Single-cell transcriptomic foundation models have demonstrated remarkable performance in various zero-shot tasks, such as cell annotation and batch correction. However, adapting these models to downstream tasks still requires substantial amounts of data to fine-tune, which hinders the generalization and utility of these foundation models in scenarios with limited data. Some methods propose releasing this burden by merging models trained on other data sources. However, integrating different models trained on disjoint semantic representations may lead to mismatches. To address this issue, we investigate the essence of the model weights in the Fourier view, where the phase is more relevant to the task and the amplitude provides generalization. Based on this intuition, the Model Synthesizer is proposed to enhance suboptimal scRNA-seq fine-tuned models by incorporating irrelevant models. Model synthesizer employs a selective integration strategy that combines the original phase with merged amplitude from different models. This strategy could preserve task-specific knowledge and retain more generalizable knowledge. We evaluate our proposed method across both zeroshot and fine-tuning tasks, and the results demonstrate the appealing performance of Model Synthesizer.

Wearable activity recognition is crucial for ubiquitous computing, enhancing human-machine interaction, medical monitoring, and personalized services. As wearable devices collect user activity data that often contain personal privacy information, federated learning (FL) is increasingly applied to protect user data privacy. However, in real-world scenarios, users' data are commonly exhibit heterogeneity, manifesting as non-independent and identically distributed (non-IID) characteristics, which presents challenges for FL methods. Traditional FL client selection approaches with heterogeneous data can cause global model drift, reducing the accuracy of activity recognition models. In this paper, we propose Data Contribution-Based Federated Learning (DCBFL) method, an adaptive FL training approach by selecting clients to counter the problem caused by heterogeneous data. Specifically, we first utilize a conditional generator on the server to construct an auxiliary dataset, which is used to train an auxiliary model as a benchmark to measure the degree of heterogeneity in each client's data. Furthermore, we reasonably differentiate the data contributions of clients based on the degree of data heterogeneity and select suitable clients for FL training, effectively utilizing heterogeneous data information, mitigating global model drift. The comprehensive experiments are conducted on five public activity recognition datasets under non-IID conditions in this work. The experimental results show that DCBFL outperforms existing baseline methods, showcasing superior performance.

Federated learning (FL) enables multiple clients to collaboratively train a global model while keeping local data decentralized. Data heterogeneity (non-IID) across clients has imposed significant challenges to FL, which makes local models re-optimize towards their own local optima and forget the global knowledge, resulting in performance degradation and convergence slowdown. Many existing works have attempted to address the non-IID issue by adding an extra global-model-based regularizing item to the local training but without an adaption scheme, which is not efficient enough to achieve high performance with deep learning models. In this paper, we propose a Selective Self-Distillation method for Federated learning (FedSSD), which imposes adaptive constraints on the local updates by self-distilling the global model’s knowledge and selectively weighting it by evaluating the credibility at both the class and sample level. The convergence guarantee of FedSSD is theoretically analyzed and extensive experiments are conducted on three public benchmark datasets, which demonstrates that FedSSD achieves better generalization and robustness in fewer communication rounds, compared with other state-of-the-art FL methods.

Benign and malignant classification of clustered microcalcifications (MCs) in digital breast tomosynthesis (DBT) is an essential task in computer-aided diagnosis. However, due to the anisotropic resolution of DBT, three-dimensional (3-D) convolutional neural network (CNN)-based methods cannot extract hierarchical features efficiently. Moreover, the sparse distribution of MC points in the cluster makes it difficult for the CNN to extract discriminative structural information for classification. To comprehensively address these challenges, we propose a novel structure-aware hierarchical network (SAH-Net) for benign and malignant classification of clustered MC in a DBT volume. Specifically, the two-dimensional (2-D) group convolution is used to extract intraslice features. The one-to-one correspondence between group convolutions and slices ensures the independence of hierarchical feature extraction. Then, a partial deformable Transformer-based 3-D structural feature learning module is proposed to capture the long-range dependency between MC points in the cluster. We evaluate the proposed method on an in-house dataset with 495 clustered MCs collected from 462 DBT images. Experimental results confirm the validity of our proposed modules. The results also show that the proposed SAH-Net outperforms several other representative methods on this topic, and achieves the best classification result, with an area under the receiver operation curve (AUC) of 86.87%. The implementation of the proposed model is available at https://github.com/sunhaotian130911/SAHNet.

Federated learning (FL) has been an effective way to train a machine learning model distributedly, holding local data without exchanging them. However, due to the inaccessibility of local data, FL with label noise would be more challenging. Most existing methods assume only open-set or closed-set noise and correspondingly propose filtering or correction solutions, ignoring that label noise can be mixed in real-world scenarios. In this article, we propose a novel FL method to discriminate the type of noise and make the FL mixed noise-robust, named FedMIN. FedMIN employs a composite framework that captures local-global differences in multiparticipant distributions to model generalized noise patterns. By determining adaptive thresholds for identifying mixed label noise in each client and assigning appropriate weights during model aggregation, FedMIN enhances the performance of the global model. Furthermore, FedMIN incorporates a loss alignment mechanism using local and global Gaussian mixture models (GMMs) to mitigate the risk of revealing samplewise loss. Extensive experiments are conducted on several public datasets, which include the simulated FL testbeds, i.e., CIFAR-10, CIFAR-100, and SVHN, and the real-world ones, i.e., Camelyon17 and multiorgan nuclei challenge (MoNuSAC). Compared to FL benchmarks, FedMIN improves model accuracy by up to 9.9% due to its superior noise estimation capabilities.

Meta-data, e.g., gender, medical history, etc., is of vital importance for dermatosis diagnosis clinically, while most existing AI-based methods ignore this information and only employ dermoscopy images. Due to the modality disequilibrium of image and meta-data, there would be unequal optimization during the modal fusion, resulting in the weak meta-data modality being suppressed and losing their contributions to the classification. To address this issue, we propose an Adaptive Multi-modal Fusion (AMF) method for dermatosis diagnosis, which reweights the fusion factor of image and meta-data by evaluating their optimization discrepancy through the Euclidean norm. What’s more, the Co-Attention mechanism is employed to align the two modalities implicitly. To evaluate the effectiveness of the proposed method, extensive experiments are conducted on multi-modal dermatosis public datasets including ISIC 2019, PDA-UFES-20, and HAM10000. The results demonstrate that the proposed method can achieve better performance in comparison to the state-of-the-art methods.

Exploring a Digital Bio-Marker (DBM) is challenging for detecting the progress of Parkinson’s Disease (PD) since the unclear disease mechanism. Traditional clinical trials first formulate a DBM hypothesis and then proceed to verify them by control study, the process of which is often limited to clinicians’ expertise and experience. Machine Learning with big data provides an opportunity to automatically discover DBMs, while recent privacy policies, such as GDPR, have created extra obstacles for collecting patient information and conducting multicenter trials. To address this issue, we propose a novel DBM discovery paradigm with federated learning, called FedDBM, which forms a closed loop consisting of model training and post hoc explanation. FedDBM employs a federated split learning to preserve patients’ privacy in a multicenter clinical trial which attempts to build a model that maps signal data to PD progress. Then, a Federated Shapley Additive exPlanations method (Fed-SHAP) is proposed to find those features of vital importance in the well-trained model, known as DBM. The proposed FedDBM was evaluated on four PD typical motor symptoms and the extensive experimental results demonstrated that FedDBM showed comparable performance with SOTA federated learning methods, and the explored DBMs were proved to be more sensitive than current clinical metrics.

Skin lesion segmentation is a fundamental procedure in computer-aided melanoma diagnosis. However, due to the diverse shape, variable size, blurry boundary, and noise interference of lesion regions, existing methods may struggle with the challenge of inconsistency within classes and indiscrimination between classes. In view of this, we propose a novel method to learn and model inter-pixel correlations from both global and local aspects, which can increase inter-class variances and intra-class similarities. Specifically, under the encoder-decoder architecture, we first design a pyramid transformer inter-pixel correlations (PTIC) module, aiming at capturing the non-local context information of different levels and further exploring the global pixel-level relationship to deal with the large variance of shape and size. Further, we devise a local neighborhood metric learning (LNML) module to strengthen the local semantic correlations learning capability and increase the separability between classes in the feature space. These two modules can complementarily strengthen the feature representation capability via exploiting the inter-pixel semantic correlations, thus further improving intra-class consistency and inter-class variance. Comprehensive experiments are performed on public skin lesion segmentation datasets: ISIC 2018, ISIC2016, and PH2, and experimental results demonstrate that the proposed method achieves better segmentation performance than other state-of-the-art methods.

Federated learning (FL) enables multiple clients to collaboratively train a globally generalized model while keeping local data decentralized. A key challenge in FL is to handle the heterogeneity of data distributions among clients. The local model will shift the global feature when fitting local data, which results in forgetting the global knowledge. Following the idea of knowledge distillation, the global model's prediction can be utilized to help local models preserve the global knowledge in FL. However, when the global model hasn't converged completely, its predictions tend to be less reliable on certain classes, which may results in distillation's misleading of local models. In this paper, we propose a class-wise adaptive self distillation (FedCAD) mechanism to ameliorate this problem. We design class-wise adaptive terms to soften the influence of distillation loss according to the global model's performance on each class and therefore avoid the misleading. Experiments show that our method outperforms other state-of-the-art FL algorithms on benchmark datasets.

Recently deep neural networks have been applied to skin lesion recognition to learn feature representations, among which segmentation masks have been proved effective in enhancing the recognition performance by making the classifier network focus on the region of interest. However, high-quality segmentations for skin images require expert doctors and are a cumbersome task in terms of time and labor. Although some open-sourced segmentation models for similar tasks could help produce segmentation masks, employing them directly would bring unpredictable mismatches because of the concept drift effect. To tackle this issue, we propose an Uncertainty-based Fusion Network (UFN) for skin lesion diagnosis, which tries to robustly fuse a trained or open-sourced segmentation model with the target skin lesion classification task. To mitigate the negative influence of the segmentation mismatch, UFN uses the Dirichlet distribution to model the uncertainty and probabilities of multi-level representations. Then modified Dempster-Shafer theory is introduced to fuse the segmentation and classification representations adaptively. Moreover, the multi-level fused representations are unequally weighted by an attention mechanism. Furthermore, UFN integrates the label supervision for every fused representation as well as the weighted representations. The extensive experiments on two public benchmark datasets demonstrated the superiority of UFN compared with state-of-the-art methods.

Surface electromyography (sEMG) is a kind of bioelectric signal of the human body, containing a wealth of action intentions. Among various gesture recognition solutions, sEMG-based solutions show irreplaceable advantages by directly sensing and parsing human muscle activities and converting them into interactive commands. However, sEMG is sensitive to many factors related to users, and there are individual differences among different users. The gesture recognition model trained based on the data of existing users has poor recognition accuracy on the data of new users directly. Excitingly, transfer learning breaks through the independently identical distribution (I.I.D.) assumption of data in different domains, so it shows great potential for cross-user sEMG gesture recognition. Therefore, we propose a Multi-Source Integration based Transfer Learning (MSITL) method to explore cross-user gesture recognition in this paper. MSITL is composed of two main parts, the Source Model Construction Strategy (SMCS) and the Target Model Adaptation Strategy (TMAS). SMCS is a layered integration model. The first layer builds a model for each user. The second layer integrates multiple models through simple majority voting. TMAS is mainly divided into three steps. The first step is to use the target domain data to evaluate the source domain model and obtain the evaluation score of the individual classifier; The second step is to fine-tune the individual classifiers under the guidance of the evaluation scores; The third step is to integrate the adjusted model. Detailed experiments are conducted on benchmark sEMG gesture recognition datasets, including NinaPro (i.e., DB1) and CapgMyo (i.e., DB-a, DB-b, and DB-c). The proposed method achieves significant improvements in performance compared with current state-of-the-art methods.

Aiming at the effectiveness of contrast feature design, we proposed a promising novel non-reference quality assessment approach in exploring Attribute-Based representation. The method generates three perceptual attribute categories tailored to contrast. The first is semantic attribute derived from deep convolutional neural network, which implements adaptive contrast prediction relevant to scenario content. Second, for perceiving Spatial channel attribute, the global and local features generated by dark channel map through the designed dual convolution structures. Third, for statistical attribute, we assume the enhanced image as “reference” and calculate the structural similarity with pristine image, and the entropy and histogram metrics are also employed to assist learning. After that, for maximizing utilization, the features are embedded and integrated hierarchically to translate into objective score. In addition, a medium-scale contrast distortion database is established to support further research, which is more challenging than existing datasets because of the sufficient content and sophisticated changes. We demonstrate the availability of structures quantitatively and verify the rationality of hypothesis. Extensive experiments reveal that the proposed method outperforms advanced methods and achieves the state-of-the-art on the created database and CSIQ, TID2013, CCID2014.

Federated learning (FL) is a novel distributed learning framework where multiple participants collaboratively train a global model without sharing any raw data to preserve privacy. However, data quality may vary among the participants, the most typical of which is label noise. The incorrect label would significantly damage the performance of the global model. In FL, the inaccessibility of raw data makes this issue more challenging. Previously published studies are limited to using a task-specific benchmark-trained model to evaluate the relevance between the benchmark dataset in the server and the local one on the participants’ side. However, such approaches have failed to exploit the cooperative nature of FL itself and are not practical. This paper proposes a Consensus-based Label Correction approach (CLC) in FL, which tries to correct the noisy labels using the developed consensus method among the FL participants. The consensus-defined class-wise information is used to identify the noisy labels and correct them with pseudo-labels. Extensive experiments are conducted on several public datasets in various settings. The experimental results prove the advantage over the state-of-art methods. The link to the source code is https://github.com/bixiao-zeng/CLC.git .

Perioperative complications are associated with increased patient morbidity and mortality, and result in substantial healthcare resource utilization. With the aim to facilitate medical decision-making and improve health outcomes, machine learning methods are used to train prediction models to inform healthcare professionals and patients about the risks, which require both inner- and outer-hospital information, e.g., daily performance and clinical tests. For sake of data security and privacy, the Hospital Information System (HIS) is usually isolated from the public Internet and the raw patient samples are forbidden to transfer directly, which limits the integration of inner-and outer-hospital information. In this paper, we propose a learning framework named FedIO which bridges Inner- and Outer-hospital information via vertical Federated Learning for perioperative complications prognostic prediction. Instead of transmitting data into one cloud center, FedIO leverages the locally kept data to train a prediction model, during which only the intermediate parameters are transmitted and integrated. Extensive experiments are conducted on real-world datasets and the results manifest that combining inner- and outer-hospital knowledge is better than either of them, and FedIO shows the same-level performance as the cloud-based methods but without sharing raw data.

False positives (FPs) reduction is indispensable for clustered microcalcifications (MCs) detection in digital breast tomosynthesis (DBT), since there might be excessive false candidates in the detection stage. Considering that DBT volume has an anisotropic resolution, we proposed a novel 3D context-aware convolutional neural network (CNN) to reduce FPs, which consists of a 2D intra-slices feature extraction branch and a 3D inter-slice features fusion branch. In particular, 3D anisotropic convolutions were designed to learn representations from DBT volumes and inter-slice information fusion is only performed on the feature map level, which could avoid the influence of anisotropic resolution of DBT volume. The proposed method was evaluated on a large-scale Chinese women population of 877 cases with 1754 DBT volumes and compared with 8 related methods. Experimental results show that the proposed network achieved the best performance with an accuracy of 92.68% for FPs reduction with an AUC of 97.65%, and the FPs are 0.0512 per DBT volume at a sensitivity of 90%. This also proved that making full use of 3D contextual information of DBT volume can improve the performance of the classification algorithm.

Human Activity Recognition (HAR) is an important application of smart wearable/mobile systems for many human-centric problems such as healthcare. The multi-sensor synchronous measurement has shown better performance for HAR than a single sensor. However, the multi-sensor setting increases the costs of data transmission, computation and energy. Therefore, the efficient sensor selection to balance recognition accuracy and sensor cost is the critical challenge. In this paper, we propose an Instance-wise Dynamic Sensor Selection (IDSS) method for HAR. Firstly, we formalize this problem as minimizing both activity classification loss and sensor number by dynamically selecting a sparse subset for each instance. Then, IDSS solves the above minimization problem via Markov Decision Process whose policy for sensor selection is learned by exploiting the instance-wise states using Imitation Learning. In order to optimize the parameters of the activity classification model and the sensor selection policy, an algorithm named Mutual DAgger is proposed to alternatively enhance their learning process. To evaluate the performance of IDSS, we conduct experiments on three real-world HAR datasets. The experimental results show that IDSS can effectively reduce the overall sensor number without losing accuracy and outperforms the state-of-the-art methods regarding the combined measurement of accuracy and sensor number.

Discriminating pathologic cognitive decline from the expected decline of normal aging is an important research topic for elderly care and health monitoring. However, most cognitive assessment methods only work when data distributions of the training set and testing set are consistent. Enabling existing cognitive assessment models to adapt to the data in new cognitive assessment tasks is a significant challenge. In this paper, we propose a novel domain adaptation method, namely the Fine-Grained Adaptation Random Forest (FAT), to bridge the cognitive assessment gap when the data distribution is changed. FAT is composed of two essential parts 1) information gain based model evaluation strategy (IGME) and 2) domain adaptation tree growing mechanism (DATG). IGME is used to evaluate every individual tree, and DATG is used to transfer the source model to the target domain. To evaluate the performance of FAT, we conduct experiments in real clinical environments. Experimental results demonstrate that FAT is significantly more accurate and efficient compared with other state-of-the-art methods.

Surface electromyography (sEMG) armband-based gesture recognition is an active research topic that aims to identify hand gestures with a single row of sEMG electrodes. As a typical type of biological signal, sEMG on one channel is nonstationary temporally and related to multiple adjacent muscles spatially, which hinders the effective representation in gesture recognition. To tackle these aspects, we propose a spatial-temporal features-based gesture recognition method (STF-GR) in this article. Specifically, STF-GR first decomposes the nonstationary multichannel sEMG by multivariate empirical mode decomposition, which jointly transforms each channel into a series of stationary subsignals. It can keep the temporal stationarity within-channel as well as the spatial independence across-channel. Then, by the convolutional recurrent neural network, STF-GR extracts and merges spatial-temporal features of decomposed sEMG signal. Finally, a negative log-likelihood-based cost function is used to make the final gesture decision. To evaluate the performance of STF-GR, we conduct experiments on three data sets, noninvasive adaptive hand prosthetic (NinaPro), CapgMyo, and BandMyo. The first two are publicly available, and BandMyo is collected by ourselves. Experimental evaluations with within-subject tests show that STF-GR exceeds the performance of other state-of-the-art methods, including deep learning algorithms that are not focused on spatial-temporal features and traditional machine learning algorithms that use handcrafted features.

In clinical settings, the assessment of Parkinson's disease (PD) mainly depends on the doctor's experience and observation. Such assessment often lacks of unified standards and may result in varied diagnoses among different doctors. To cope with this problem, we propose an objective and quantified symptom assessment tool of PD on mobile devices, based on the Unified Parkinson's Disease Rating Scale (UPDRS). The mobile PD assessment tool, PdAssit, assesses PD symptoms by actively delivering six tasks and generates scores equivalent to UPDRS using machines learning models. PdAssit is applied in three essential applications, including medication response detection, symptom self-tracking and diagnostic assistance. The feasibility and effectiveness of PdAssit is demonstrated through clinical experiments.