With large-scale language models demonstrating superior capabilities in a wide range of downstream natural language processing tasks, the future trajectory of research in the field of text categorization faces increasing uncertainty. In this evolving paradigm of open-ended language modeling, where task delimitations are increasingly blurred, a pressing question arises: to what extent has text classification advanced under the full potential of large language model (LLM)? To address this pivotal inquiry, we introduce recurrent generative pre-trained transformer (RGPT), an adaptive boosting framework meticulously designed to craft a dedicated LLM for text classification. RGPT constructs a sequence of base learners by dynamically modulating the training data distribution and iteratively fine-tuning LLMs. These base learners are then progressively integrated, leveraging historical prediction trajectories to form a highly specialized text classification model. Extensive empirical evaluations demonstrate that RGPT surpasses eight state-of-the-art pretrained language models and seven cutting-edge LLMs across four benchmark datasets, achieving an average performance gain of 2.90%. © 2012 IEEE.

Smart healthcare aims to revolutionize med-ical services by integrating artificial intelligence (AI). The limitations of classical machine learning include privacy concerns that prevent direct data sharing among medical institutions, untimely updates, and long training times. To address these issues, this study proposes a digital twin-assisted quantum federated learning algorithm (DTQFL). By leveraging the 5G mobile network, digital twins (DT) of patients can be created instantly using data from various Internet of Medical Things (IoMT) devices and simultane-ously reduce communication time in federated learning (FL) at the same time. DTQFL generates DT for patients with specific diseases, allowing for synchronous training and updating of the variational quantum neural network (VQNN) without disrupting the VQNN in the real world. This study utilized DTQFL to train its own personalized VQNN for each hospital, considering privacy security and training speed. Simultaneously, the personalized VQNN of each hospital was obtained through further local iterations of the final global parameters. The results indicate that DTQFL can train a good VQNN without collecting local data while achieving accuracy comparable to that of data-centralized algorithms. In addition, after personalized train-ing, the VQNN can achieve higher accuracy than that with-out personalized training. © 2026 IEEE.

Nanorobots have been used in smart health to collect time series data such as electrocardiograms and electroencephalograms. Real-time classification of dynamic time series signals in nanorobots is a challenging task. Nanorobots in the nanoscale range require a classification algorithm with low computational complexity. First, the classification algorithm should be able to dynamically analyze time series signals and update itself to process the concept drifts (CD). Second, the classification algorithm should have the ability to handle catastrophic forgetting (CF) and classify historical data. Most importantly, the classification algorithm should be energy-efficient to use less computing power and memory to classify signals in real-time on a smart nanorobot. To solve these challenges, we design an algorithm that can Prevent Concept Drift in Online continual Learning for time series classification (PCDOL). The prototype suppression item in PCDOL can reduce the impact caused by CD. It also solves the CF problem through the replay feature. The computation per second and the memory consumed by PCDOL are only 3.572M and 1KB, respectively. The experimental results show that PCDOL is better than several state-of-the-art methods for dealing with CD and CF in energy-efficient nanorobots. © 2026 IEEE.

Anticancer peptides (ACPs) are short bioactive sequences that selectively target tumor cells with minimal toxicity, positioning them as promising candidates for next-generation cancer therapies. However, existing computational models face limitations in sequence representation and class imbalance. To address these challenges, we propose UACD-ACPs, a unified fusion-driven framework that integrates a diffusion-inspired noise-conditioned classifier for ACP prediction and a diffusion-based peptide generation module with cancer-type-aware organization for targeted downstream screening. The classification module integrates ProtBERT-based semantic embeddings with physicochemical descriptors via the Multiscale Embedding Compression Strategy (MECS) and a diffusion-inspired noise-conditioned encoder, substantially enhancing predictive robustness and accuracy, particularly under challenging imbalanced multi-class settings. In the generative pipeline, we introduce a denoising diffusion-based generative framework augmented by two novel fusion modules: the Bitemporal Fusion Module (BFM) and the Temporal Feature Attention Module (TFAM). These modules perform multi-scale temporal and semantic fusion to promote the generation of structurally coherent and functionally relevant peptide candidates. Experimental results demonstrate that UACD-ACPs outperforms state-of-the-art methods in terms of accuracy, F1-score, and AUC-ROC. The generated peptides exhibit favorable physicochemical properties, diverse secondary structures, and strong structural stability, as validated by molecular dynamics simulations and membrane-binding analyses. Overall, this study highlights the potential of fusion-driven diffusion-based frameworks for alleviating class imbalance and data heterogeneity in anticancer peptide modeling, paving the way for scalable and biologically grounded ACP discovery. © 2026 Li et al.

The subject of this work is to check how different types of music affect human emotions. While listening to music, a subjective survey and brain activity measurements were carried out using an EEG helmet. The aim is to demonstrate the impact of different music genres on emotions. The research involved a diverse group of participants of different gender and musical preferences. This had the effect of capturing a wide range of emotional responses to music. After the experiment, a relationship analysis of the respondents’ questionnaires with EEG signals was performed. The analysis revealed connections between emotions and observed brain activity. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2026.

Major depressive disorder is projected to become the leading contributor to mental illness by 2030. While resting-state functional magnetic resonance imaging (rs-fMRI) has emerged as a non-invasive solution for depression detection, two significant challenges remain. First, due to medical data privacy regulations and the high costs associated with acquiring the necessary equipment, individual medical institutions struggle to obtain sufficient annotated data. Second, domain shifts, caused by discrepancies in scanner parameters and acquisition protocols across multi-center datasets, significantly hinder model generalization. To address these challenges, we propose a federated domain adaptation (FDA) method that integrates co-activation patterns and a multimodal Mamba network, termed FDA-CAPMA, for fMRI-based depression detection. Specifically, a federated learning architecture ensures both physical data isolation and patient privacy through parameter aggregation. A state-space model-based Mamba network captures cross-modal correlations between fMRI time-series features and non-imaging features. Additionally, a local maximum mean discrepancy (LMMD) module aligns source and target domain distributions in both feature and prediction spaces. Extensive experiments on the largest multi-center depression dataset (Rest-meta-MDD, 1813 participants) and ABIDE dataset, our method achieves an accuracy of 67.16%, and 65.72%, respectively. This work establishes a new paradigm for privacy-preserving depression recognition. Code will be available at: https://github.com/helang818/FDA-CAPMA/ © 2026 Elsevier B.V.

Depression profoundly impacts multiple dimensions of an individual's life, including personal and social functioning, academic achievement, occupational productivity, and overall quality of life. With recent advancements in affective computing, deep learning technologies have been increasingly adopted to identify patterns indicative of depression. However, due to concerns over participant privacy, data in this domain remain scarce, posing significant challenges for the development of robust discriminative models for depression detection. To address this limitation, we build a Large-scale Multimodal Vlog Dataset (LMVD) for depression recognition in real-world settings. The LMVD dataset comprises 1,823 video samples, totaling approximately 214 h of content, collected from 1,475 participants across four major multimedia platforms: Sina Weibo, Bilibili, TikTok, and YouTube. In addition, we introduce a novel architecture, MDDformer, specifically designed to capture non-verbal behavioral cues associated with depressive states. Extensive experimental evaluations conducted on LMVD demonstrate the superior performance of MDDformer in depression detection tasks. We anticipate that LMVD will become a valuable benchmark resource for the research community, facilitating progress in multimodal, real-world depression recognition. The dataset and source code will be made publicly available at: https://github.com/helang818/LMVD. © 2025 Elsevier B.V., All rights reserved.

Next point-of-interest (POI) recommendation aims to predict the next location a user will visit based on historical behavioral data. Given their strong natural language understanding and reasoning capabilities, large language models (LLMs) have been increasingly introduced into next POI recommendation, as conventional methods are limited by their numerical representations, which often obscure the inherent semantic information embedded in contextual features However, LLM-based approaches struggle to effectively comprehend multiple factors holistically and express them in natural language. To address this issue, we propose Multi-Factor Collaboration for Next POI Recommendation (MFC4POI), a pretraining and fine-tuning framework that reformulates the next POI recommendation task as a question-answering problem. It leverages the strong comprehension and reasoning capabilities of LLMs to analyze user’s current behavior based on historical trajectories, and real time on-site situation, including the dynamic preferences and nearby geographic information in natural language. Due to the inherent limitations of LLMs in accurately modeling temporal and geographical information, MFC4POI introduces an intent identification agent to capture users’ dynamic preferences based on Spatio-temporal Intent-based Knowledge Graph (STIKG). It alleviates the weak spatial reasoning capability of LLMs while providing clear reasoning paths that make the recommendations more intuitive and coherent. To further improve understanding of users’s real-time on-site situation, MFC4POI integrates geographic information with dynamic preferences by providing nearby POIs that align with these dynamic preferences, thereby further enhancing LLMs’ performance. Experiments on three real-world datasets, containing up to 4000 users and 405000 check-ins, show that MFC4POI consistently outperforms strong baselines. Specifically, it achieves better performance than most baselines on Acc@1 and surpasses all baselines on Acc@5 and Acc@10 across all datasets. © 2026 Published by Elsevier Ltd.

As a new trend in Natural Language Processing (NLP), prompt tuning has been explored to provide a reliable answer without requiring massive labeled samples and training learning in sentiment and emotion detection tasks. However, how to effectively encode the commonness and uniquesness across difference affections into prompts sets a limit to the potential of multi-affection joint detection. To fill this gap, we propose a multi-affection prompt (MAP) learning framework that takes both the commonness of multiple affections and the uniqueness of specific affection into consideration. More specifically, two different prompt encoders are first proposed to elaborate the multi-task shared prompt and the task-specific prompt, respectively. Second, a multi-task prompt interaction learning layer is proposed to capture the correlation between the multi-task and task-specific prompts. MAP adopts separate multi-task and task-specific prompts to learn different vectors for different affection tasks, thus mitigating the affection discrepancy of the [MASK] token in the masked language modeling task. Extensive experiments on two benchmark datasets show that our proposed method can significantly improve the multi-task generalization capability of PLMs, and yield better results than other state-of-the-art (SOTA) baselines, by the margin of 2.7% and 3.4%. © The author(s) 2026. The articles published in this open access journal are distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/).

With the accumulation of resources in the era of big data and the rise of pre-trained models in deep learning, optimizing neural networks for various tasks often involves different strategies for fine-tuning pre-trained models versus training from scratch. However, existing optimizers primarily focus on reducing the loss function by updating model parameters, without fully addressing the unique demands of these two major paradigms. In this paper, we propose DualOpt, a novel approach that decouples optimization techniques specifically tailored for these distinct training scenarios. For training from scratch, we introduce real-time layer-wise weight decay, designed to enhance both convergence and generalization by aligning with the characteristics of weight updates and network architecture. For more importantly fine-tuning, we integrate weight rollback with the optimizer, incorporating a rollback term into each weight update step. This ensures consistency in the weight distribution between upstream and downstream models, effectively mitigating knowledge forgetting and improving fine-tuning performance. Additionally, we extend the layer-wise weight decay to dynamically adjust the rollback levels across layers, adapting to the varying demands of different downstream tasks. Extensive experiments across diverse tasks, including image classification, object detection, semantic segmentation, and instance segmentation, demonstrate the broad applicability and state-of-the-art performance of DualOpt. © 2026 IEEE.