Haiping Lu

Prasun C Tripathi, Sina Tabakhi, Mohammod N I Suvon et al.

Pulmonary Arterial Wedge Pressure (PAWP) is an essential cardiovascular hemodynamics marker to detect heart failure. In clinical practice, Right Heart Catheterization is considered a gold standard for assessing cardiac hemodynamics while non-invasive methods are often needed to screen high-risk patients from a large population. In this paper, we propose a multimodal learning pipeline to predict PAWP marker. We utilize complementary information from Cardiac Magnetic Resonance Imaging (CMR) scans (short-axis and four-chamber) and Electronic Health Records (EHRs). We extract spatio-temporal features from CMR scans using tensor-based learning. We propose a graph attention network to select important EHR features for prediction, where we model subjects as graph nodes and feature relationships as graph edges using the attention mechanism. We design four feature fusion strategies: early, intermediate, late, and hybrid fusion. With a linear classifier and linear fusion strategies, our pipeline is interpretable. We validate our pipeline on a large dataset of $2,641$ subjects from our ASPIRE registry. The comparative study against state-of-the-art methods confirms the superiority of our pipeline. The decision curve analysis further validates that our pipeline can be applied to screen a large population. The code is available at https://github.com/prasunc/hemodynamics.

Wenrui Fan, Mohammod N. I. Suvon, Shuo Zhou et al.

Pathology and anatomy are two essential groups of semantics in medical data. Pathology describes what the diseases are, while anatomy explains where the diseases occur. They describe diseases from different perspectives, providing complementary insights into diseases. Thus, properly understanding these semantics and their relationships can enhance medical vision-language models (VLMs). However, pathology and anatomy semantics are usually entangled in medical data, hindering VLMs from explicitly modeling these semantics and their relationships. To address this challenge, we propose MeDSLIP, a novel Medical Dual-Stream Language-Image Pre-training pipeline, to disentangle pathology and anatomy semantics and model the relationships between them. We introduce a dual-stream mechanism in MeDSLIP to explicitly disentangle medical semantics into pathology-relevant and anatomy-relevant streams and align visual and textual information within each stream. Furthermore, we propose an interaction modeling module with prototypical contrastive learning loss and intra-image contrastive learning loss to regularize the relationships between pathology and anatomy semantics. We apply MeDSLIP to chest X-ray analysis and conduct comprehensive evaluations with four benchmark datasets: NIH CXR14, RSNA Pneumonia, SIIM-ACR Pneumothorax, and COVIDx CXR-4. The results demonstrate MeDSLIP's superior generalizability and transferability across different scenarios. The code is available at https://github.com/Shef-AIRE/MeDSLIP, and the pre-trained model is released at https://huggingface.co/pykale/MeDSLIP.

Mohammod N. I. Suvon, Shuo Zhou, Prasun C. Tripathi et al.

Recent advancements in early assessment of pulmonary hypertension (PH) primarily focus on applying machine learning methods to centralized diagnostic modalities, such as 12-lead electrocardiogram (12L-ECG). Despite their potential, these approaches fall short in decentralized clinical settings, e.g., point-of-care and general practice, where handheld 6-lead ECG (6L-ECG) can offer an alternative but is limited by the scarcity of labeled data for developing reliable models. To address this, we propose a lead-specific electrocardiogram multimodal variational autoencoder (\textsc{LS-EMVAE}), which incorporates a hierarchical modality expert (HiME) fusion mechanism and a latent representation alignment loss. HiME combines mixture-of-experts and product-of-experts to enable flexible, adaptive latent fusion, while the alignment loss improves coherence among lead-specific and shared representations. To alleviate data scarcity and enhance representation learning, we adopt a transfer learning strategy: the model is first pre-trained on a large unlabeled 12L-ECG dataset and then fine-tuned on smaller task-specific labeled 6L-ECG datasets. We validate \textsc{LS-EMVAE} across two retrospective cohorts in a 6L-ECG setting: 892 subjects from the ASPIRE registry for (1) PH detection and (2) phenotyping pre-/post-capillary PH, and 16,416 subjects from UK Biobank for (3) predicting elevated pulmonary atrial wedge pressure, where it consistently outperforms unimodal and multimodal baseline methods and demonstrates strong generalizability and interpretability. The code is available at https://github.com/Shef-AIRE/LS-EMVAE.

Mohammod N. I. Suvon, Prasun C. Tripathi, Wenrui Fan et al.

Recent advancements in non-invasive detection of cardiac hemodynamic instability (CHDI) primarily focus on applying machine learning techniques to a single data modality, e.g. cardiac magnetic resonance imaging (MRI). Despite their potential, these approaches often fall short especially when the size of labeled patient data is limited, a common challenge in the medical domain. Furthermore, only a few studies have explored multimodal methods to study CHDI, which mostly rely on costly modalities such as cardiac MRI and echocardiogram. In response to these limitations, we propose a novel multimodal variational autoencoder ($\text{CardioVAE}_\text{X,G}$) to integrate low-cost chest X-ray (CXR) and electrocardiogram (ECG) modalities with pre-training on a large unlabeled dataset. Specifically, $\text{CardioVAE}_\text{X,G}$ introduces a novel tri-stream pre-training strategy to learn both shared and modality-specific features, thus enabling fine-tuning with both unimodal and multimodal datasets. We pre-train $\text{CardioVAE}_\text{X,G}$ on a large, unlabeled dataset of $50,982$ subjects from a subset of MIMIC database and then fine-tune the pre-trained model on a labeled dataset of $795$ subjects from the ASPIRE registry. Comprehensive evaluations against existing methods show that $\text{CardioVAE}_\text{X,G}$ offers promising performance (AUROC $=0.79$ and Accuracy $=0.77$), representing a significant step forward in non-invasive prediction of CHDI. Our model also excels in producing fine interpretations of predictions directly associated with clinical features, thereby supporting clinical decision-making.

Peizhen Bai, Filip Miljković, Xianyuan Liu et al.

Inverse protein folding generates valid amino acid sequences that can fold into a desired protein structure, with recent deep-learning advances showing strong potential and competitive performance. However, challenges remain, such as predicting elements with high structural uncertainty, including disordered regions. To tackle such low-confidence residue prediction, we propose a Mask-prior-guided denoising Diffusion (MapDiff) framework that accurately captures both structural information and residue interactions for inverse protein folding. MapDiff is a discrete diffusion probabilistic model that iteratively generates amino acid sequences with reduced noise, conditioned on a given protein backbone. To incorporate structural information and residue interactions, we develop a graph-based denoising network with a mask-prior pre-training strategy. Moreover, in the generative process, we combine the denoising diffusion implicit model with Monte-Carlo dropout to reduce uncertainty. Evaluation on four challenging sequence design benchmarks shows that MapDiff substantially outperforms state-of-the-art methods. Furthermore, the in silico sequences generated by MapDiff closely resemble the physico-chemical and structural characteristics of native proteins across different protein families and architectures.

Xiaokun Liu, Sayedmohammadreza Rastegari, Yijun Huang et al.

In cancer therapeutics, protein-metal binding mechanisms critically govern the pharmacokinetics and targeting efficacy of drugs, thereby fundamentally shaping the rational design of anticancer metallodrugs. While conventional laboratory methods used to study such mechanisms are often costly, low throughput, and limited in capturing dynamic biological processes, machine learning (ML) has emerged as a promising alternative. Despite increasing efforts to develop protein-metal binding datasets and ML algorithms, the application of ML in tumor protein-metal binding remains limited. Key challenges include a shortage of high-quality, tumor-specific datasets, insufficient consideration of multiple data modalities, and the complexity of interpreting results due to the ''black box'' nature of complex ML models. This paper summarizes recent progress and ongoing challenges in using ML to predict tumor protein-metal binding, focusing on data, modeling, and interpretability. We present multimodal protein-metal binding datasets and outline strategies for acquiring, curating, and preprocessing them for training ML models. Moreover, we explore the complementary value provided by different data modalities and examine methods for their integration. We also review approaches for improving model interpretability to support more trustworthy decisions in cancer research. Finally, we offer our perspective on research opportunities and propose strategies to address the scarcity of tumor protein data and the limited number of predictive models for tumor protein-metal binding. We also highlight two promising directions for effective metal-based drug design: integrating protein-protein interaction data to provide structural insights into metal-binding events and predicting structural changes in tumor proteins after metal binding.

Hong Lu, Yali Bian, Rahul C. Shah

High-quality annotations are essential for object detection models, but ensuring label accuracy - especially for bounding boxes - remains both challenging and costly. This paper introduces ClipGrader, a novel approach that leverages vision-language models to automatically assess the accuracy of bounding box annotations. By adapting CLIP (Contrastive Language-Image Pre-training) to evaluate both class label correctness and spatial precision of bounding box, ClipGrader offers an effective solution for grading object detection labels. Tested on modified object detection datasets with artificially disturbed bounding boxes, ClipGrader achieves 91% accuracy on COCO with a 1.8% false positive rate. Moreover, it maintains 87% accuracy with a 2.1% false positive rate when trained on just 10% of the COCO data. ClipGrader also scales effectively to larger datasets such as LVIS, achieving 79% accuracy across 1,203 classes. Our experiments demonstrate ClipGrader's ability to identify errors in existing COCO annotations, highlighting its potential for dataset refinement. When integrated into a semi-supervised object detection (SSOD) model, ClipGrader readily improves the pseudo label quality, helping achieve higher mAP (mean Average Precision) throughout the training process. ClipGrader thus provides a scalable AI-assisted tool for enhancing annotation quality control and verifying annotations in large-scale object detection datasets.

Wenrui Fan, L. M. Riza Rizky, Jiayang Zhang et al.

Neuropathic pain, affecting up to 10% of adults, remains difficult to treat due to limited therapeutic efficacy and tolerability. Although resting-state functional MRI (rs-fMRI) is a promising non-invasive measurement of brain biomarkers to predict drug response in therapeutic development, the complexity of fMRI demands machine learning models with substantial capacity. However, extreme data scarcity in neuropathic pain research limits the application of high-capacity models. To address the challenge of data scarcity, we propose FMM$_{TC}$, a Foundation-Model-boosted Multimodal learning framework for fMRI-based neuropathic pain drug response prediction, which leverages both internal multimodal information in pain-specific data and external knowledge from large pain-agnostic data. Specifically, to maximize the value of limited pain-specific data, FMM$_{TC}$ integrates complementary information from two rs-fMRI modalities: Time series and functional Connectivity. FMM$_{TC}$ is further boosted by an fMRI foundation model with its external knowledge from extensive pain-agnostic fMRI datasets enriching limited pain-specific information. Evaluations with an in-house dataset and a public dataset from OpenNeuro demonstrate FMM$_{TC}$'s superior representation ability, generalizability, and cross-dataset adaptability over existing unimodal fMRI models that only consider one of the rs-fMRI modalities. The ablation study validates the effectiveness of multimodal learning and foundation-model-powered external knowledge transfer in FMM$_{TC}$. An integrated gradient-based interpretation study explains how FMM$_{TC}$'s cross-dataset dynamic behaviors enhance its adaptability. In conclusion, FMM$_{TC}$ boosts clinical trials in neuropathic pain therapeutic development by accurately predicting drug responses to improve the participant stratification efficiency.

Shuo Zhou, Junhao Luo, Yaya Jiang et al.

Lateralization is a fundamental feature of the human brain, where sex differences have been observed. Conventional studies in neuroscience on sex-specific lateralization are typically conducted on univariate statistical comparisons between male and female groups. However, these analyses often lack effective validation of group specificity. Here, we formulate modeling sex differences in lateralization of functional networks as a dual-classification problem, consisting of first-order classification for left vs. right functional networks and second-order classification for male vs. female models. To capture sex-specific patterns, we develop the Group-Specific Discriminant Analysis (GSDA) for first-order classification. The evaluation on two public neuroimaging datasets demonstrates the efficacy of GSDA in learning sex-specific models from functional networks, achieving a significant improvement in group specificity over baseline methods. The major sex differences are in the strength of lateralization and the interactions within and between lobes. The GSDA-based method is generic in nature and can be adapted to other group-specific analyses such as handedness-specific or disease-specific analyses.

Raivo Koot, Markus Hennerbichler, Haiping Lu

In video action recognition, transformers consistently reach state-of-the-art accuracy. However, many models are too heavyweight for the average researcher with limited hardware resources. In this work, we explore the limitations of video transformers for lightweight action recognition. We benchmark 13 video transformers and baselines across 3 large-scale datasets and 10 hardware devices. Our study is the first to evaluate the efficiency of action recognition models in depth across multiple devices and train a wide range of video transformers under the same conditions. We categorize current methods into three classes and show that composite transformers that augment convolutional backbones are best at lightweight action recognition, despite lacking accuracy. Meanwhile, attention-only models need more motion modeling capabilities and stand-alone attention block models currently incur too much latency overhead. Our experiments conclude that current video transformers are not yet capable of lightweight action recognition on par with traditional convolutional baselines, and that the previously mentioned shortcomings need to be addressed to bridge this gap. Code to reproduce our experiments will be made publicly available.

Xianyuan Liu, Shuo Zhou, Tao Lei et al.

Unsupervised Domain Adaptation (UDA) can transfer knowledge from labeled source data to unlabeled target data of the same categories. However, UDA for first-person action recognition is an under-explored problem, with lack of datasets and limited consideration of first-person video characteristics. This paper focuses on addressing this problem. Firstly, we propose two small-scale first-person video domain adaptation datasets: ADL$_{small}$ and GTEA-KITCHEN. Secondly, we introduce channel-temporal attention blocks to capture the channel-wise and temporal-wise relationships and model their inter-dependencies important to first-person vision. Finally, we propose a Channel-Temporal Attention Network (CTAN) to integrate these blocks into existing architectures. CTAN outperforms baselines on the two proposed datasets and one existing dataset EPIC$_{cvpr20}$.

Raivo Koot, Haiping Lu

Efficient video action recognition remains a challenging problem. One large model after another takes the place of the state-of-the-art on the Kinetics dataset, but real-world efficiency evaluations are often lacking. In this work, we fill this gap and investigate the use of transformers for efficient action recognition. We propose a novel, lightweight action recognition architecture, VideoLightFormer. In a factorized fashion, we carefully extend the 2D convolutional Temporal Segment Network with transformers, while maintaining spatial and temporal video structure throughout the entire model. Existing methods often resort to one of the two extremes, where they either apply huge transformers to video features, or minimal transformers on highly pooled video features. Our method differs from them by keeping the transformer models small, but leveraging full spatiotemporal feature structure. We evaluate VideoLightFormer in a high-efficiency setting on the temporally-demanding EPIC-KITCHENS-100 and Something-Something-V2 (SSV2) datasets and find that it achieves a better mix of efficiency and accuracy than existing state-of-the-art models, apart from the Temporal Shift Module on SSV2.

Xianyuan Liu, Raivo Koot, Shuo Zhou et al.

This report describes the technical details of our submission to the EPIC-Kitchens 2021 Unsupervised Domain Adaptation Challenge for Action Recognition. The EPIC-Kitchens dataset is more difficult than other video domain adaptation datasets due to multi-tasks with more modalities. Firstly, to participate in the challenge, we employ a transformer to capture the spatial information from each modality. Secondly, we employ a temporal attention module to model temporal-wise inter-dependency. Thirdly, we employ the adversarial domain adaptation network to learn the general features between labeled source and unlabeled target domain. Finally, we incorporate multiple modalities to improve the performance by a three-stream network with late fusion. Our network achieves the comparable performance with the state-of-the-art baseline T$A^3$N and outperforms the baseline on top-1 accuracy for verb class and top-5 accuracies for all three tasks which are verb, noun and action. Under the team name xy9, our submission achieved 5th place in terms of top-1 accuracy for verb class and all top-5 accuracies.

Haiping Lu, Xianyuan Liu, Robert Turner et al.

Machine learning is a general-purpose technology holding promises for many interdisciplinary research problems. However, significant barriers exist in crossing disciplinary boundaries when most machine learning tools are developed in different areas separately. We present Pykale - a Python library for knowledge-aware machine learning on graphs, images, texts, and videos to enable and accelerate interdisciplinary research. We formulate new green machine learning guidelines based on standard software engineering practices and propose a novel pipeline-based application programming interface (API). PyKale focuses on leveraging knowledge from multiple sources for accurate and interpretable prediction, thus supporting multimodal learning and transfer learning (particularly domain adaptation) with latest deep learning and dimensionality reduction models. We build PyKale on PyTorch and leverage the rich PyTorch ecosystem. Our pipeline-based API design enforces standardization and minimalism, embracing green machine learning concepts via reducing repetitions and redundancy, reusing existing resources, and recycling learning models across areas. We demonstrate its interdisciplinary nature via examples in bioinformatics, knowledge graph, image/video recognition, and medical imaging.

Yan Ge, Jun Ma, Li Zhang et al.

Higher-order proximity (HOP) is fundamental for most network embedding methods due to its significant effects on the quality of node embedding and performance on downstream network analysis tasks. Most existing HOP definitions are based on either homophily to place close and highly interconnected nodes tightly in embedding space or heterophily to place distant but structurally similar nodes together after embedding. In real-world networks, both can co-exist, and thus considering only one could limit the prediction performance and interpretability. However, there is no general and universal solution that takes both into consideration. In this paper, we propose such a simple yet powerful framework called homophily and heterophliy preserving network transformation (H2NT) to capture HOP that flexibly unifies homophily and heterophily. Specifically, H2NT utilises motif representations to transform a network into a new network with a hybrid assumption via micro-level and macro-level walk paths. H2NT can be used as an enhancer to be integrated with any existing network embedding methods without requiring any changes to latter methods. Because H2NT can sparsify networks with motif structures, it can also improve the computational efficiency of existing network embedding methods when integrated. We conduct experiments on node classification, structural role classification and motif prediction to show the superior prediction performance and computational efficiency over state-of-the-art methods. In particular, DeepWalk-based H2 NT achieves 24% improvement in terms of precision on motif prediction, while reducing 46% computational time compared to the original DeepWalk.

Li Zhang, Yan Ge, Haiping Lu

Graph Neural Networks (GNNs) are widely used in graph representation learning. However, most GNN methods are designed for either homogeneous or heterogeneous graphs. In this paper, we propose a new model, Hop-Hop Relation-aware Graph Neural Network (HHR-GNN), to unify representation learning for these two types of graphs. HHR-GNN learns a personalized receptive field for each node by leveraging knowledge graph embedding to learn relation scores between the central node's representations at different hops. In neighborhood aggregation, our model simultaneously allows for hop-aware projection and aggregation. This mechanism enables the central node to learn a hop-wise neighborhood mixing that can be applied to both homogeneous and heterogeneous graphs. Experimental results on five benchmarks show the competitive performance of our model compared to state-of-the-art GNNs, e.g., up to 13K faster in terms of time cost per training epoch on large heterogeneous graphs.

Hao Xu, Shengqi Sang, Haiping Lu

The use of drug combinations often leads to polypharmacy side effects (POSE). A recent method formulates POSE prediction as a link prediction problem on a graph of drugs and proteins, and solves it with Graph Convolutional Networks (GCNs). However, due to the complex relationships in POSE, this method has high computational cost and memory demand. This paper proposes a flexible Tri-graph Information Propagation (TIP) model that operates on three subgraphs to learn representations progressively by propagation from protein-protein graph to drug-drug graph via protein-drug graph. Experiments show that TIP improves accuracy by 7%+, time efficiency by 83$\times$, and space efficiency by 3$\times$.

Shuo Zhou, Wenwen Li, Christopher R. Cox et al.

Brain imaging data are important in brain sciences yet expensive to obtain, with big volume (i.e., large p) but small sample size (i.e., small n). To tackle this problem, transfer learning is a promising direction that leverages source data to improve performance on related, target data. Most transfer learning methods focus on minimizing data distribution mismatch. However, a big challenge in brain imaging is the large domain discrepancies in cognitive experiment designs and subject-specific structures and functions. A recent transfer learning approach minimizes domain dependence to learn common features across domains, via the Hilbert-Schmidt Independence Criterion (HSIC). Inspired by this method, we propose a new Domain Independent Support Vector Machine (DI-SVM) for transfer learning in brain condition decoding. Specifically, DI-SVM simultaneously minimizes the SVM empirical risk and the dependence on domain information via a simplified HSIC. We use public data to construct 13 transfer learning tasks in brain decoding, including three interesting multi-source transfer tasks. Experiments show that DI-SVM's superior performance over eight competing methods on these tasks, particularly an improvement of more than 24% on multi-source transfer tasks.

Yan Ge, Haiping Lu, Pan Peng

Clustering is fundamental for gaining insights from complex networks, and spectral clustering (SC) is a popular approach. Conventional SC focuses on second-order structures (e.g., edges connecting two nodes) without direct consideration of higher-order structures (e.g., triangles and cliques). This has motivated SC extensions that directly consider higher-order structures. However, both approaches are limited to considering a single order. This paper proposes a new Mixed-Order Spectral Clustering (MOSC) approach to model both second-order and third-order structures simultaneously, with two MOSC methods developed based on Graph Laplacian (GL) and Random Walks (RW). MOSC-GL combines edge and triangle adjacency matrices, with theoretical performance guarantee. MOSC-RW combines first-order and second-order random walks for a probabilistic interpretation. We automatically determine the mixing parameter based on cut criteria or triangle density, and construct new structure-aware error metrics for performance evaluation. Experiments on real-world networks show 1) the superior performance of two MOSC methods over existing SC methods, 2) the effectiveness of the mixing parameter determination strategy, and 3) insights offered by the structure-aware error metrics.

Wenwen Li, Jian Lou, Shuo Zhou et al.

While functional magnetic resonance imaging (fMRI) is important for healthcare/neuroscience applications, it is challenging to classify or interpret due to its multi-dimensional structure, high dimensionality, and small number of samples available. Recent sparse multilinear regression methods based on tensor are emerging as promising solutions for fMRI, yet existing works rely on unfolding/folding operations and a tensor rank relaxation with limited tightness. The newly proposed tensor singular value decomposition (t-SVD) sheds light on new directions. In this work, we study t-SVD for sparse multilinear regression and propose a Sparse tubal-regularized multilinear regression (Sturm) method for fMRI. Specifically, the Sturm model performs multilinear regression with two regularization terms: a tubal tensor nuclear norm based on t-SVD and a standard L1 norm. We further derive the algorithm under the alternating direction method of multipliers framework. We perform experiments on four classification problems, including both resting-state fMRI for disease diagnosis and task-based fMRI for neural decoding. The results show the superior performance of Sturm in classifying fMRI using just a small number of voxels.

Li Zhang, Heda Song, Nikolaos Aletras et al.

Graph convolutional network (GCN) is an emerging neural network approach. It learns new representation of a node by aggregating feature vectors of all neighbors in the aggregation process without considering whether the neighbors or features are useful or not. Recent methods have improved solutions by sampling a fixed size set of neighbors, or assigning different weights to different neighbors in the aggregation process, but features within a feature vector are still treated equally in the aggregation process. In this paper, we introduce a new convolution operation on regular size feature maps constructed from features of a fixed node bandwidth via sampling to get the first-level node representation, which is then passed to a standard GCN to learn the second-level node representation. Experiments show that our method outperforms competing methods in semi-supervised node classification tasks. Furthermore, our method opens new doors for exploring new GCN architectures, particularly deeper GCN models.

Qiquan Shi, Haiping Lu

Principal component analysis (PCA) is an unsupervised method for learning low-dimensional features with orthogonal projections. Multilinear PCA methods extend PCA to deal with multidimensional data (tensors) directly via tensor-to-tensor projection or tensor-to-vector projection (TVP). However, under the TVP setting, it is difficult to develop an effective multilinear PCA method with the orthogonality constraint. This paper tackles this problem by proposing a novel Semi-Orthogonal Multilinear PCA (SO-MPCA) approach. SO-MPCA learns low-dimensional features directly from tensors via TVP by imposing the orthogonality constraint in only one mode. This formulation results in more captured variance and more learned features than full orthogonality. For better generalization, we further introduce a relaxed start (RS) strategy to get SO-MPCA-RS by fixing the starting projection vectors, which increases the bias and reduces the variance of the learning model. Experiments on both face (2D) and gait (3D) data demonstrate that SO-MPCA-RS outperforms other competing algorithms on the whole, and the relaxed start strategy is also effective for other TVP-based PCA methods.