Search Results (95)

This paper proposes a technological framework designed to mitigate the inherent risks associated with the deployment of artificial intelligence (AI) in decision-making and task execution within the management processes. The Agreement Validation Interface (AVI) functions as a modular Application Programming Interface (API) Gateway positioned between user applications and LLMs. This gateway architecture is designed to be LLM-agnostic, meaning it can operate with various underlying LLMs without requiring specific modifications for each model. This universality is achieved by standardizing the interface for requests and responses and applying a consistent set of validation and enhancement processes irrespective of the chosen LLM provider, thus offering a consistent governance layer across a diverse LLM ecosystem. AVI facilitates the orchestration of multiple AI subcomponents for input–output validation, response evaluation, and contextual reasoning, thereby enabling real-time, bidirectional filtering of user interactions. A proof-of-concept (PoC) implementation of AVI was developed and rigorously evaluated using industry-standard benchmarks. The system was tested for its effectiveness in mitigating adversarial prompts, reducing toxic outputs, detecting personally identifiable information (PII), and enhancing factual consistency. The results demonstrated that AVI reduced successful fast injection attacks by 82%, decreased toxic content generation by 75%, and achieved high PII detection performance (F1-score ≈ 0.95). Furthermore, the contextual reasoning module significantly improved the neutrality and factual validity of model outputs. Although the integration of AVI introduced a moderate increase in latency, the overall framework effectively enhanced the reliability, safety, and interpretability of LLM-driven applications. AVI provides a scalable and adaptable architectural template for the responsible deployment of generative AI in high-stakes domains such as finance, healthcare, and education, promoting safer and more ethical use of AI technologies. Full article

Data workflows are an important component of modern analytical systems, enabling structured data extraction, transformation, integration, and delivery across diverse applications. Despite their importance, these workflows are often developed using ad hoc approaches, leading to scalability and maintenance challenges. This paper proposes a structured, three-level methodology—conceptual, logical, and physical—for modeling data workflows using Business Process Model and Notation (BPMN). A custom BPMN metamodel is introduced, along with a tool built on BPMN.io, that enforces modeling constraints and supports translation from high-level workflow designs to executable implementations. Logical models are further enriched through blueprint definitions, specified in a formal, implementation-agnostic JSON schema. The methodology is validated through a case study, demonstrating its applicability across ETL and machine learning domains, promoting clarity, reuse, and automation in data pipeline development. Full article

The rapid advancement of Large Language Models (LLMs) has led to substantial investment in enhancing their capabilities and expanding their feature sets. Despite these developments, a critical gap remains between model sophistication and their dependable deployment in real-world applications. A key concern is the inconsistency of LLM-generated outputs in production environments, which hinders scalability and reliability. In response to these challenges, we propose a novel framework that integrates custom-defined, rule-based logic to constrain and guide LLM behavior effectively. This framework enforces deterministic response boundaries while considering the model’s reasoning capabilities. Furthermore, we introduce a quantitative performance scoring mechanism that achieves an 85.5% improvement in response consistency, facilitating more predictable and accountable model outputs. The proposed system is industry-agnostic and can be generalized to any domain with a well-defined validation schema. This work contributes to the growing research on aligning LLMs with structured, operational constraints to ensure safe, robust, and scalable deployment. Full article

Retrieval-Augmented Generation (RAG) has become a central approach to enhance the factual consistency and domain specificity of large language models (LLMs) by incorporating external context at inference time. However, most existing RAG systems rely on dense vector-based similarity, which fails to capture complex semantic structures, relational dependencies, and multimodal content. In this paper, we introduce Graph Codes—a matrix-based encoding of Multimedia Feature Graphs—as an alternative retrieval paradigm. Graph Codes preserve semantic topology by explicitly encoding entities and their typed relationships from multimodal documents, enabling structure-aware and interpretable retrieval. We evaluate our system in two domains: multimodal scene understanding (200 annotated image-question pairs) and clinical question answering (150 real-world medical queries with 10,000 structured knowledge snippets). Results show that our method outperforms dense retrieval baselines in precision (+9–15%), reduces hallucination rates by over 30%, and yields higher expert-rated answer quality. Theoretically, this work demonstrates that symbolic similarity over typed semantic graphs provides a more faithful alignment mechanism than latent embeddings. Practically, it enables interpretable, modality-agnostic retrieval pipelines deployable in high-stakes domains such as medicine or law. We conclude that Graph Code-based RAG bridges the gap between structured knowledge representation and neural generation, offering a robust and explainable alternative to existing approaches. Full article

Autoimmune diseases such as multiple sclerosis (MS) are characterized by a loss of self-tolerance, driven by diminished regulatory T cell (Treg) function and elevated Th1/Th17 responses. Existing therapies broadly suppress the immune system without correcting this imbalance, often leading to adverse effects. LPX3, a novel small-molecule T cell immunoglobulin and mucin domain-containing 3 and 4 (Tim-3/4) receptor agonist, was developed to restore immune tolerance via Treg induction. In this study, LPX3 was formulated into a liposomal oral delivery system, enabling efficient uptake through the gastrointestinal tract and lymphatic targeting. In vitro and in vivo analyses confirmed LPX3’s ability to expand CD4⁺Foxp3⁺ Tregs in a dose-dependent manner. In a MOG-induced experimental autoimmune encephalomyelitis (EAE) mouse model of MS, both prophylactic and therapeutic oral administration of LPX3 significantly delayed disease onset, reduced symptom severity, and improved survival. Importantly, efficacy was achieved without antigen co-delivery, indicating an antigen-independent mechanism of immune modulation. LPX3 liposomes showed deep lymph node penetration and colocalization with immune cells, supporting its functional delivery to key immunological sites. These findings suggest LPX3 is a promising candidate for treating autoimmune diseases by re-establishing immune regulation through oral, antigen-agnostic tolerance induction. Full article

This study addresses the critical need for effective data fusion strategies in pest prediction for pitahaya (dragon fruit) cultivation in the Ecuadorian Amazon, where heterogeneous data sources—such as environmental sensors and chlorophyll measurements—offer complementary but fragmented insights. Current agricultural monitoring systems often fail to integrate these data streams, limiting early pest detection accuracy. To overcome this, we compared early and late fusion approaches using comprehensive experiments. Multidimensionality is a central challenge: the datasets span temporal (hourly sensor readings), spatial (plot-level chlorophyll samples), and spectral (chlorophyll reflectance) dimensions. We applied dimensionality reduction techniques—PCA, KPCA (linear, polynomial, RBF), t-SNE, and UMAP—to preserve relevant structure and enhance interpretability. Evaluation metrics included the proportion of information retained (score) and cluster separability (silhouette score). Our results demonstrate that early fusion yields superior integrated representations, with PCA and KPCA-linear achieving the highest scores (0.96 vs. 0.94), and KPCA-poly achieving the best cluster definition (silhouette: 0.32 vs. 0.31). Statistical validation using the Friedman test (${\chi }^2$ = 12.00, p = 0.02) and Nemenyi post hoc comparisons (p < 0.05) confirmed significant performance differences. KPCA-RBF performed poorly (score: 0.83; silhouette: 0.05), and although t-SNE and UMAP offered visual insights, they underperformed in clustering (silhouette < 0.12). These findings make three key contributions. First, early fusion better captures cross-domain interactions before dimensionality reduction, improving prediction robustness. Second, KPCA-poly offers an effective non-linear mapping suitable for tropical agroecosystem complexity. Third, our framework, when deployed in Joya de los Sachas, improved pest prediction accuracy by 12.60% over manual inspection, leading to more targeted pesticide use. This contributes to precision agriculture by providing low-cost, scalable strategies for smallholder farmers. Future work will explore hybrid fusion pipelines and sensor-agnostic models to extend generalizability. Full article

The construction industry frequently encounters safety hazards, with falls related to undetected openings being a major cause of fatalities. Identifying unstructured openings using computer vision is challenging due to their unpredictable nature and the difficulty of acquiring large labeled datasets in dynamic construction environments. Conventional deep learning methods require substantial data, limiting their applicability. Few-shot learning (FSL) offers a promising alternative by enabling models to learn from limited examples. This study investigates the effectiveness of an FSL approach, specifically model-agnostic meta-learning (MAML), enhanced with domain-specific attributes, for identifying unstructured openings with minimal labeled data. We developed and evaluated an attribute-enhanced MAML framework under various few-shot conditions (k-way, n-shot) and compared its performance against conventional supervised fi-ne-tuning. The results demonstrate that the proposed FSL model achieved high classification accuracy (over 90.5%) and recall (over 85.5%) using only five support shots per class. Notably, the FSL approach significantly outperformed supervised fine-tuning methods under the same limited data conditions, exhibiting substantially higher recall crucial for safety monitoring. These findings validate that FSL, augmented with relevant attributes, provides a data-efficient and effective solution for monitoring unpredictable hazards like unstructured openings, reducing the reliance on extensive data annotation. This research contributes valuable insights for developing adaptive and robust AI-powered safety monitoring systems in the construction domain. Full article

This study proposes an Aspect-Enhanced Prompting (AEP) method for unsupervised Multi-Source Domain Adaptation in Aspect Sentiment Classification, where data from the target domain are completely unavailable for model training. The proposed AEP is based on two generative language models: one generates a prompt from a given review, while the other follows the prompt and classifies the sentiment of an aspect. The first model extracts Aspect-Related Features (ARFs), which are words closely related to the aspect, from the review and incorporates them into the prompt in a domain-agnostic manner, thereby directing the second model to identify the sentiment accurately. Our framework incorporates an innovative rescoring mechanism and a cluster-based prompt expansion strategy. Both are intended to enhance the robustness of the generation of the prompt and the adaptability of the model to diverse domains. The results of experiments conducted on five datasets (Restaurant, Laptop, Device, Service, and Location) demonstrate that our method outperforms the baselines, including a state-of-the-art unsupervised domain adaptation method. The effectiveness of both the rescoring mechanism and the cluster-based prompt expansion is also validated through an ablation study. Full article

Despite the growing adoption of Property Graph Databases, like Neo4j, interacting with them remains difficult for non-technical users due to the reliance on formal query languages. Natural Language Interfaces (NLIs) address this by translating natural language (NL) into Cypher. However, existing solutions are typically limited to high-resource languages; are difficult to adapt to evolving domains with limited annotated data; and often depend on Machine Learning (ML) approaches, including Large Language Models (LLMs), that demand substantial computational resources and advanced expertise for training and maintenance. We address these limitations by introducing a novel dependency-based, training-free, schema-agnostic Natural Language Interface (NLI) that converts NL queries into Cypher for querying Property Graphs. Our system employs a modular pipeline-integrating entity and relationship extraction, Named Entity Recognition (NER), semantic mapping, triple creation via syntactic dependencies, and validation against an automatically extracted Schema Graph. The distinctive feature of this approach is the reduction in candidate entity pairs using syntactic analysis and schema validation, eliminating the need for candidate query generation and ranking. The schema-agnostic design enables adaptation across domains and languages. Our system supports single- and multi-hop queries, conjunctions, comparisons, aggregations, and complex questions through an explainable process. Evaluations on real-world queries demonstrate reliable translation results. Full article

With the widespread application of machine learning techniques in time series analysis, the interpretability of models trained on time series data has attracted increasing attention. Most existing explanation methods are based on time-domain features, making it difficult to reveal how complex models focus on time–frequency information. To address this, this paper proposes a time–frequency domain-based time series interpretation method aimed at enhancing the interpretability of models at the time–frequency domain. This method extends the traditional LIME algorithm by combining the ideas of short-time Fourier transform (STFT), inverse STFT, and local interpretable model-agnostic explanations (LIME), and introduces a self-designed TFHS (time–frequency homogeneous segmentation) algorithm. The TFHS algorithm achieves precise homogeneous segmentation of the time–frequency matrix through peak detection and clustering analysis, incorporating the distribution characteristics of signals in both frequency and time dimensions. The experiment verified the effectiveness of the TFHS algorithm on Synthetic Dataset 1 and the effectiveness of the TF-LIME algorithm on Synthetic Dataset 2, and then further evaluated the interpretability performance on the MIT-BIH dataset. The results demonstrate that the proposed method significantly improves the interpretability of time-series models in the time–frequency domain, exhibiting strong generalization capabilities and promising application prospects. Full article

In this study, we present a novel concept termed open-vocabulary domain generalization (OVDG), which we investigate within the context of semantic segmentation. OVDG presents greater difficulty compared to conventional domain generalization, yet it offers greater practicality. It jointly considers (1) recognizing both base and novel classes and (2) generalizing to unseen domains. In OVDG, only the labels of base classes and the images from source domains are available to learn a robust model. Then, the model could be generalized to images from novel classes and target domains directly. In this paper, we propose a dual-branch FreeMix module to implement the OVDG task effectively in a universal framework: the base segmentation branch (BSB) and the entity segmentation branch (ESB). First, the entity mask is introduced as a novel concept for segmentation generalization. Additionally, semantic logits are learned for both the base mask and the entity mask, enhancing the diversity and completeness of masks for both base classes and novel classes. Second, the FreeMix utilizes pretrained self-supervised learning on large-scale remote-sensing data (RS_SSL) to extract domain-agnostic visual features for decoding masks and semantic logits. Third, a training tactic called dataset-aware sampling (DAS) is introduced for multi-source domain learning, aimed at improving the overall performance. In summary, RS_SSL, ESB, and DAS can significantly improve the generalization ability of the model on both a class level and a domain level. Experiments demonstrate that our method produces state-of-the-art results on several remote-sensing semantic-segmentation datasets, including Potsdam, GID5, DeepGlobe, and URUR, for OVDG. Full article

Cohesive human–robot collaboration can be achieved through seamless communication between human and robot partners. We posit that the design aspects of human–robot communication (HRCom) can take inspiration from human communication to create more intuitive systems. A key component of HRCom systems is perception models developed using machine learning. Being data-driven, these models suffer from the dearth of comprehensive, labelled datasets while models trained on standard, publicly available datasets do not generalize well to application-specific scenarios. Complex interactions and real-world variability lead to shifts in data that require domain adaptation by the models. Existing domain adaptation techniques do not account for incommensurable modes of communication between humans and robot perception systems. Taking into account these challenges, a novel framework is presented that leverages existing domain adaptation techniques off-the-shelf and uses statistical measures to start and stop the training of models when they encounter domain-shifted data. Statistically informed multimodal (domain adaptation by transfer) learning (SIMLea) takes inspiration from human communication to use human feedback to auto-label for iterative domain adaptation. The framework can handle incommensurable multimodal inputs, is mode and model agnostic, and allows statistically informed extension of datasets, leading to more intuitive and naturalistic HRCom systems. Full article

Model-agnostic meta-learning (MAML), coupled with digital twins, is transformative for predictive maintenance (PdM), especially in robotic arms in assembly lines, where rapid and accurate fault classification of arms is essential. Despite gaining significant traction, the framework faces significant challenges, like hypersensitivity to learning parameters and limited generalization during meta-testing. To address these challenges, we proposed an ensemble-based meta-learning approach integrating majority voting with model-agnostic meta-learning (MAML), and operational grouping was implemented via Latin hypercube sampling (LHS) to enhance few-shot learning ability and generalization along with maintaining stable output. This approach demonstrates superior accuracy in classifying a significantly larger number of defective mechanical classes, particularly in cross-domain few-shot (CDFS) learning scenarios. The proposed methodology is validated using a synthetic vibration signal dataset of robotic arm faults generated via a digital twin. Comparative analysis with existing frameworks, including ANIL, Protonet, and Reptile, confirms that our approach achieves higher accuracy in the given scenario. Full article

Traditional text classification models predominantly rely on static text representations, failing to capture temporal variations in language usage and evolving semantic meanings. This limitation reduces their ability to accurately classify time-sensitive texts, where understanding context, detecting trends, and addressing semantic shifts over time are critical. This paper introduces a novel time-aware short text classification model incorporating temporal information, enabling tracking of and adaptation to evolving language semantics. The proposed model enhances contextual understanding by leveraging timestamps and significantly improves classification accuracy, particularly for time-sensitive applications such as News topic classification. The model employs a hybrid architecture combining Convolutional Neural Networks (CNNs) and Bidirectional Long Short-Term Memory (BiLSTM) networks, enriched with attention mechanisms to capture both local and global dependencies. To further refine semantic representation and mitigate the effects of semantic drift, the model fine-tunes GloVe embeddings and employs synonym-based data augmentation. The proposed approach is evaluated on three benchmark dynamic datasets, achieving superior performance with classification accuracy reaching 92% for the first two datasets and 85% for the third dataset. Furthermore, the model is applied to a different-fields categorization and trend analysis task, demonstrating its capability to capture temporal patterns and perform detailed trend analysis of domain-agnostic textual content. These results underscore the potential of the proposed framework to provide deeper insights into the evolving nature of language and its impact on short-text classification. This work advances natural language processing by offering a comprehensive time-aware classification framework, addressing the challenges of temporal dynamics in language semantics. Full article

Neural networks (NNs) are essential in advancing modern safety-critical systems. Lightweight NN architectures are deployed on resource-constrained devices using hardware accelerators like Graphics Processing Units (GPUs) for fast responses. However, the latest semiconductor technologies may be affected by physical faults that can jeopardize the NN computations, making fault mitigation crucial for safety-critical domains. The recent studies propose software-based Hardening Techniques (HTs) to address these faults. However, the proposed fault countermeasures are evaluated through different hardware-agnostic error models neglecting the effort required for their implementation and different test benches. Comparing application-level HTs across different studies is challenging, leaving it unclear (i) their effectiveness against hardware-aware error models on any NN and (ii) which HTs provide the best trade-off between reliability enhancement and implementation cost. In this study, application-level HTs are evaluated homogeneously and independently by performing a study on the feasibility of implementation and a reliability assessment under two hardware-aware error models: (i) weight single bit-flips and (ii) neuron bit error rate. Our results indicate that not all HTs suit every NN architecture, and their effectiveness varies depending on the evaluated error model. Techniques based on the range restriction of activation function consistently outperform others, achieving up to 58.23% greater mitigation effectiveness while keeping the introduced overhead at inference time low while requiring a contained effort in their implementation. Full article