Adapting large language models to specialized domains remains challenging due to the computational cost of full finetuning and the limited availability of domain-specific parallel data. We present a systematic framework for parameter-efficient domain adaptation using Low-Rank Adaptation (LoRA) geared towards efficient learning in low-resource scenarios. Our method combines data-scaling analysis, dual-track hyperparameter optimization, and competitive benchmarking. We evaluate our approach on the low-resource English–Norwegian petroleum translation domain using a distilled version of NLLB and parallel data from the Norwegian Petroleum Directorate. Our adapted model achieves 61.48 BLEU (+24.62 over the base model) and 0.9298 COMET, while updating <0.4% of parameters. Our results provide a reproducible and computationally efficient blueprint for domain adaptation in neural machine translation, particularly for specialized and resource-constrained domains.

Fire situation monitoring is critical for protecting lives and property, offering real-time data on fire location, intensity, and progression. Traditional vision-based methods often treat fires as rigid objects, overlooking their fuzzy boundaries. To address this, we propose the subtractive counterfactual-based fire segmentation (SCF) method, which enhances edge detail focus for precise fire profiling. The approach constructs a counterfactual image by masking flames. Factual and counterfactual features are extracted and hypothetical counterfactual features are introduced during inference. Subtracting factual features from hypothetical counterfactual features yields unbiased fire features for segmentation. Experiments on five datasets demonstrate SCF's superior accuracy and reliability in fire monitoring.

Biohybrid microrobots with autonomous movement capabilities have broad application prospects in targeted delivery, attracting researchers to study their movement characteristics. However, its automatic control is still challenging, and exploring real-time detection of its environment for path planning to achieve stable closed-loop control is highly important for its practical application. Here, we applied deep learning for the detection of biohybrid microrobots and their targets and obstacles, followed by real-time path planning and trajectory tracking of biohybrid microrobots for targeted delivery. The proposed detection algorithm introduces attention and multi-scale feature fusion mechanisms in YOLOv7 algorithm (AM-YOLOv7) with the aim of enhancing the precision of detecting small-scale targets when robots, obstacles and targets are displayed globally, and the detection capabilities are verified through simulations and experiments. The proposed planning algorithm introduces a turning penalty function and a path smoothing strategy into A* algorithm (PS-A*) to make the planned path short and smooth, which has been verified through simulation and experiments. The adaptive fuzzy PID method is used to track the robot's trajectory, and experiments and simulations show that the biohybrid microrobot can move according to the preset trajectory better. The final cell scene experimental results show that the biohybrid microrobot using this system can effectively avoid obstacle cells and be delivered to target cells. The system can detect biohybrid microrobots, obstacle cells and target cells, plan short and smooth trajectories, and track them accurately. The proposed method has certain generalizability and broad application prospects in targeted delivery.

Among various disasters, fire poses one of the most widespread threats to public safety. The early stage of fire, marked by small and slow-spreading fire objects, is the ideal time for firefighting intervention. Therefore, early fire detection is crucial to prevent potential hazards and reduce loss of life and property. However, due to the variable shapes and small size, existing methods struggle to precisely locate smoke and flame in the initial stage. Compared to rigid objects, flame has unique local morphology, such as sharp tip, irregular edge and cavity. These prior knowledge could guide the model to focus on the local morphological features of the flame, thereby improving early fire detection ability. To address the current challenges, the paper proposes a Local Morphological Knowledge Matching based Early Fire Detector(LMKMFD), which accurately detects early fires and locates the multi-scale fires by exploring local flame morphology in fire images. Firstly, the local morphological features are extracted by matching the input fire image with the knowledge templates of the designed local morphological knowledge base. Secondly, fire multi-scale semantic features at four scales are mined by a Transformer-based backbone. Finally, fire local morphological features and multi-scale semantic features are aggregated by single-level depth prediction module to achieve region-level localization of fire objects. Experimental results on private and public datasets show that LMKMFD exhibits high detection precision for fires of different scales, particularly small early fires. LMKMFD outperforms baseline models, with mean Average Precision(mAP) of 88.38% and 83.12% on two datasets. Notably, due to the significance of local morphology of flame, the method performs better in detecting flame than smoke.

Systematic investigation of tumor-infiltrating immune (TII) cells is important to the development of immunotherapies, and the clinical response prediction in cancers. There exists complex transcriptional regulation within TII cells, and different immune cell types display specific regulation patterns. To dissect transcriptional regulation in TII cells, we first integrated the gene expression profiles from single-cell datasets, and proposed a computational pipeline to identify TII cell type-specific transcription factor (TF) mediated activity immune modules (TF-AIMs). Our analysis revealed key TFs, such as BACH2 and NFKB1 play important roles in B and NK cells, respectively. We also found some of these TF-AIMs may contribute to tumor pathogenesis. Based on TII cell type-specific TF-AIMs, we identified eight CD8+ T cell subtypes. In particular, we found the PD1 + CD8+ T cell subset and its specific TF-AIMs associated with immunotherapy response. Furthermore, the TII cell type-specific TF-AIMs displayed the potential to be used as predictive markers for immunotherapy response of cancer patients. At the pan-cancer level, we also identified and characterized six molecular subtypes across 9680 samples based on the activation status of TII cell type-specific TF-AIMs. Finally, we constructed a user-friendly web interface CellTF-AIMs (http://bio-bigdata.hrbmu.edu.cn/CellTF-AIMs/) for exploring transcriptional regulatory pattern in various TII cell types. Our study provides valuable implications and a rich resource for understanding the mechanisms involved in cancer microenvironment and immunotherapy.

In the contemporary social environment, social crisis events occur frequently with significant impacts. Effective management of these events requires comprehensive group intention mining, which encompasses intention detection and intention attribution. Knowledge graph inference facilitates the detection of group intention in crisis events. This is supported by the construction of crisis knowledge graphs, which organize crisis elements and inter-element relations into structured semantic information. This paper provides a comprehensive overview of the research about knowledge graph in social crisis management, focusing on three key areas: knowledge graph construction and inference, knowledge graph-based interpretable crisis attribution, and risk management. Specifically, the interpretable semantics in crisis knowledge graphs enables attribution of intention. To illustrate the significance of knowledge graphs in group intention mining, the COVID-19 and China–US game events are selected as two case studies. Finally, the paper proposes future research directions to solve the limitations of existing knowledge graph-related methods in social crises.

Multimodal emotion recognition comprehensively identifies the emotion contained in multimodal data by bridging the gaps between heterogeneous dataset. In recent years, multimodal emotion recognition methods have gained significant attention and been shown to surpass single-modal approaches. Most of the existing multimodal emotion analysis methods simply combine different modalities to improve the recognition capability of consistent emotion expressions across multimodal data. However, it remains challenging to recognize the right emotion when multiple modalities' contents carry inconsistent or even contradictory emotions. To solve this problem, we propose a novel image-text emotion recognition model named Multimodal Cross-Attention Bayesian Network(MCABN). The entire network exploits Bayesian theory to learn the distribution of its weight parameters, making optimization directions and results of parameters interpretable. What's more, the model leverages the consistency and complementarity between visual content and textual description to arrive at accurate decisions. Specifically, for each modality, multiple explainable features (color, texture, and shape feature in the image, while adjective, adverb, verb, noun, and negative feature in the text) and one unexplainable feature are fused as its feature representation to reinforce emotion-related features. Then two single-modal attention modules(Visual Attention Module and Textual Attention Module) capture the most discriminative features in a single image and text; Two cross-modal attention modules(Image-guided Text Attention Module and Text-guided Image Attention Module) extract the complementary and dominant features between two modalities by interactive learning. Finally, the outputs of four attention modules are integrated through intermediate fusion to predict the final emotion. The experimental results on the NVTD and MVSA-Multiple dataset indicate that the proposed MCABN outperforms state-of-the-art baselines by substantial margins.

In the current social environment, social crisis events occur frequently with significant impacts.Group intention mining through automatic technologies for managing social crises has gained extensive attention.This paper presents an overview of research on group intention mining in social crisis events, covering three areas: knowledge graph inference, intention attribution, and risk management.Knowledge graph inference facilitates the detection of group intention in crisis events.It is supported by the construction of crisis knowledge graphs, which organize crisis elements and inter-element relations into structured semantic knowledge.The interpretable semantics in the crisis knowledge graphs enables attribution of intention.Group intention mining consists of intention detection and intention attribution, serving the risk management of social crisis events.To gain insights into the process of group intention mining in social crises, the Covid-19 event is selected as a case study.Finally, the paper proposes future research directions to solve the limitations of existing intention mining methods in social crises.

In this article, the event-triggered multistep model predictive control for the discrete-time nonlinear system over communication networks under the influence of packet dropouts and cyber attacks is studied. First, the interval type-2 Takagi-Sugeno fuzzy model is applied to express the discrete-time nonlinear system and an event-triggered mode, which is capable of determining whether the sampled signal ought to be delivered into the unreliable network, is designed to economize communication resources. Second, two Bernoulli processes are introduced to represent the randomly happening packet dropouts in the unreliable network and the randomly occurring deception attacks on the actuator side from the adversaries. Third, under the assumption that the system states are unmeasurable, a multistep parameter-dependent model predictive controller is synthesized via optimizing one series of feedback laws for a given period of time, which leads to improved control performance than that of the one-step approach. Moreover, the results on the recursive feasibility and closed-loop stability related to the networked system are achieved, which explicitly consider the external disturbance and input constraint. Finally, simulation experiments on the mass-spring-damping system are carried out to illustrate the rationality and effectiveness of the provided control strategy.

RNA-sequencing enables accurate and low-cost transcriptome-wide detection. However, expression estimates vary as reference genomes and gene annotations are updated, confounding existing expression-based prognostic signatures. Herein, prognostic 9-gene pair signature (GPS) was applied to 197 patients with stage I lung adenocarcinoma derived from previous and latest data from The Cancer Genome Atlas (TCGA) processed with different reference genomes and annotations. For 9-GPS, 6.6% of patients exhibited discordant risk classifications between the two TCGA versions. Similar results were observed for other prognostic signatures, including IRGPI, 15-gene and ORACLE. We found that conflicting annotations for gene length and overlap were the major cause of their discordant risk classification. Therefore, we constructed a prognostic 40-GPS based on stable genes across GENCODE v20-v30 and validated it using public data of 471 stage I samples (log-rank P < 0.0010). Risk classification was still stable in RNA-sequencing data processed with the newest GENCODE v32 versus GENCODE v20-v30. Specifically, 40-GPS could predict survival for 30 stage I samples with formalin-fixed paraffin-embedded tissues (log-rank P = 0.0177). In conclusion, this method overcomes the vulnerability of existing prognostic signatures due to reference genome and annotation updates. 40-GPS may offer individualized clinical applications due to its prognostic accuracy and classification stability.

To improve an intercultural communication between foreigners and Chinese in mainland China, an app called INTER is designed help them to find common communication topics. Questions & Answers (Q&A) and information push are utilized in the app application. A User interface is designed and preliminary tested by 30 participants studied in mainland china, who come from different cultural environments. During the test, they are given one actual user interface, and asked to fill in a questionnaire, followed by an interview for user experience collection and their personal opinions about function designs with application flow chat. INTER is demonstrated as a digital intercultural communication tool which can be utilized as software in mobile phone for an interactive accessible communication. Interface design would be further improved in the future study.