Sign language is the primary communication mode for 72 million hearing-impaired individuals worldwide, necessitating effective bidirectional Sign Language Production and Sign Language Translation systems. However, functional bidirectional systems require a unified linguistic environment, hindered by the lack of suitable unified datasets, particularly those providing the necessary pose information for accurate Sign Language Production (SLP) evaluation. Concurrently, current SLP evaluation methods like back-translation ignore pose accuracy, and high-quality coordinated generation remains challenging. To create this crucial environment and overcome these challenges, we introduce CNText2Sign and CNSign, which together constitute the first unified dataset aimed at supporting bidirectional accessibility systems for Chinese sign language; CNText2Sign provides 15,000 natural language-to-sign mappings and standardized skeletal keypoints for 8,643 vocabulary items supporting pose assessment. Building upon this foundation, we propose the AuraLLM model, which leverages a decoupled architecture with CNText2Sign's pose data for novel direct gesture accuracy assessment. The model employs retrieval augmentation and Cascading Vocabulary Resolution to handle semantic mapping and out-of-vocabulary words, and achieves all-scenario production with controllable coordination of gestures and facial expressions via pose-conditioned video synthesis. Concurrently, our Sign Language Translation model SignMST-C employs targeted self-supervised pretraining for dynamic feature capture, achieving new SOTA results on PHOENIX2014-T with BLEU-4 scores up to 32.08. AuraLLM establishes a strong performance baseline on CNText2Sign with a BLEU-4 score of 50.41 under direct evaluation.

Explaining the decision-making process of Graph Neural Networks (GNNs) is essential for improving their transparency and reliability. However, real-world graphs are often heterogeneous and subject to structural noise, posing severe challenges to the robustness of existing explanation methods. To address these issues, we propose RoHeX, a Robust Heterogeneous GNN Explainer that enhances explanation quality under noisy conditions. RoHeX begins with a theoretical analysis revealing how different heterogeneous GNN architectures amplify structural perturbations through message passing. Building on this insight, we design a denoising variational inference framework that filters noisy structures and learns robust latent graph representations. Furthermore, we incorporate relation-aware heterogeneous semantics into the explanation generation process, formulating explanation as an optimization problem under the graph information bottleneck principle. This formulation enables RoHeX to balance fidelity and compactness, producing explanations that are both semantically meaningful and structurally stable. Comprehensive experiments on multiple real-world heterogeneous graphs demonstrate that RoHeX consistently surpasses state-of-the-art baselines in explanation fidelity, robustness to structural perturbations, and explainability.

Retrieval-augmented generation (RAG) has improved large language models (LLMs) on knowledge-intensive tasks, yet most systems assume static facts and struggle when answers depend on serialized and dynamic data, like time--e.g., ordering events, aligning facts to valid intervals, or planning actions under evolving conditions. This paper presents Knowledge-Graph Iterative Retrieval-Augmented Generation (KG-iRAG), a framework specialized for temporal reasoning. KG-iRAG couples a time-aware planner with a knowledge graph (KG) to iteratively fetch and compose evidence along a temporal axis. Concretely, it (i) represents events and facts with explicit timestamps and validity intervals; (ii) propagates temporal constraints through iterative retrieval using operators; and (iii) verifies temporal consistency while refining intermediate hypotheses, enabling step-by-step deduction for queries that mix knowledge retrieval with inference. Across public temporal QA benchmarks, KG-iRAG consistently improves accuracy and calibration over strong RAG baselines while reducing unnecessary retrieval through targeted, constraint-guided steps. To stress-test real-time decision queries, three application-oriented datasets (weatherQA-Irish, ~ weatherQA-Sydney, and~ trafficQA-TFNSW) are additionally constructed and tested alongside existing temporal benchmarks. The results demonstrate that injecting temporal structure into KG-driven RAG yields robust gains on multi-step, time-dependent queries, advancing the state of temporal reasoning with LLMs.

Personality detection from text is commonly performed by analysing users' social media posts. However, existing methods heavily rely on large-scale annotated datasets, making it challenging to obtain high-quality personality labels. Moreover, most studies treat emotion and personality as independent variables, overlooking their interactions. In this paper, we propose a novel self-supervised framework, EmoPerso, which improves personality detection through emotion-aware modelling. EmoPerso first leverages generative mechanisms for synthetic data augmentation and rich representation learning. It then extracts pseudo-labeled emotion features and jointly optimizes them with personality prediction via multi-task learning. A cross-attention module is employed to capture fine-grained interactions between personality traits and the inferred emotional representations. To further refine relational reasoning, EmoPerso adopts a self-taught strategy to enhance the model's reasoning capabilities iteratively. Extensive experiments on two benchmark datasets demonstrate that EmoPerso surpasses state-of-the-art models. The source code is available at https://github.com/slz0925/EmoPerso.

In critical domains such as healthcare and law, accurately modelling the uncertainty of automatic computational models is essential. For instance, healthcare models must produce reliable estimates to guide human decision-making. However, modelling uncertainty remains challenging, particularly for models handling low-resource datasets and complex, domain-specific vocabulary. Most existing predictive models model point estimates rather than probability distributions, limiting our ability to quantify model uncertainty. This paper introduces a novel model, BAKER, designed to address these limitations. BAKER combines the strengths of Bayesian inference, known for its effectiveness in modelling uncertainty, and kernel methods, which excel at capturing complex data relationships. Incorporating kernel functions enhances model performance, particularly by reducing overfitting in data-limited scenarios. Our experimental analysis shows that BAKER significantly improves uncertainty reasoning compared to existing models.

Multimodal fake news detection often involves modelling heterogeneous data sources, such as vision and language. Existing detection methods typically rely on fusion effectiveness and cross-modal consistency to model the content, complicating understanding how each modality affects prediction accuracy. Additionally, these methods are primarily based on static feature modelling, making it difficult to adapt to the dynamic changes and relationships between different data modalities. This paper develops a significantly novel approach, GAMED, for multimodal modelling, which focuses on generating distinctive and discriminative features through modal decoupling to enhance cross-modal synergies, thereby optimizing overall performance in the detection process. GAMED leverages multiple parallel expert networks to refine features and pre-embed semantic knowledge to improve the experts' ability in information selection and viewpoint sharing. Subsequently, the feature distribution of each modality is adaptively adjusted based on the respective experts' opinions. GAMED also introduces a novel classification technique to dynamically manage contributions from different modalities, while improving the explainability of decisions. Experimental results on the Fakeddit and Yang datasets demonstrate that GAMED performs better than recently developed state-of-the-art models. The source code can be accessed at https://github.com/slz0925/GAMED.

Federated learning facilitates collaborative training of machine learning models on data distributed across multiple locations, effectively addressing the privacy concerns by eliminating the need for data centralization—a critical consideration in medical image analysis. In healthcare applications like cardiovascular segmentation, datasets from individual sites often feature annotations for specific heart regions, leading to partial overlaps. To address this challenge, we present a two-step partial annotation framework for federated learning, featuring a hybrid 3D multi-encoding UNet enhanced with focal modulation layers in the second stage. This architecture enables specialized subnetworks to act as experts, extracting features tailored to specific regions of interest based on each client’s data. To further improve feature extraction and differentiation, we incorporate focal modulation blocks and apply regularization by introducing an auxiliary generic decoder during training. Comprehensive experiments on diverse cardiac MRI datasets demonstrate that our approach significantly outperforms centralized learning models.

Mental health concerns are a critical issue in today’s digital age, posing a threat to both individual and societal well-being and making the identification of at-risk individuals crucial. Analyzing an individual’s social media post history can offer insights into their mental health state and help identify the presence of suicidal ideation. However, the complexity of linguistic and temporal data, along with sparsity and time irregularities, poses a formidable challenge in machine learning. Previous methods in this domain either rely on Euclidean space for processing which does not adequately model the power-law properties of social media posts, or lose information due to the discretization of the time axis. To address these challenges, we propose a novel framework, THYMES, which leverages pre-trained encoders and a rich representation learning paradigm with hyperbolic learning to model power-law features for enhanced sequence modeling. We perform experiments on two datasets and demonstrate that THYMES outperforms previously proposed methods while maintaining classification fairness under heavy data imbalances. Additionally, we qualitatively analyze commonly misclassified samples to reveal the shortcomings of models in this domain.

The rising rate of mental health issues in the digital age underscores the critical need for proactive interventions to assess an individual’s well-being. This problem is further exacerbated by the social stigma surrounding the subject, which suppresses the willingness of victims to seek help. Social media can serve as an outlet for such individuals to express their negative emotions or thoughts of self-harm. The social media account of an individual can offer a plethora of valuable information that can be used to predict their mental health. By unifying principles of robust classifier training and selective classification, we propose a novel framework, SAFENet, to predict the suicide risk of users by using their historical social media posts. When the confidence of prediction is low or the individual is classified as a high-risk user, SAFENet delegates the analysis of the posts to a human evaluator for further intervention. Our experiments show that SAFENet outperforms existing state-of-the-art frameworks. We further qualitatively analyze predictions from SAFENet and demonstrate that it performs robustly on difficult samples that may cause contemporary methods to make errors. Our system addresses the urgent need for efficient and effective mental health intervention in the digital era.

Language models such as Bidirectional Encoder Representations from Transformers (BERT) have been very effective in various Natural Language Processing (NLP) and text mining tasks including text classification. However, some tasks still pose challenges for these models, including text classification with limited labels. This can result in a cold-start problem. Although some approaches have attempted to address this problem through single-stage clustering as an intermediate training step coupled with a pre-trained language model, which generates pseudo-labels to improve classification, these methods are often error-prone due to the limitations of the clustering algorithms. To overcome this, we have developed a novel two-stage intermediate clustering with subsequent fine-tuning that models the pseudo-labels reliably, resulting in reduced prediction errors. The key novelty in our model, IDoFew, is that the two-stage clustering coupled with two different clustering algorithms helps exploit the advantages of the complementary algorithms that reduce the errors in generating reliable pseudo-labels for fine-tuning. Our approach has shown significant improvements compared to strong comparative models.

This paper proposes a novel mask-based output layer for multi-level hierarchical classification, addressing the limitations of existing methods which (i) often do not embed the taxonomy structure being used, (ii) use a complex backbone neural network with n disjoint output layers that do not constraint each other, (iii) may output predictions that are often inconsistent with the taxonomy in place, and (iv) have often a fixed value of n. Specifically, we propose a model agnostic output layer that embeds the taxonomy and that can be combined with any model. Our proposed output layer implements a top-down divide-and-conquer strategy through a masking mechanism to enforce that predictions comply with the embedded hierarchy structure. Focusing on image classification, we evaluate the performance of our proposed output layer on three different datasets, each with a three-level hierarchical structure. Experiments on these datasets show that our proposed mask-based output layer allows to improve several multi-level hierarchical classification models using various performance metrics.

This paper presents Marine-tree, a large-scale hierarchical annotated dataset for marine organism classification. Marine-tree contains more than 160k annotated images divided into 60 classes organised in a hierarchy-tree structure using an adapted CATAMI (Collaborative and Automated Tools for the Analysis of Marine Imagery and video) classification scheme. Images were meticulously collected by scuba divers using the RLS (Reef Life Survey) methodology and later annotated by experts in the field. We also propose a hierarchical loss function that can be applied to any multi-level hierarchical classification model, which takes into account the parent-child relationship between predictions and uses it to penalize inconsistent predictions. Experimental results demonstrate thatMarine-tree and the proposed hierarchical loss function are a good contribution for both research in underwater imagery and hierarchical classification.

In this paper we introduce Jarvis, a context-as-a-service mobile tool, which enables context-aware data collection, service discovery, and computer-aided situational awareness through a conversational User Interface (UI). At the core of Jarvis are two main components: (i) a voice-based UI to translate speech to Context Definition and Query Language (Speech-to-CDQL), and (ii) an operational component called Context-as-a-Service (CoaaS), which enables smart things and IoT silos to discover, validate and share relevant and dependable context. The UI is based on two machine learning models: a Speech-to-Text model and a Text-to-CDQL model based on an encoder-decoder architecture. Jarvis is developed as a mobile application that allows people with different backgrounds to interact with various IoT devices. Our demo shows how easy Jarvis can be used for context-aware data collection and to interact with diverse objects in a smart home environment through voice.

The modern industrial sector generates enormous amounts of high-dimensional heterogeneous data daily. However, mostly the vectored data (rank-one tensor) have been considered for anomaly detection, whereas the data in real-life is high dimensional. The expressive power of methods based on vector data is restrictive as they may destroy the structural information embedded in data and lead to the curse-of-dimensionality and overfitting. In this paper, we present a novel anomaly detection approach for large-scale tensor data. We first present novel one-class support tensor machines (OCSTM) with bounded loss function. We further extend it by leveraging the randomness to design a scalable approach that can also be used for large-scale anomaly detection. To solve the corresponding optimization of the objective function, we utilize half-quadratic optimization followed by solving it like a traditional OCSTM optimization at each iteration. We demonstrate the proposed randomized OCSTM with bounded hinge loss through experiments on 14 benchmark data sets. Experimental results demonstrate the effectiveness of the proposed approach against anomalies and a significant reduction in the computational complexity.

Alzheimer's disease (AD) is an irreversible and progressive disorder where a large number of brain cells and their connections degenerate and die, eventually destroy the memory and other important mental functions that affect memory, thinking, language, judgment, and behavior. Not a single test can effectively determine AD; however, CT and magnetic resonance imaging (MRI) can be used to observe the decrease in size of different areas (mainly temporal and parietal lobes). This paper proposes an integrative deep ensemble learning framework to obtain better predictive performance for AD diagnosis. Unlike DenseNet, we present a multiresolutional ensemble PartialNet tailored to Alzheimer detection using brain MRIs. PartialNet incorporates the properties of identity mappings, diversified depth as well as deep supervision, thus, considers feature reuse that in turn results in better learning. Additionally, the proposed ensemble PartialNet demonstrates better characteristics in terms of vanishing gradient, diminishing forward flow with better training time, and a low number of parameters compared with DenseNet. Experiments performed on benchmark AD neuroimaging initiative data set that showed considerable performance gain (2 + % ↑ $\uparrow $ ) and (1.2 + % ↑ $\uparrow $ ) for multiclass and binary class in AD detection in comparison to state-of-the-art methods.

Twitter is currently a popular online social media platform which allows users to share their user-generated content. This publicly-generated user data is also crucial to healthcare technologies because the discovered patterns would hugely benefit them in several ways. One of the applications is in automatically discovering mental health problems, e.g., depression. Previous studies to automatically detect a depressed user on online social media have largely relied upon the user behaviour and their linguistic patterns including user's social interactions. The downside is that these models are trained on several irrelevant content which might not be crucial towards detecting a depressed user. Besides, these content have a negative impact on the overall efficiency and effectiveness of the model. To overcome the shortcomings in the existing automatic depression detection methods, we propose a novel computational framework for automatic depression detection that initially selects relevant content through a hybrid extractive and abstractive summarization strategy on the sequence of all user tweets leading to a more fine-grained and relevant content. The content then goes to our novel deep learning framework comprising of a unified learning machinery comprising of Convolutional Neural Network (CNN) coupled with attention-enhanced Gated Recurrent Units (GRU) models leading to better empirical performance than existing strong baselines.