instance learning (CoD-MIL) framework for whole slide image classification. Specifically, the chain-of-diagnosis text prompt decomposes the complex diagnostic process in WSI into progressive sub-processes from low to high magnification. Additionally, we propose a text-guided contrastive masking module to accurately localize the tumor region by masking the most discriminative instances and introducing the guidance of normal tissue texts in a contrastive way. Extensive experiments conducted on three real-world subtyping datasets demonstrate the effectiveness and superiority of CoD-MIL.

The matrix-based Rényi's entropy (MBRE) has recently been introduced as a substitute for the original Rényi's entropy that could be directly obtained from data samples, avoiding the expensive intermediate step of density estimation. Despite its remarkable success in a broad of information-related tasks, the computational cost of MBRE, however, becomes a bottleneck for large-scale applications. The challenge, when facing sequential data, is further amplified due to the requirement of large-scale eigenvalue decomposition on multiple dense kernel matrices constructed by sliding windows in the region of interest, resulting in overall time complexity, where and denote the number and the size of windows, respectively. To overcome this issue, we adopt the static MBRE estimator together with a variance reduction criterion to develop randomized approximations for the target entropy, leading to high accuracy with substantially lower query complexity by utilizing the historical estimation results. Specifically, assuming that the changes of adjacent sliding windows are bounded by , which is a trivial case in domains, e.g., time-series analysis, we lower the complexity by a factor of . Polynomial approximation techniques are further adopted to support arbitrary orders. In general, our algorithms achieve total computational complexity, where denote the number of vector queries and the polynomial degrees, respectively. Theoretical upper and lower bounds are established in terms of the convergence rate for both and , and large-scale experiments on both simulation and real-world data are conducted to validate the effectiveness of our algorithms. The results show that our methods achieve promising speedup with only a trivial loss in performance.

Accurate histopathological image segmentation is crucial for precise disease diagnosis and prognosis. Yet, challenges like staining variations, imaging conditions, and tissue diversity impede model generalization across domains, such as different institutes or organs. Traditional domain generalization (DG) techniques, such as data augmentation and feature alignment, excel in classification tasks but face challenges in segmentation tasks due to their dense prediction requirements. These tasks are particularly computationally demanding, and are complicated due to the fine-grained feature variability that arises from the domain differences in histopathological images. To tackle this, we propose the Shallow-Deep Synergy (SDS) approach for the U-Net-based segmentation framework, which capitalizes on the distinctive characteristics of both shallow and deep layers of the U-Net. Specifically, we introduce the fine-grained domain variations in image intensities and textures for shallow layers, while focusing on aligning the pixel-level classification decision boundaries in deep layers by adjusting the optimization trajectory through class-wise gradient and feature alignment. Moreover, the SDS is equipped with a big-batch strategy further boosting alignment efficiency, achieving high accuracy without substantial GPU memory. Extensive experiments conducted on two histopathological segmentation datasets, each representing different domain types, demonstrate that the proposed SDS achieves superior generalization performance compared to existing domain generalization methods, even being competitive with intra-domain models in some cases.

Multiple instance learning (MIL)-based framework has become the mainstream for processing the whole slide image (WSI) with giga-pixel size and hierarchical image context in digital pathology. However, these methods heavily depend on a substantial number of bag-level labels and solely learn from the original slides, which are easily affected by variations in data distribution. Recently, vision language model (VLM)-based methods introduced the language prior by pre-training on large-scale pathological image-text pairs. However, the previous text prompt lacks the consideration of pathological prior knowledge, there-fore does not substantially boost the model's performance. Moreover, the collection of such pairs and the pre-training process are very time-consuming and source-intensive. To solve the above problems, we propose a dual-scale vision-language multiple instance learning (ViLa-MIL) framework for whole slide image classification. Specifically, we propose a dual-scale visual descriptive text prompt based on the frozen large language model (LLM) to boost the performance of VLM effectively. To transfer the VLM to process WSI efficiently, for the image branch, we propose a prototype-guided patch decoder to aggregate the patch features progressively by grouping similar patches into the same prototype; for the text branch, we introduce a context-guided text decoder to enhance the text features by incorporating the multi-granular image contexts. Extensive studies on three multi-cancer and multi-center subtyping datasets demonstrate the superiority of ViLa-MIL.

Information-theoretic generalization analysis has achieved astonishing success in characterizing the generalization capabilities of noisy and iterative learning algorithms. However, current advancements are mostly restricted to average-case scenarios and necessitate the stringent bounded loss assumption, leaving a gap with regard to computationally tractable PAC generalization analysis, especially for long-tailed loss distributions. In this paper, we bridge this gap by introducing a novel class of PAC bounds through leveraging loss entropies. These bounds simplify the computation of key information metrics in previous PAC information-theoretic bounds to one-dimensional variables, thereby enhancing computational tractability. Moreover, our data-independent bounds provide novel insights into the generalization behavior of the minimum error entropy criterion, while our data-dependent bounds improve over previous results by alleviating the bounded loss assumption under both leave-one-out and supersample settings. Extensive numerical studies indicate strong correlations between the generalization error and the induced loss entropy, showing that the presented bounds adeptly capture the patterns of the true generalization gap under various learning scenarios.

The recent surge in contrastive learning has intensified the interest in understanding the generalization of non-pointwise learning paradigms. While information-theoretic analysis achieves remarkable success in characterizing the generalization behavior of learning algorithms, its applicability is largely confined to pointwise learning, with extensions to the simplest pairwise settings remaining unexplored due to the challenges of non-i.i.d losses and dimensionality explosion. In this paper, we develop the first series of information-theoretic bounds extending beyond pointwise scenarios, encompassing pointwise, pairwise, triplet, quadruplet, and higher-order scenarios, all within a unified framework. Specifically, our hypothesis-based bounds elucidate the generalization behavior of iterative and noisy learning algorithms via gradient covariance analysis, and our prediction-based bounds accurately estimate the generalization gap with computationally tractable low-dimensional information metrics. Comprehensive numerical studies then demonstrate the effectiveness of our bounds in capturing the generalization dynamics across diverse learning scenarios.

Non-maximum suppression (NMS) is an indispensable post-processing step in object detection. With the continuous optimization of network models, NMS has become the ``last mile'' to enhance the efficiency of object detection. This paper systematically analyzes NMS from a graph theory perspective for the first time, revealing its intrinsic structure. Consequently, we propose two optimization methods, namely QSI-NMS and BOE-NMS. The former is a fast recursive divide-and-conquer algorithm with negligible mAP loss, and its extended version (eQSI-NMS) achieves optimal complexity of $\mathcal{O}(n\log n)$. The latter, concentrating on the locality of NMS, achieves an optimization at a constant level without an mAP loss penalty. Moreover, to facilitate rapid evaluation of NMS methods for researchers, we introduce NMS-Bench, the first benchmark designed to comprehensively assess various NMS methods. Taking the YOLOv8-N model on MS COCO 2017 as the benchmark setup, our method QSI-NMS provides $6.2\times$ speed of original NMS on the benchmark, with a $0.1\%$ decrease in mAP. The optimal eQSI-NMS, with only a $0.3\%$ mAP decrease, achieves $10.7\times$ speed. Meanwhile, BOE-NMS exhibits $5.1\times$ speed with no compromise in mAP.

Text-to-image diffusion models benefit artists with high-quality image generation. Yet their stochastic nature hinders artists from creating consistent images of the same subject. Existing methods try to tackle this challenge and generate consistent content in various ways. However, they either depend on external restricted data or require expensive tuning of the diffusion model. For this issue, we propose a novel one-shot tuning paradigm, termed OneActor. It efficiently performs consistent subject generation solely driven by prompts via a learned semantic guidance to bypass the laborious backbone tuning. We lead the way to formalize the objective of consistent subject generation from a clustering perspective, and thus design a cluster-conditioned model. To mitigate the overfitting challenge shared by one-shot tuning pipelines, we augment the tuning with auxiliary samples and devise two inference strategies: semantic interpolation and cluster guidance. These techniques are later verified to significantly improve the generation quality. Comprehensive experiments show that our method outperforms a variety of baselines with satisfactory subject consistency, superior prompt conformity as well as high image quality. Our method is capable of multi-subject generation and compatible with popular diffusion extensions. Besides, we achieve a $4\times$ faster tuning speed than tuning-based baselines and, if desired, avoid increasing the inference time. Furthermore, our method can be naturally utilized to pre-train a consistent subject generation network from scratch, which will implement this research task into more practical applications. (Project page: https://johnneywang.github.io/OneActor-webpage/)

Biomedical Coreference Resolution focuses on identifying the coreferences in biomedical texts, which normally consists of two parts: (i) mention detection to identify textual representation of biological entities and (ii) finding their coreference links. Recently, a popular approach to enhance the task is to embed knowledge base into deep neural networks. However, the way in which these methods integrate knowledge leads to the shortcoming that such knowledge may play a larger role in mention detection than coreference resolution. Specifically, they tend to integrate knowledge prior to mention detection, as part of the embeddings. Besides, they primarily focus on mention-dependent knowledge (KBase), i.e., knowledge entities directly related to mentions, while ignores the correlated knowledge (K+) between mentions in the mention-pair. For mentions with significant differences in word form, this may limit their ability to extract potential correlations between those mentions. Thus, this paper develops a novel model to integrate both KBase and K+ entities and achieves the state-of-the-art performance on BioNLP and CRAFT-CR datasets. Empirical studies on mention detection with different length reveals the effectiveness of the KBase entities. The evaluation on cross-sentence and match/mismatch coreference further demonstrate the superiority of the K+ entities in extracting background potential correlation between mentions.

Subsampling is an important technique to tackle the computational challenges brought by big data. Many subsampling procedures fall within the framework of importance sampling, which assigns high sampling probabilities to the samples appearing to have big impacts. When the noise level is high, those sampling procedures tend to pick many outliers and thus often do not perform satisfactorily in practice. To tackle this issue, we design a new Markov subsampling strategy based on Huber criterion (HMS) to construct an informative subset from the noisy full data; the constructed subset then serves as refined working data for efficient processing. HMS is built upon a Metropolis-Hasting procedure, where the inclusion probability of each sampling unit is determined using the Huber criterion to prevent over scoring the outliers. Under mild conditions, we show that the estimator based on the subsamples selected by HMS is statistically consistent with a sub-Gaussian deviation bound. The promising performance of HMS is demonstrated by extensive studies on large-scale simulations and real data examples.

Prompt tuning has achieved great success in various sentence-level classification tasks by using elaborated label word mappings and prompt templates. However, for solving token-level classification tasks, e.g., named entity recognition (NER), previous research, which utilizes N-gram traversal for prompting all spans with all possible entity types, is time-consuming. To this end, we propose a novel prompt-based contrastive learning method for few-shot NER without template construction and label word mappings. First, we leverage external knowledge to initialize semantic anchors for each entity type. These anchors are simply appended with input sentence embeddings as template-free prompts (TFPs). Then, the prompts and sentence embeddings are in-context optimized with our proposed semantic-enhanced contrastive loss. Our proposed loss function enables contrastive learning in few-shot scenarios without requiring a significant number of negative samples. Moreover, it effectively addresses the issue of conventional contrastive learning, where negative instances with similar semantics are erroneously pushed apart in natural language processing (NLP)-related tasks. We examine our method in label extension (LE), domain-adaption (DA), and low-resource generalization evaluation tasks with six public datasets and different settings, achieving state-of-the-art (SOTA) results in most cases.

Triplet learning, i.e. learning from triplet data, has attracted much attention in computer vision tasks with an extremely large number of categories, e.g., face recognition and person re-identification. Albeit with rapid progress in designing and applying triplet learning algorithms, there is a lacking study on the theoretical understanding of their generalization performance. To fill this gap, this paper investigates the generalization guarantees of triplet learning by leveraging the stability analysis. Specifically, we establish the first general high-probability generalization bound for the triplet learning algorithm satisfying the uniform stability, and then obtain the excess risk bounds of the order O(log(n)/(√n) ) for both stochastic gradient descent (SGD) and regularized risk minimization (RRM), where 2n is approximately equal to the number of training samples. Moreover, an optimistic generalization bound in expectation as fast as O(1/n) is derived for RRM in a low noise case via the on-average stability analysis. Finally, our results are applied to triplet metric learning to characterize its theoretical underpinning.

The matrix-based Rényi's entropy allows us to directly quantify information measures from given data, without explicit estimation of the underlying probability distribution. This intriguing property makes it widely applied in statistical inference and machine learning tasks. However, this information theoretical quantity is not robust against noise in the data, and is computationally prohibitive in large-scale applications. To address these issues, we propose a novel measure of information, termed low-rank matrix-based Rényi's entropy, based on low-rank representations of infinitely divisible kernel matrices. The proposed entropy functional inherits the specialty of of the original definition to directly quantify information from data, but enjoys additional advantages including robustness and effective calculation. Specifically, our low-rank variant is more sensitive to informative perturbations induced by changes in underlying distributions, while being insensitive to uninformative ones caused by noises. Moreover, low-rank Rényi's entropy can be efficiently approximated by random projection and Lanczos iteration techniques, reducing the overall complexity from O(n³) to O(n²s) or even O(ns²), where n is the number of data samples and s ≪ n. We conduct large-scale experiments to evaluate the effectiveness of this new information measure, demonstrating superior results compared to matrix-based Rényi's entropy in terms of both performance and computational efficiency.

Additive models have been burgeoning in data analysis due to their flexible representation and desirable interpretability. However, most existing approaches are constructed under empirical risk minimization (ERM), and thus perform poorly in situations where average performance is not a suitable criterion for the problems of interest, e.g., data with complex non-Gaussian noise, imbalanced labels or both of them. In this paper, a novel class of sparse additive models is proposed under tilted empirical risk minimization (TERM), which addresses the deficiencies in ERM by imposing tilted impact on individual losses, and is flexibly capable of achieving a variety of learning objectives, e.g., variable selection, robust estimation, imbalanced classification and multiobjective learning. On the theoretical side, a learning theory analysis which is centered around the generalization bound and function approximation error bound (under some specific data distributions) is conducted rigorously. On the practical side, an accelerated optimization algorithm is designed by integrating Prox-SVRG and random Fourier acceleration technique. The empirical assessments verify the competitive performance of our approach on both synthetic and real data.

Recently, information-theoretic analysis has become a popular framework for understanding the generalization behavior of deep neural networks. It allows a direct analysis for stochastic gradient / Langevin descent (SGD/SGLD) learning algorithms without strong assumptions such as Lipschitz or convexity conditions. However, the current generalization error bounds within this framework are still far from optimal, while substantial improvements on these bounds are quite challenging due to the intractability of high-dimensional information quantities. To address this issue, we first propose a novel information theoretical measure: kernelized Rényi's entropy, by utilizing operator representation in Hilbert space. It inherits the properties of Shannon's entropy and can be effectively calculated via simple random sampling, while remaining independent of the input dimension. We then establish the generalization error bounds for SGD/SGLD under kernelized Rényi's entropy, where the mutual information quantities can be directly calculated, enabling evaluation of the tightness of each intermediate step. We show that our information-theoretical bounds depend on the statistics of the stochastic gradients evaluated along with the iterates, and are rigorously tighter than the current state-of-the-art (SOTA) results. The theoretical findings are also supported by large-scale empirical studies.

MOTIVATION: Artificially making clinical decisions for patients with multi-morbidity has long been considered a thorny problem due to the complexity of the disease. Drug recommendations can assist doctors in automatically providing effective and safe drug combinations conducive to treatment and reducing adverse reactions. However, the existing drug recommendation works ignored two critical information. (i) Different types of medical information and their interrelationships in the patient's visit history can be used to construct a comprehensive patient representation. (ii) Patients with similar disease characteristics and their corresponding medication information can be used as a reference for predicting drug combinations. RESULTS: To address these limitations, we propose DAPSNet, which encodes multi-type medical codes into patient representations through code- and visit-level attention mechanisms, while integrating drug information corresponding to similar patient states to improve the performance of drug recommendation. Specifically, our DAPSNet is enlightened by the decision-making process of human doctors. Given a patient, DAPSNet first learns the importance of patient history records between diagnosis, procedure and drug in different visits, then retrieves the drug information corresponding to similar patient disease states for assisting drug combination prediction. Moreover, in the training stage, we introduce a novel information constraint loss function based on the information bottleneck principle to constrain the learned representation and enhance the robustness of DAPSNet. We evaluate the proposed DAPSNet on the public MIMIC-III dataset, our model achieves relative improvements of 1.33%, 1.20% and 2.03% in Jaccard, F1 and PR-AUC scores, respectively, compared to state-of-the-art methods. AVAILABILITY AND IMPLEMENTATION: The source code is available at the github repository: https://github.com/andylun96/DAPSNet.

Aspect-based sentiment analysis (ABSA) means to identify fine-grained aspects, opinions, and sentiment polarities. Recent ABSA research focuses on utilizing multi-task learning (MTL) to achieve less computational costs and better performance. However, there are certain limits in MTL-based ABSA. For example, unbalanced labels and sub-task learning difficulties may result in the biases that some labels and sub-tasks are overfitting, while the others are underfitting. To address these issues, inspired by neuro-symbolic learning systems, we propose a meta-based self-training method with a meta-weighter (MSM). We believe that a generalizable model can be achieved by appropriate symbolic representation selection (in-domain knowledge) and effective learning control (regulation) in a neural system. Thus, MSM trains a teacher model to generate in-domain knowledge (e.g., unlabeled data selection and pseudo-label generation), where the generated pseudo-labels are used by a student model for supervised learning. Then, the meta-weighter of MSM is jointly trained with the student model to provide each instance with sub-task-specific weights to coordinate their convergence rates, balancing class labels, and alleviating noise impacts introduced from self-training. The following experiments indicate that MSM can utilize 50% labeled data to achieve comparable results to state-of-arts models in ABSA and outperform them with all labeled data.

The Matrix-based Rényi’s entropy enables us to directly measure information quantities from given data without the costly probability density estimation of underlying distributions, thus has been widely adopted in numerous statistical learning and inference tasks. However, exactly calculating this new information quantity requires access to the eigenspectrum of a semi-positive definite (SPD) matrix$A$which grows linearly with the number of samples$n$, resulting in a$O(n^{3})$time complexity that is prohibitive for large-scale applications. To address this issue, this paper takes advantage of stochastic trace approximations for matrix-based Rényi’s entropy with arbitrary$\alpha \in \mathbb {R}^{+}$orders, lowering the complexity by converting the entropy approximation to a matrix-vector multiplication problem. Specifically, we develop random approximations for integer-order$\alpha $cases and polynomial series approximations (Taylor and Chebyshev) for fractional$\alpha $cases, leading to a$O(n^{2}sm)$overall time complexity, where$s, m \ll n$denote the number of vector queries and the polynomial order respectively. We theoretically establish statistical guarantees for all approximation algorithms and give explicit order of$s$and$m$with respect to the approximation error$\epsilon $, showing optimal convergence rate for both parameters up to a logarithmic factor. Large-scale simulations and real-world applications validate the effectiveness of the developed approximations, demonstrating remarkable speedup with negligible loss in accuracy.

Accurate tissue segmentation in histopathological images is essential for promoting the development of precision pathology. However, the size of the digital pathological image is great, which needs to be tiled into small patches containing limited semantic information. To imitate the pathologist's diagnosis process and model the semantic relation of the whole slide image, We propose a semi-supervised pixel contrastive learning framework (SSPCL) which mainly includes an uncertainty-guided mutual dual consistency learning module (UMDC) and a cross image pixel-contrastive learning module (CIPC). The UMDC module enables efficient learning from unlabeled data through mutual dual-consistency and consensus-based uncertainty. The CIPC module aims at capturing the cross-patch semantic relationship by optimizing a contrastive loss between pixel embeddings. We also propose several novel domain-related sampling methods by utilizing the continuous spatial structure of adjacent image patches, which can avoid the problem of false sampling and improve the training efficiency. In this way, SSPCL significantly reduces the labeling cost on histopathological images and realizes the accurate quantitation of tissues. Extensive experiments on three tissue segmentation datasets demonstrate the effectiveness of SSPCL, which outperforms state-of-the-art up to 5.0% in mDice.

Leukemia classification relies on a detailed cytomorphological examination of Bone Marrow (BM) smear. However, applying existing deep-learning methods to it is facing two significant limitations. Firstly, these methods require large-scale datasets with expert annotations at the cell level for good results and typically suffer from poor generalization. Secondly, they simply treat the BM cytomorphological examination as a multi-class cell classification task, thus failing to exploit the correlation among leukemia subtypes over different hierarchies. Therefore, BM cytomorphological estimation as a time-consuming and repetitive process still needs to be done manually by experienced cytologists. Recently, Multi-Instance Learning (MIL) has achieved much progress in data-efficient medical image processing, which only requires patient-level labels (which can be extracted from the clinical reports). In this paper, we propose a hierarchical MIL framework and equip it with Information Bottleneck (IB) to tackle the above limitations. First, to handle the patient-level label, our hierarchical MIL framework uses attention-based learning to identify cells with high diagnostic values for leukemia classification in different hierarchies. Then, following the information bottleneck principle, we propose a hierarchical IB to constrain and refine the representations of different hierarchies for better accuracy and generalization. By applying our framework to a large-scale childhood acute leukemia dataset with corresponding BM smear images and clinical reports, we show that it can identify diagnostic-related cells without the need for cell-level annotations and outperforms other comparison methods. Furthermore, the evaluation conducted on an independent test cohort demonstrates the high generalizability of our framework.

Multiple instance learning (MIL)-based methods have become the mainstream for processing the megapixel-sized whole slide image (WSI) with pyramid structure in the field of digital pathology. The current MIL-based methods usually crop a large number of patches from WSI at the highest magnification, resulting in a lot of redundancy in the input and feature space. Moreover, the spatial relations between patches can not be sufficiently modeled, which may weaken the model's discriminative ability on fine-grained features. To solve the above limitations, we propose a Multi-scale Graph Transformer (MG-Trans) with information bottleneck for whole slide image classification. MG-Trans is composed of three modules: patch anchoring module (PAM), dynamic structure information learning module (SILM), and multi-scale information bottleneck module (MIBM). Specifically, PAM utilizes the class attention map generated from the multi-head self-attention of vision Transformer to identify and sample the informative patches. SILM explicitly introduces the local tissue structure information into the Transformer block to sufficiently model the spatial relations between patches. MIBM effectively fuses the multi-scale patch features by utilizing the principle of information bottleneck to generate a robust and compact bag-level representation. Besides, we also propose a semantic consistency loss to stabilize the training of the whole model. Extensive studies on three subtyping datasets and seven gene mutation detection datasets demonstrate the superiority of MG-Trans.

Pathological primary tumor (pT) stage focuses on the infiltration degree of the primary tumor to surrounding tissues, which relates to the prognosis and treatment choices. The pT staging relies on the field-of-views from multiple magnifications in the gigapixel images, which makes pixel-level annotation difficult. Therefore, this task is usually formulated as a weakly supervised whole slide image (WSI) classification task with the slide-level label. Existing weakly-supervised classification methods mainly follow the multiple instance learning paradigm, which takes the patches from single magnification as the instances and extracts their morphological features independently. However, they cannot progressively represent the contextual information from multiple magnifications, which is critical for pT staging. Therefore, we propose a structure-aware hierarchical graph-based multi-instance learning framework (SGMF) inspired by the diagnostic process of pathologists. Specifically, a novel graph-based instance organization method is proposed, namely structure-aware hierarchical graph (SAHG), to represent the WSI. Based on that, we design a novel hierarchical attention-based graph representation (HAGR) network to capture the critical patterns for pT staging by learning cross-scale spatial features. Finally, the top nodes of SAHG are aggregated by a global attention layer for bag-level representation. Extensive studies on three large-scale multi-center pT staging datasets with two different cancer types demonstrate the effectiveness of SGMF, which outperforms state-of-the-art up to 5.6% in the F1 score.

Modal regression, a widely used regression protocol, has been extensively investigated in statistical and machine learning communities due to its robustness to outlier and heavy-tailed noises. Understanding modal regression's theoretical behavior can be fundamental in learning theory. Despite significant progress in characterizing its statistical property, the majority results are based on the assumption that samples are independent and identical distributed (i.i.d.), which is too restrictive for real-world applications. This paper concerns about the statistical property of regularized modal regression (RMR) within an important dependence structure - Markov dependent. Specifically, we establish the upper bound for RMR estimator under moderate conditions and give an explicit learning rate. Our results show that the Markov dependence impacts on the generalization error in the way that sample size would be discounted by a multiplicative factor depending on the spectral gap of the underlying Markov chain. This result shed a new light on characterizing the theoretical underpinning for robust regression.

Recently, the scheme of model-X knockoffs was proposed as a promising solution to address controlled feature selection under high-dimensional finite-sample settings. However, the procedure of model-X knockoffs depends heavily on the coefficient-based feature importance and only concerns the control of false discovery rate (FDR). To further improve its adaptivity and flexibility, in this paper, we propose an error-based knockoff inference method by integrating the knockoff features, the error-based feature importance statistics, and the stepdown procedure together. The proposed inference procedure does not require specifying a regression model and can handle feature selection with theoretical guarantees on controlling false discovery proportion (FDP), FDR, or k-familywise error rate (k-FWER). Empirical evaluations demonstrate the competitive performance of our approach on both simulated and real data.

Computational Pathology (CPATH) offers the possibility for highly accurate and low-cost automated pathological diagnosis. However, the high time cost of model inference is one of the main issues limiting the application of CPATH methods. Due to the large size of Whole-Slide Image (WSI), commonly used CPATH methods divided a WSI into a large number of image patches at relatively high magnification, then predicted each image patch individually, which is time-consuming. In this paper, we propose a novel Uncertainty-based Model Acceleration (UMA) method for reducing the time cost of model inference, thereby relieving the deployment burden of CPATH applications. Enlightened by the slide-viewing process of pathologists, only a few high-uncertain regions are regarded as “suspicious” regions that need to be predicted at high magnification, and most of the regions in WSI are predicted at low magnification, thereby reducing the times of image patch extraction and prediction. Meanwhile, uncertainty estimation ensures prediction accuracy at low magnification. We take two fundamental CPATH classification tasks (i.e., cancer region detection and subtyping) as examples. Extensive experiments on two large-scale renal cell carcinoma classification datasets demonstrate that our UMA can significantly reduce the time cost of model inference while maintaining competitive classification performance.

Coreference resolution aims at linking all mentions that refer to the same entity, which are widely adopted in many biomedical and bioinformatics tasks, such as biomedical knowledge graph construction and metabolic pathway integration. Many recent studies focus on improving neural model structures. However, we argue that a practical method that integrates commonsense knowledge can further improve coreference resolution performance, because commonsense delivers extra prior knowledge for reasoning and can enhance related representations, rather than naive mention-context occurrence modeling. In this work, we propose an effective method to integrate external commonsense knowledge into a neural coreference resolution model. Specially, a gated attention mechanism is employed in our method to leverage commonsense according to different contexts. By using ConceptNet as the knowledge base in three span-ranking backbone models, the models can yield significant performance gains on used datasets. We also achieve improvements in tasks of long-term mention detection and cross-sentence coreferences after incorporating knowledge.

Biomedical relation extraction seeks to automatically extract biomedical relations from biomedical text, which plays an important role in biomedical studies. However, constructing high-quality biomedical annotation data is not only time-consuming but also requires a high level of knowledge in the biomedical field. To alleviate this problem, Semi-supervised Biomedical Relation Extraction aims to extract relation facts from the limited labeled data and the more readily available unlabeled samples. Existing works can be roughly categorized as self-training methods and self-ensembling methods. The former aims to generate pseudo labels, which may lead to the gradual drift problem. The latter aims to encourage the output of one model to be consistent with the other model, where the acquisition of the model is tedious. To alleviate these issues, we propose a novel Uncertainty-Guided Mutual Consistency Training framework(UG-MCT) for semi-supervised Biomedical relation extraction. Specifically, our framework consists of two models with the same structure, which differ only when updating their weights, and then an intersecting pseudo-label mechanism is designed to convert the prediction discrepancies of the two models into mutual consistency training loss, thus promoting the consistency of model predictions. In addition, we utilize uncertainty as guided information to assist the model in focusing on the confident pseudo labels and mitigate the noise of inaccurate pseudo labeling during training. Thus, our model is very simple and efficient while mitigating the noise introduced by pseudo-labels. UG-MCT is evaluated on multiple datasets in different settings and the experimental results demonstrate that our method is highly effective in semi-supervised biomedical relation extraction compared to the state-of-the-art.

Predicting drug combinations according to patients' electronic health records is an essential task in intelligent healthcare systems, which can assist clinicians in ordering safe and effective prescriptions. However, existing work either missed/underutilized the important information lying in the drug molecule structure in drug encoding or has insufficient control over Drug-Drug Interactions (DDIs) rates within the predictions. To address these limitations, we propose CSEDrug, which enhances the drug encoding and DDIs controlling by leveraging multi-faceted drug knowledge, including molecule structures of drugs, Synergistic DDIs (SDDIs), and Antagonistic DDIs (ADDIs). We integrate these types of knowledge into CSEDrug by a graph-based drug encoder and multiple loss functions, including a novel triplet learning loss and a comprehensive DDI controllable loss. We evaluate the performance of CSEDrug in terms of accuracy, effectiveness, and safety on the public MIMIC-III dataset. The experimental results demonstrate that CSEDrug outperforms several state-of-the-art methods and achieves a 2.93% and a 2.77% increase in the Jaccard similarity scores and F1 scores, meanwhile, a 0.68% reduction of the ADDI rate (safer drug combinations), and 0.69% improvement of the SDDI rate (more effective drug combinations).

Distance metric learning has become a popular solution for few-shot Named Entity Recognition (NER). The typical setup aims to learn a similarity metric for measuring the semantic similarity between test samples and referents, where each referent represents an entity class. The effect of this setup may, however, be compromised for two reasons. First, there is typically a limited optimization exerted on the representations of entity tokens after initing by pre-trained language models. Second, the referents may be far from representing corresponding entity classes due to the label scarcity in the few-shot setting. To address these challenges, we propose a novel approach named COntrastive learning with Prompt guiding for few-shot NER (COPNER). We introduce a novel prompt composed of class-specific words to COPNER to serve as 1) supervision signals for conducting contrastive learning to optimize token representations; 2) metric referents for distance-metric inference on test samples. Experimental results demonstrate that COPNER outperforms state-of-the-art models with a significant margin in most cases. Moreover, COPNER shows great potential in the zero-shot setting.

Extracting knowledge from heterogeneous data sources is fundamental for the construction of structured biomedical knowledge graphs (BKGs), where entities and relations are represented as nodes and edges in the graphs, respectively. Previous biomedical knowledge extraction methods simply considered limited entity types and relations by using a task-specific training set, which is insufficient for large-scale BKGs development and downstream task applications in different scenarios. To alleviate this issue, we propose a joint continual learning biomedical information extraction (JCBIE) network to extract entities and relations from different biomedical information datasets. By empirically studying different joint learning and continual learning strategies, the proposed JCBIE can learn and expand different types of entities and relations from different datasets. JCBIE uses two separated encoders in joint-feature extraction, hence can effectively avoid the feature confusion problem comparing with using one hard-parameter sharing encoder. Specifically, it allows us to adopt entity augmented inputs to establish the interaction between named entity recognition and relation extraction. Finally, a novel evaluation mechanism is proposed for measuring cross-corpus generalization errors, which was ignored by traditional evaluation methods. Our empirical studies show that JCBIE achieves promising performance when continual learning strategy is adopted with multiple corpora.

Tissue segmentation is an essential task in computational pathology. However, relevant datasets for such a pixel-level classification task are hard to obtain due to the difficulty of annotation, bringing obstacles for training a deep learning-based segmentation model. Recently, contrastive learning has provided a feasible solution for mitigating the heavy reliance of deep learning models on annotation. Nevertheless, applying contrastive loss to the most abstract image representations, existing contrastive learning frameworks focus on global features, therefore, are less capable of encoding finer-grained features (e.g., pixel-level discrimination) for the tissue segmentation task. Enlightened by domain knowledge, we design three contrastive learning tasks with multi-granularity views (from global to local) for encoding necessary features into representations without accessing annotations. Specifically, we construct: (1) an image-level task to capture the difference between tissue components, i.e., encoding the component discrimination; (2) a superpixel-level task to learn discriminative representations of local regions with different tissue components, i.e., encoding the prototype discrimination; (3) a pixel-level task to encourage similar representations of different tissue components within a local region, i.e., encoding the spatial smoothness. Through our global-to-local pre-training strategy, the learned representations can reasonably capture the domain-specific and fine-grained patterns, making them easily transferable to various tissue segmentation tasks in histopathological images. We conduct extensive experiments on two tissue segmentation datasets, while considering two real-world scenarios with limited or sparse annotations. The experimental results demonstrate that our framework is superior to existing contrastive learning methods and can be easily combined with weakly supervised and semi-supervised segmentation methods.

Healthcare Representation learning has been a key element to achieving state-of-the-art performance on healthcare prediction. Recent advances based Electronic Healthcare Records(EHRs) are mostly devoted to extracting temporal progression patterns with temporal model and their variants. Although these works have shown excellent performances in healthcare prediction, the unified temporal pattern may not be suitable for individuals in all healthcare conditions. Moreover, some studies ususally introduce complex Deep Neural Networks models and medical prior knowledge to get compact representation, causing great computational burden. In this paper, we propose a general health care representation model, named AEFNet. We only leverage three simple convolution operations and a set of up and down sampling to ensure performance and model complexity equally, which achieves adaptively extract distinct individual key feature in a light manner. AEFNet can shrink and refine highly suitable scale information adaptively and comletely. Breaking traditional fixed convolution scale or multi-scale, AEFNet achieves scale adaptively to extract the most significant information and context relationship. Finally, We validate our method on the public dataset MIMIC-III, and the evaluation results indicate that our method can significantly outperform other remarkable baseline models.

Pathological diagnosis is the gold standard for cancer diagnosis, but it is labor-intensive, in which tasks such as cell detection, classification, a nd c ounting a re particularly prominent. A common solution for automating these tasks is using nucleus segmentation technology. However, it is hard to train a robust nucleus segmentation model, due to several challenging problems,i.e., the nucleus adhesion, stacking, and excessive fusion with the background. Recently, some researchers proposed a series of automatic nucleus segmentation methods based on point annotation, which can significant i mprove t he m odel performance. Nevertheless, the point annotation needs to be marked by experienced pathologists. In order to take advantage of segmentation methods based on point annotation, further alleviate the manual workload, and make cancer diagnosis more efficient and accurate, it is necessary to develop an automatic nucleus detection algorithm, which can automatically and efficiently l ocate the position of the nucleus in the pathological image and extract valuable information for pathologists. In this paper, we propose a W-shaped network for automatic nucleus detection. Different from the traditional U-Net based method, mapping the original pathology image to the target mask directly, our proposed method split the detection task into two sub-tasks. The first sub-task maps the original pathology image to the binary mask, then the binary mask is mapped to the density mask in the second subtask. After the task is split, the task's difficulty i s significantly reduced, and the network's overall performance is improved. Our proposed network can automatic identify the center of each nucleus. Combined with the NuClick, a semi-supervised recognition model based on point annotation, we implement a fully automatic nucleus annotation framework.

Nuclei segmentation is a fundamental task in digital pathology analysis and can be automated by deep learning-based methods. However, the development of such an automated method requires a large amount of data with precisely annotated masks which is hard to obtain. Training with weakly labeled data is a popular solution for reducing the workload of annotation. In this paper, we propose a novel meta-learning-based nuclei segmentation method which follows the label correction paradigm to leverage data with noisy masks. Specifically, we design a fully conventional meta-model that can correct noisy masks using a small amount of clean meta-data. Then the corrected masks can be used to supervise the training of the segmentation model. Meanwhile, a bi-level optimization method is adopted to alternately update the parameters of the main segmentation model and the meta-model in an end-to-end way. Extensive experimental results on two nuclear segmentation datasets show that our method achieves the state-of-the-art result. It even achieves comparable performance with the model training on supervised data in some noisy settings.

Digital pathology plays a crucial role in the development of artificial intelligence in the medical field. The digital pathology platform can make the pathological resources digital and networked, and realize the permanent storage of visual data and the synchronous browsing processing without the limitation of time and space. It has been widely used in various fields of pathology. However, there is still a lack of an open and universal digital pathology platform to assist doctors in the management and analysis of digital pathological sections, as well as the management and structured description of relevant patient information. Most platforms cannot integrate image viewing, annotation and analysis, and text information management. To solve the above problems, we propose a comprehensive and extensible platform, PIMIP (Pathology Information Management & Integration Platform). PIMIP has developed the image annotation functions based on the visualization of digital pathological sections. Our annotation functions support multi-user collaborative annotation and multi-device annotation, and realize the automation of some annotation tasks. In the annotation task, we invited a professional pathologist for guidance. We introduce a machine learning module for image analysis. The data we collected included public data from local hospitals and clinical examples. Our platform is more clinical and suitable for clinical use. In addition to image data, we also structured the management and display of text information. So our platform is comprehensive. The platform framework is built in a modular way to support users to add machine learning modules independently, which makes our platform extensible.

Diagnostic pathology, which is the basis and gold standard of cancer diagnosis, provides essential information on the prognosis of the disease and vital evidence for clinical treatment. However, pathological diagnosis is subjective, and differences in observation and diagnosis between pathologists are common. This phenomenon is more evident in hospitals with insufficient medical resources. Deep learning (DL) can be used to identify and classify structures in digital pathology. In order to solve the above difficulties, in this work, we propose a DL framework for generating pathological diagnosis by analyzing histopathological images of renal cell carcinoma. A deep neural network is trained on a large high-quality annotated dataset for accurate tumor area detection, subtyping, and grading. The results show that our framework has achieved pathologist-level accuracy in diagnosis, can generate pathology reports with tumor indicators, and provide pathologists with interpretable auxiliary diagnoses

Constructing large-scaled medical knowledge graphs (MKGs) can significantly boost healthcare applications for medical surveillance, bring much attention from recent research. An essential step in constructing large-scale MKG is extracting information from medical reports. Recently, information extraction techniques have been proposed and show promising performance in biomedical information extraction. However, these methods only consider limited types of entity and relation due to the noisy biomedical text data with complex entity correlations. Thus, they fail to provide enough information for constructing MKGs and restrict the downstream applications. To address this issue, we propose Biomedical Information Extraction (BioIE), a hybrid neural network to extract relations from biomedical text and unstructured medical reports. Our model utilizes a multi-head attention enhanced graph convolutional network (GCN) to capture the complex relations and context information while resisting the noise from the data. We evaluate our model on two major biomedical relationship extraction tasks, chemical-disease relation (CDR) and chemical-protein interaction (CPI), and a cross-hospital pan-cancer pathology report corpus. The results show that our method achieves superior performance than baselines. Furthermore, we evaluate the applicability of our method under a transfer learning setting and show that BioIE achieves promising performance in processing medical text from different formats and writing styles.

Personalized diagnoses have not been possible due to a sear amount of data pathologists have to bear during the day-to-day routine, leading to the current generalized standards being continuously updated as new findings are reported. It is noticeable that these practical standards are developed based on multi-source heterogeneous data, including whole-slide images and pathology and clinical reports. In this study, we propose a framework that combines pathological images and medical reports to generate a personalized diagnosis result for an individual patient. We use nuclei-level image feature similarity and content-based deep learning method to search for a personalized group of populations with similar pathological characteristics, extract structured prognostic information from descriptive pathology reports of the similar patient population, and assign importance of different prognostic factors to generate a personalized pathological diagnosis result. We use multi-source heterogeneous data from TCGA (The Cancer Genome Atlas) database. The result demonstrates that our framework matches the performance of pathologists in the diagnosis of renal cell carcinoma. This framework is designed to be generic, and this could be applied to other types of cancer. The weights could provide insights into the known prognostic factors and further guide more precise clinical treatment protocols.

Subsampling is a widely used and effective method to deal with the challenges brought by big data. Most subsampling procedures are designed based on the importance sampling framework, where samples with high importance measures are given corresponding sampling probabilities. However, in the highly noisy case, these samples may cause an unstable estimator which could lead to a misleading result. To tackle this issue, we propose a gradient-based Markov subsampling (GMS) algorithm to achieve robust estimation. The core idea is to construct a subset which allows us to conservatively correct a crude initial estimate towards the true signal. Specifically, GMS selects samples with small gradients via a probabilistic procedure, constructing a subset that is likely to exclude noisy samples and provide a safe improvement over the initial estimate. We show that the GMS estimator is statistically consistent at a rate which matches the optimal in the minimax sense. The promising performance of GMS is supported by simulation studies and real data examples.

Additive models have attracted much attention for high-dimensional regression estimation and variable selection. However, the existing models are usually limited to the single-task learning framework under the mean squared error (MSE) criterion, where the utilization of variable structure depends heavily on priori knowledge among variables. For high-dimensional observations in real environment, e.g., Coronal Mass Ejections (CMEs) data, the learning performance of previous methods may be degraded seriously due to the complex non-Gaussian noise and the insufficiency of prior knowledge on variable structure. To tackle this problem, we propose a new class of additive models, called Multi-task Additive Models (MAM), by integrating the mode-induced metric, the structure-based regularizer, and additive hypothesis spaces into a bilevel optimization framework. Our approach does not require any priori knowledge of variable structure and suits for high-dimensional data with complex noise, e.g., skewed noise, heavy-tailed noise, and outliers. A smooth iterative optimization algorithm with convergence guarantees is provided to implement MAM efficiently. Experiments on simulations and the CMEs analysis demonstrate the competitive performance of our approach for robust estimation and automatic structure discovery.

The PU learning problem concerns about learning from positive and unlabeled data. A popular heuristic is to iteratively enlarge training set based on some margin-based criterion. However, little theoretical analysis has been conducted to support the success of these heuristic methods. In this work, we show that not all margin-based heuristic rules are able to improve the learned classifiers iteratively. We find that a so-called large positive margin oracle is necessary to guarantee the success of PU learning. Under this oracle, a provable positive-margin based PU learning algorithm is proposed for linear regression and classification under the truncated Gaussian distributions. The proposed algorithm is able to reduce the recovering error geometrically proportional to the positive margin. Extensive experiments on real-world datasets verify our theory and the state-of-the-art performance of the proposed PU learning algorithm.

Recently, a new framework, Fredholm learning, was proposed for semisupervised learning problems based on solving a regularized Fredholm integral equation. It allows a natural way to incorporate unlabeled data into learning algorithms to improve their prediction performance. Despite rapid progress on implementable algorithms with theoretical guarantees, the generalization ability of Fredholm kernel learning has not been studied. In this letter, we focus on investigating the generalization performance of a family of classification algorithms, referred to as Fredholm kernel regularized classifiers. We prove that the corresponding learning rate can achieve [Formula: see text] ([Formula: see text] is the number of labeled samples) in a limiting case. In addition, a representer theorem is provided for the proposed regularized scheme, which underlies its applications.

Learning with l1 -regularizer has brought about a great deal of research in learning theory community. Previous known results for the learning with l1 -regularizer are based on the assumption that samples are independent and identically distributed (i.i.d.), and the best obtained learning rate for the l1 -regularization type algorithms is O(1/√m) , where m is the samples size. This paper goes beyond the classic i.i.d. framework and investigates the generalization performance of least square regression with l1 -regularizer ( l1 -LSR) based on uniformly ergodic Markov chain (u.e.M.c) samples. On the theoretical side, we prove that the learning rate of l1 -LSR for u.e.M.c samples l1 -LSR(M) is with the order of O(1/m) , which is faster than O(1/√m) for the i.i.d. counterpart. On the practical side, we propose an algorithm based on resampling scheme to generate u.e.M.c samples. We show that the proposed l1 -LSR(M) improves on the l1 -LSR(i.i.d.) in generalization error at the low cost of u.e.M.c resampling.

Recently, much attention has been drawn to the problem of matrix completion, which arises in a number of fields, including computer vision, pattern recognition, sensor network, and recommendation systems. This paper proposes a novel algorithm, named robust alternative minimization (RAM), which is based on the constraint of low rank to complete an unknown matrix. The proposed RAM algorithm can effectively reduce the relative reconstruction error of the recovered matrix. It is numerically easier to minimize the objective function and more stable for large-scale matrix completion compared with other existing methods. It is robust and efficient for low-rank matrix completion, and the convergence of the RAM algorithm is also established. Numerical results showed that both the recovery accuracy and running time of the RAM algorithm are competitive with other reported methods. Moreover, the applications of the RAM algorithm to low-rank image recovery demonstrated that it achieves satisfactory performance.