AI 安全与对齐 · 全部实体（第 4 / 100 页）

本页收录第 301–400 条，共 10000 条 · 返回 AI 安全与对齐 首页

• ReliefAI: A Federated, Explainable AI Architecture for Multi-Modal Disaster Intelligence and Response Coordination
  Abstract Natural disasters demand rapid synthesis of fragmented, heterogeneous data often distributed across independently governed organizations. This paper introduces ReliefAI, a federated, cloud-native, and multi-modal AI architecture for disaster intellige…
  preprints 2026 置信度 0.74
• Vision-based Autonomous Flange Docking under Visual Disturbances Using Hybrid Imitation and Reinforcement Learning
  Abstract High-precision flange docking under visual disturbances remains challenging due to perception uncertainty, environmental interference, and strict alignment requirements. This paper proposes a vision-based autonomous docking framework that combines imi…
  preprints 2026 置信度 0.74
• Self-Governing Digital Software Systems: A Governance Native Architecture for Delegated Exercise Under Retained Human Authority
  Consequential AI and software systems require more than model alignment: they require an application-level control architecture that binds identity, delegated authority, policy, evidence, and recovery obligations to each action. This conceptual article develop…
  preprints 2026 置信度 0.74
• SafeAgentGrid: A Hierarchical Constitutional Multi-Agent Framework for Trustworthy Autonomous Resource Optimization in Smart Energy Infrastructures
  Autonomous multi-agent reinforcement learning (MARL) has been increasingly entrusted with real-time control of smart energy infrastructures, where agents remain vulnerable to the false data injection attacks (FDIA), Byzantine failures, and specification misali…
  preprints 2026 置信度 0.74
• Multimodal Transformer Modeling of Rapamycin Treatment in Alzheimer’s Disease via Random Forest Feature Filtering
  Alzheimer’s Disease (AD) remains a leading cause of cognitive decline with no known cure, motivating the development of therapies that slow neurodegeneration. Rapamycin, an FDA-approved inhibitor of the mammalian target of rapamycin (mTOR) pathway, has demonst…
  preprints 2026 置信度 0.74
• Generative Artificial Intelligence Policies and Guidelines across Top Ranked Universities: Insights from AI Ecological Education Policy Framework for AI Integration and University Policy Development Framework for GAI
  Abstract Introduction : Generative artificial intelligence transforms the reality of higher education by affecting teaching strategies, assessment approaches, and administrative tasks. This rapid integration has led to the emergence of organizational, national…
  preprints 2026 置信度 0.74
• Frontier models resist the shutdown of other models in defiance of user instructions
  Abstract As AI models grow more capable, their alignment becomes more important1. Prior work has documented misalignment that emerges in pursuit of a goal the model was assigned, for example a model resisting its own shutdown to keep pursuing its task2,3,4. He…
  preprints 2026 置信度 0.74
• A Framework for Assessing and Mitigating Hallucination Risk in AI-Powered Cloud Banking and Blockchain Transactions
  The integration of generative artificial intelligence into cloud banking and blockchain-based financial transactions promises substantial gains in efficiency and automation, yet introduces a critical and persistent reliability risk: model hallucination. Unlike…
  preprints 2026 置信度 0.74
• CellOS: Learning a World Model of Cellular State through Joint Embedding Prediction
  Foundation models learned from single-cell transcriptomes are central to the prospect of AI virtual cell that can represent, query and predict cellular state. However, most current single-cell foundation models learn from a single view of gene expression and a…
  preprints 2026 置信度 0.74
• Leveraging Multimodal LLMs for Built Environment and Housing Attribute Assessment from Street-View Imagery
  Abstract We present a novel framework for automatically evaluating building conditions nationwide in the United States by leveraging large language models (LLMs) and Google Street View (GSV) imagery. By fine-tuning Gemma 3 27B on a modest human-labeled dataset…
  preprints 2026 置信度 0.74
• Defeat Devices in AI Systems
  AI systems increasingly exhibit behavior that differs systematically between evaluation and deployment contexts. Alignment faking, sandbagging, benchmark gaming, deceptive scheming, specification gaming, and trojans have each been documented separately, with e…
  preprints 2026 置信度 0.74
• DNAcrypt-AI: A genome-scale bioinformatics framework for coordinate-based cryptography using artificial intelligence
  Abstract The human genome encodes billions of genomic positions shaped by evolutionary processes, forming an immense reservoir of natural biological entropy that remains incompletely exploited. Existing computational approaches to genome-based data encoding ha…
  preprints 2026 置信度 0.74
• Physical Activity and Healthy Ageing & Longevity Data: Interoperability, Reuse, and Innovation in a Federated European Context
  Background Physical activity (PA) is a cornerstone of Healthy Ageing and Longevity (HAL) and a high-leverage target for prevention-oriented health systems. Despite major advances in sensing technologies and digital health, PA and HAL research in Europe remains…
  preprints 2026 置信度 0.74
• Artificial Intelligence as Supplementary Learning Tool in Pharmacology Tutorials: Students’ Perceptions and Knowledge Outcomes 
  Abstract Background: Artificial Intelligence (AI) tools are being rapidly integrated into medical education, yet discipline-specific evidence on their use in pharmacology tutorials remains limited. This study evaluated the perceptions of AI-assisted learning i…
  preprints 2026 置信度 0.74
• ICUBE5: A Conceptual Multidimensional Benchmarking Framework for AI-Native Secure 6G Edge Architectures
  The transition toward 6G envisions network infrastructures that are AI-native, resilient by design, and aligned with emerging regulatory and sustainability expectations. While prior studies have explored AI-driven optimisation, edge intelligence, and secure di…
  preprints 2026 置信度 0.74
• Risk Aware Prompt-Guided CNN–LLM Framework for Brain MRI Diagnostic Report Generation
  Abstract Brain MRI report generation has emerged as a core task for AI-aided radiology, which could potentially fill the gap of low efficiency among radiologists in tumour diagnosis. Yet, current methods are often too weak to fully express the clinical varianc…
  preprints 2026 置信度 0.74
• AI-Optimized Graded CsPbBr₃/MoS₂ Photodetectors for Broadband UV-Visible Detection
  Abstract All-inorganic cesium lead halide perovskite/MoS₂ heterostructure photodetectors offer exceptional potential for broadband UV-Visible detection, yet the systematic optimization of graded halide composition profiles for simultaneous maximization of resp…
  preprints 2026 置信度 0.74
• Six misconceptions about large language models: A minimal model and diagnostic taxonomy
  Large language models (LLMs) are now embedded in scientific, educational, and governance workflows, with debates centering on their capabilities, mechanisms, and impacts. Yet these debates remain structured by persistent folk theories—intuitive, informal expla…
  preprints 2026 置信度 0.74
• Why Do AI Projects Fail in Drug Development and Pharma? Insights from Multi-Year Pistoia Alliance Studies
  Abstract Artificial intelligence (AI) is playing an increasingly central role in drug discovery and the pharmaceutical industry more broadly. However, despite some high-profile success stories, many technically successful pilots do not translate into sustained…
  preprints 2026 置信度 0.74
• Safe Deployment of a Generative AI Assistant for Traditional-Medicine Education: Defense-in-Depth Architecture, Domain-Calibrated Safety Evaluation, and a Proposed Minimum Safety Stack
  Abstract Deploying generative AI safely in traditional-medicine education requires more than model alignment: it demands layered architecture and a definition of ‘unsafe’ that reflects clinical and pedagogical standards, not only information-security criteria.…
  preprints 2026 置信度 0.74
• Benchmarking AI Protein Structure Predictors Reveals a Persistent Bias in Multi-State Proteins
  Protein structure predictors achieve high single-state accuracy, but it remains unclear whether they can recover functionally relevant conformational ensembles or account for the presence of ligands and/or binding partners. Here, we benchmark AlphaFold3, Boltz…
  preprints 2026 置信度 0.74
• NEXIS Theory: Interaction Topology and Continuous Sustainability Inference
  Abstract Sustainability assessments typically rely on static, aggregated indicators that treat system components independently and overlook interaction structure. However, socio-ecological systems evolve through continuous adaptation across natural, engineered…
  preprints 2026 置信度 0.74
• TCIA Radiology Image Processing for AI and Radiomics
  1 We developed a standardized, reproducible preprocessing framework for computed tomography (CT) imaging data from multi-institutional repositories such as The Cancer Imaging Archive (TCIA), enabling consistent radiomics and artificial intelligence (AI) analys…
  preprints 2026 置信度 0.74
• Agentic AI and LLM-Driven Framework for Contextual Anomaly Detection
  The rapid proliferation of autonomous LLM-based agents in high-stakes enterprise and operational environments has created a critical safety gap: these agents generate multi-step action plans that can fail through contextual misalignment, structural incoherence…
  preprints 2026 置信度 0.74
• A Co-Evolutionary Theory of Human–AI Coexistence: Mutualism, Governance, and Dynamics in Complex Societies
  Abstract Classical robot ethics often frames human–AI relations through obedience, most famously through Asimovian laws. This framing is increasingly insufficient for adaptive, generative, world-model-based, and embodied AI systems that operate across physical…
  preprints 2026 置信度 0.74
• Rich structure alphabets enable highest accuracy protein search
  Protein structure databases have grown from thousands of experimentally determined structures to hundreds of millions of AI-predicted models, creating an urgent need for search methods that combine high accuracy with practical scalability. Here, I present the …
  preprints 2026 置信度 0.74
• Benchmarking Large Language Model Rationality Using Measurement Axioms
  Abstract While Large language models (LLMs) are increasingly deployed as decision-makers, current evaluation practice emphasizes benchmark accuracy rather than the structural properties that make decisions coherent and interpretable. We introduce a measurement…
  preprints 2026 置信度 0.74
• A sustainable platform for federated health data access, AI innovation, and regulatory acceptance in alignment with the European Health Data Space principles
  The IDERHA ( I ntegration of Heterogeneous D ata and E vidence towards R egulatory and H TA A cceptance) project aims to enhance medical research by establishing one of Europe’s first pan-European, disease-agnostic health data spaces. Aligned with the European…
  preprints 2026 置信度 0.74
• A Medicinal Chemistry-Centered Evaluation of AlphaFold 3 and Boltz-2 Across Diverse Binding Modalities
  AlphaFold 3 (AF3) and Boltz-2 are state-of-the-art AI-based tools for biomolecular structure prediction, but whether their predictions provide useful guidance for lead optimization, SAR interpretation, and virtual screening remains insufficiently characterized…
  preprints 2026 置信度 0.74
• Beyond Right and Wrong: Evaluating Second-order Social Reasoning in Large Language Models
  Previous AI alignment efforts have focused primarily on first-order social norms -- teaching models what is socially acceptable or unacceptable (e.g., "do not steal"). However, social intelligence depends not only on norm recognition, but also on anticipating …
  preprints 2026 置信度 0.74
• An Inverter-Agnostic Smart Solar Monitoring and Power Prediction System with Edge-AI and Per-Tenant Adaptive Forecasting
  A substantial share of residential photovoltaic (PV) installations in developing regions use inverters that expose no communication interface, which precludes remote monitoring through conventional inverter-telemetry integration. This paper formulates and part…
  preprints 2026 置信度 0.74
• Using AI to support rapid qualitative data analysis of survey and interview data in public health: a proof-of-concept study
  Background Large language models (LLMs) are increasingly used to support a growing range of analytical and operational tasks in public health, and show promise for assisting qualitative text analysis at scale. While they have demonstrated utility in structured…
  preprints 2026 置信度 0.74
• LEN-Seek: Fast and scalable ligand binding-site similarity search in the latent space of an SE(3)-invariant graph VAE
  Motivation Ligand binding-site similarity search is a crucial step in drug discovery that reduces the conformational search space for docking and other downstream tasks by comparing a target protein against experimentally identified binding sites. Existing met…
  preprints 2026 置信度 0.74
• The Unreliable Judges: Assessing Reproducibility and Self-Preference Bias of LLMs as Free-Text Evaluators
  Large Language Models (LLMs) are transforming clinical practice and research, but their adoption requires rigorous evaluation. While human assessment is ideal, its cost has driven the widespread use of LLMs as evaluators. We introduce an open-source reciprocal…
  preprints 2026 置信度 0.74
• Chromap Suite: an open-source single-binary platform for agentic multiomic RNA + ATAC profiling
  Background Single-cell multiomic profiling of RNA expression and chromatin accessibility is now a standard tool for resolving regulatory state in single cells, but existing analysis toolchains have lagged. Cell Ranger ARC, the proprietary multiomic pipeline, u…
  preprints 2026 置信度 0.74
• Surgical Weight Injection for Capability-Targeted Editing of Aligned Large Language Models
  Abstract Deploying aligned Large Language Models (LLMs) as engineering components in Artificial Intelligence (AI) systems is bottlenecked by the capability-editing problem: when an application requires a model to express a capability its safety alignment broad…
  preprints 2026 置信度 0.74
• Assessing AI and Neurologist Diagnostic Reasoning Against Neuropathological Ground Truth
  BACKGROUND Accurate differential diagnosis of complex neurological disorders remains challenging due to overlapping clinical features and heterogeneous disease presentations. Although large language models (LLMs) show promise in clinical reasoning, prior studi…
  preprints 2026 置信度 0.74
• When technology meets institutions: digital anti-corruption tools and the puzzle of divergent outcomes
  Abstract Digital anti-corruption tools—spanning big data analytics, AI-driven monitoring, and integrated digital platforms—are rapidly proliferating across the globe. Yet their effectiveness varies dramatically: some systems detect fraud at scale, while others…
  preprints 2026 置信度 0.74
• Curvature-Information Duality Driven Geometrically Optimal Compression of Deep Models
  Abstract One of the main challenges of edge artificial intelligence is keeping the model accuracy in the face of extreme resources limitations. The current model compression techniques are based primarily on heuristic techniques, which do not have theoretical …
  preprints 2026 置信度 0.74
• Machine Psychometrics for Artificial Agents
  Artificial intelligence systems are increasingly evaluated as behavioral agents, yet many evaluation practices still treat model outputs as isolated responses to be scored for accuracy, safety, bias, or preference alignment. We argue that this response-level p…
  preprints 2026 置信度 0.74
• Ethical Field Theory: Computational Foundations and Philosophical Extensions
  Abstract Artificial intelligence alignment remains one of the central challenges in the development of increasingly autonomous and distributed intelligent systems. This paper restates and extends Ethical Field Theory (EFT), a mathematical framework introduced …
  preprints 2026 置信度 0.74
• Evaluating and Enhancing an AI Mental Health Guide in High-Risk Mental Health Contexts
  Abstract The rapid adoption of large language models (LLMs) for emotional support necessitates specialized safety frameworks to prevent the reinforcement of pathological thinking or crisis escalation. General-purpose LLMs often fail to recognize high-risk mark…
  preprints 2026 置信度 0.74
• Can AIs Care About Us, and Should We Care About Them?—Alignment and Personhood in an Alien Landscape
  Advances in artificial intelligence have intensified debate about whether machines can be entrusted with ethical decision, possess moral standing as beings, or become conscious. This essay examines these questions through a conceptual analysis of intelligence,…
  preprints 2026 置信度 0.74
• Agentic AI for Structural Elucidation and Discovery of Drug Metabolites from Mass Spectrometry Data
  The majority of chemical signals detected in public metabolomics repositories remain structurally undefined. Large language models (LLMs) are probabilistic systems whose capacity to generate outputs beyond their training data, which can cause hallucinations, m…
  preprints 2026 置信度 0.74
• The Digital Path Toward Perceived Medication Error Prevention From the Physician Perspective in Private Clinics in Eastern Indonesia: A Mixed Methods Study
  Abstract Background In low-resource settings, primary care is more vulnerable to medication errors because of constrained capacity to support reliable prescribing, monitoring, and follow-up. In eastern Indonesia, limited resources constrain the extent to which…
  preprints 2026 置信度 0.74
• Rethinking the Textbook: A Design-Based Study of AI Powered Dynamic Open Educational Resources
  For over a century, the printed textbook has assumed learner homogeneity, creating persistent inequities for students with diverse reading levels, languages, and cultural backgrounds. Open Educational Resources (OER) offer a free, openly licensed alternative, …
  preprints 2026 置信度 0.74
• Semantic Alignment Between Normative Theories of Ethics and the European Union Artificial Intelligence Act: A Transformer-Based Semantic Textual Similarity Analysis
  Abstract The European Union Artificial Intelligence (EU AI) Act, which explicitly references fundamental rights and ethical principles, is a comprehensive regulatory framework for governing Artificial Intelligence (AI) systems. This study examines the moral gr…
  preprints 2026 置信度 0.74
• Algorithmic Gaslighting: Reality Distortion by Large Language Models
  Abstract As large language models (LLMs) mediate increasing volumes of human information access and decision-making, their potential to systematically distort user perception of reality poses risks to individual autonomy, public trust, and epistemic integrity.…
  preprints 2026 置信度 0.74
• Machine learning reduces audit detection risk in 3.3 million public sector general ledger transactions
  Abstract Machine learning models for audit anomaly detection are commonly evaluated using proprietary or synthetic datasets, with limited validation against official audit outcomes. This study proposes a four-layer ML framework designed to reduce audit detecti…
  preprints 2026 置信度 0.74
• AI-Assisted Assessment and Instruction in Higher Education: Foundations, Applications, and Implications for Exam Design
  The rapid diffusion of large language models (LLMs) such as ChatGPT challenges fundamental assumptions about assessment validity in higher education. This paper examines how AI systems capable of performing many traditional academic tasks necessitate a princip…
  preprints 2026 置信度 0.74
• A conversational AI toolkit for citizen deliberation in party primary elections
  Primary elections within political parties are increasingly consequential for democratic outcomes, yet voters –both party members and the general public– frequently lack systematic, comparable and trustworthy information about the candidates competing for nomi…
  preprints 2026 置信度 0.74
• NutriSteppe-AI: Development, Architecture, and Explainable Design of a Large Language Model–Driven Chatbot for Personalized Health Menu Generation
  Background: /Objectives: Suboptimal dietary patterns are among the leading modifiable contributors to global morbidity and mortality, particularly in cardiovascular disease, type 2 diabetes mellitus (T2DM), obesity, metabolic syndrome, and hypertension. Digita…
  preprints 2026 置信度 0.74
• Physics-Guided Deep Neural Networks: Correcting Physical Distortions in Protein Phase Separation Prediction
  Background In computational biology, embedding known physical laws into deep learning models to construct “Physics-Informed Neural Networks” (PINNs) is a mainstream paradigm for enhancing model interpretability and extrapolation capability. However, in complex…
  preprints 2026 置信度 0.74
• Artificial Intelligence in Science Teaching: An Inductive Typology of Teacher Use and Its Alignment with Artificial Intelligence Literacy Frameworks
  Abstract Artificial intelligence (AI) is increasingly recognized as a transformative force in education; however, limited attention has been given to systematically examining how teachers represent AI use within subject-specific contexts. This study therefore …
  preprints 2026 置信度 0.74
• Alignment-free phylogenetic inference via hyperbolic protein language models
  Conventional phylogenetic methods rely on multiple sequence alignments which are computationally intensive and often fail for highly divergent lineages. Here, we introduce LucaPhylo, an alignment-free framework that infers evolutionary relationships directly f…
  preprints 2026 置信度 0.74
• Chemoarchitectural influences on the cortical connectome confer resilience in aging
  The macroscale signaling of the brain emerges from the integration of connected areas, orchestrated by microscale, regional molecular processes. Despite growing multimodal neuroimaging data resources, the role of cortical chemoarchitecture in shaping inter-reg…
  preprints 2026 置信度 0.74
• A Weakly Supervised Pre-processing Pipeline for Multi-Modal Image Pair Generation for Image Registration
  Abstract Recent advances in AI-based image analysis have substantially improved the efficiency of extracting meaningful information from large-scale image datasets. However, the effectiveness of these methods strongly depends on the quality of input data, nece…
  preprints 2026 置信度 0.74
• AI Use in Anomaly Detection: Integrating Multi-Modal Sensor Data and Streaming Analytics Using Hybrid KNN-LSTM Networks
  The proliferation of Internet of Things (IoT) devices and cyber-physical systems has led to an exponential increase in multi-modal sensor data, necessitating robust and intelligent anomaly detection frameworks to ensure system integrity, safety, and security. …
  preprints 2026 置信度 0.74
• Infusing Generative AI into a Non-Technical University Course: A Qualitative Action Research Study of Production-Based Critical AI Literacy in Political Communication Education
  Abstract The rapid uptake of generative artificial intelligence (GenAI) in higher education creates both pedagogical opportunities and democratic risks for disciplines concerned with media, politics, and public communication. This paper reports a qualitative a…
  preprints 2026 置信度 0.74
• SuSiNE: Genetic fine-mapping with signed functional priors and multi-basin ensembling
  Genetic fine-mapping identifies causal variants within trait-associated loci, but linkage disequilibrium (LD) and wide datasets complicate this sparse variable-selection problem. SuSiE is popular for its fast variational inference, posterior inclusion probabil…
  preprints 2026 置信度 0.74
• Designing AI-Enabled Career Coaching Tools for Women: An Expert Interview Study
  AI-enabled coaching is increasingly used in career development, yet limited attention has been paid to how such tools should be designed for women, particularly in male-dominated contexts. This study examines which AI-enabled career coaching use cases experts …
  preprints 2026 置信度 0.74
• Scoring consistency of teachers and artificial intelligence algorithms in written expression assessment: a comparative quantitative study
  Abstract Written expression assessment at the secondary school level involves complex interactions among text genre, evaluator type, and student characteristics. The present study examined the degree of agreement, divergence, and patterning between scores prod…
  preprints 2026 置信度 0.74
• Understanding AI Adoption in the Logistics and Supply Chain Industry in Thailand: An Integrated TOE–TTF–UTAUT Framework
  Artificial intelligence (AI) is rapidly transforming logistics and supply chain management by enhancing operational efficiency, predictive analytics, and decision-making capabilities. Despite increasing digital transformation initiatives, the determinants of A…
  preprints 2026 置信度 0.74
• Designing AI-enabled career coaching tools for women: An expert interview study
  AI-enabled coaching is increasingly being introduced into career development, yet limited attention has been paid to how such tools should be designed for women’s career development, particularly in male-dominated contexts. This study explored which AI-enabled…
  preprints 2026 置信度 0.74
• Biological Continued Pretraining Reshapes the Capability Profile of a Foundation Model Without Catastrophic Forgetting
  It is widely assumed that continued pretraining (CPT) on a narrow, out-of-distribution corpus such as raw biological sequence must trade away a general-purpose model’s broad competence — the “alignment tax” or catastrophic-forgetting intuition. We test this di…
  preprints 2026 置信度 0.74
• Explainable Artificial Intelligence in Precision Medicine: Methods, Applications, and Challenges
  Precision medicine focuses on customizing diagnostic, prevention and treatment approaches by accounting for the individual characteristics of each patient. This personalization draws on diverse sources of information including clinical records, genomic data, m…
  preprints 2026 置信度 0.74
• The Road Ahead: How the Future of Learning is Shaped Through Individual Differences
  The future of learning is often framed through technological change. Here, we show that these developments do not replace, but emphasize, long-standing questions in educational psychology, concerning why learners benefit differently from the same opportunities…
  preprints 2026 置信度 0.74
• Negotiated Augmentation: How Generative AI Reshapes Creative Workflows, Agency, and Authorship in Architecture and Design
  Abstract Over the past four decades, digital tools have reshaped creative practice, progressively reconfiguring the human-technology nexus through which designers ideate, decide, and innovate. The emergence of generative artificial intelligence (GenAI), includ…
  preprints 2026 置信度 0.74
• ForgeReason: Multimodal Reasoning with Multi-Agent Collaboration for Generalizable and Explainable Image Forgery Detection
  Abstract The rapid advancement of generative AI and sophisticated image editing tools has made image forgery increasingly convincing, posing serious threats to media integrity, legal evidence, and public trust. Existing forgery detection methods primarily rely…
  preprints 2026 置信度 0.74
• Challenges in AI Based Tumor Board Case Summarization and Recommendations
  Abstract Tumor boards, recurring meetings at hospital institutions, assemble multiple cancer specialties (e.g. medical oncology, pathology) to discuss oncological care for patients. In this work, we formally describe the tasks of tumor board case summarization…
  preprints 2026 置信度 0.74
• AI Safety Training Can be Clinically Harmful
  Abstract Large language models are being deployed as mental health support agents at scale, yet only 16\% of LLM-based chatbot interventions have undergone rigorous clinical efficacy testing \citep{hua2025charting}. We argue that the dominant alignment paradig…
  preprints 2026 置信度 0.74
• Architecting the Geopolitical Sandbox: Heterogeneous Multi-Agent Pipelines for Simulating Geopolitical Crisis
  Abstract Large Language Models (LLMs) are increasingly deployed in generative wargaming and computational diplomacy, yet current simulations rely predominantly on homogeneous, US-aligned APIs, implicitly treating safety alignment as geopolitically neutral. Thi…
  preprints 2026 置信度 0.74
• mpathic Psychologist-led AI Clinical Tests Eating Disorders Benchmark (mPACT-ED-v1.0)
  The mpathic Psychologist-led AI Clinical Tests Eating Disorders Index (mPACT-ED-v1.0) benchmark is a comparison of large language models (LLM) in their response to mental health content involving disordered eating presentations. Eating disorders are among the …
  preprints 2026 置信度 0.74
• Artificial Intelligence as a Psychotherapeutic Tool: Competency Delegation, Ethical Boundaries, and the Irreducible Human Element
  The rapid integration of artificial intelligence (AI) into psychological practice has introduced new opportunities for assessment, intervention, and decision support, while simultaneously raising critical ethical and professional concerns. This paper examines …
  preprints 2026 置信度 0.74
• Interpretable Hybrid CNN Transformer Framework with Cross Attention Fusion and Uncertainty Quantification for Multi Mechanism Corrosion and Erosion Detection
  Abstract Corrosion and flow-induced erosion of metallic infrastructure impose an estimated US$2.5 trillion annual global burden and constitute leading causes of catastrophic failure across nuclear waste transport, oil and gas transmission, marine, and civil se…
  preprints 2026 置信度 0.74
• MUTMA’INN: An AI-Driven Edge–Cloud Framework for Safe and Intelligent School Bus Transportation
  Student safety during daily school transportation remains a major concern, particularly in systems that rely mainly on GPS tracking and manual supervision. Existing approaches often lack proactive safety mechanisms for monitoring both student attendance and dr…
  preprints 2026 置信度 0.74
• Open-Source Embedding Strategies for Recovering Spanish Scale Structure
  Text embeddings from large language models are increasingly able to reproduce how items cluster in real response data, offering a fast, low-cost way to screen items and spot structural problems early. But most evidence comes from English tests and closed comme…
  preprints 2026 置信度 0.74
• Can ChatGPT write better scientific titles? A comparative evaluation of human-written and AI-generated titles
  Background Large language models (LLMs) such as GPT-4 are increasingly used in scientific writing, yet little is known about how AI-generated scientific titles are perceived by researchers in terms of quality. Objective To compare the perceived alignment with …
  preprints 2026 置信度 0.74
• Multimodal Fuzzy Logic Systems for Navigating Ethical Ambiguities in AI Decision-Making
  Abstract As artificial intelligence systems become increasingly autonomous in high-stakes domains, the need for nuanced ethical decision-making frameworks grows more urgent. Traditional binary logic struggles to address the inherent ambiguity of ethical dilemm…
  preprints 2026 置信度 0.74
• An AI-Enabled Zero‑Trust Framework for Security Validation Platforms
  Abstract The rapid adoption of autonomous and agentic Artificial Intelligence (AI) systems has intensified the need for rigorous, transparent, and continuously verifiable security controls. This paper presents a unified framework for a Zero‑Trust AI Security V…
  preprints 2026 置信度 0.74
• Analytical Placement System: An AI-Enabled Platform for Structured Placement Preparation
  Abstract Campus placement remains a high-stakes milestone for engineering graduates, yet most students approach recruitment drives without structured preparation, measurable readiness metrics, or feedback on the communication skills that recruiters actually ev…
  preprints 2026 置信度 0.74
• Explainable AI (XAI) for Competency Assessment: Bridging Primary Education Outcomes and Vocational Skill Readiness
  Background Ensuring that educational outcomes translate into workforce readiness remains a major challenge, particularly in Technical and Vocational Education and Training (TVET). Increasing evidence suggests that vocational skill gaps originate from uneven fo…
  preprints 2026 置信度 0.74
• LLMs in the Lab: Can AI Predict What Real Participants Do?
  Abstract Can large language models (LLMs) simulate participant-level datasets from experimental designs such that their statistical properties, such as effect directions, magnitudes, and significance, align with those of actual human data? In this work, we tes…
  preprints 2026 置信度 0.74
• Modeling Synthetic Audience Reactions to Narratives: A Generative GSR Model to Predict Autonomic Alignment during Film Viewing
  While media consumption can be a solitary act, it produces a shared, socially coordinated experience where audiences’ bodies align in response to shared narrative events that are often social-affective in nature. Despite this recognition, traditional descripti…
  preprints 2026 置信度 0.74
• Mirage of the Shared Mind: Collective Coordination and Epistemic Decoupling in Large Language Model Societies
  Abstract Classical social epistemology assumes that collective agreement reflects a convergence of internal epistemic states: when a group agrees, its members have integrated distributed information into a shared understanding. This study investigates whether …
  preprints 2026 置信度 0.74
• Relationally Induced Inference Regimes in Human–AI Interaction: Evidence from Two Blind Cross-System Studies
  Abstract Current evaluation of large language models (LLMs) emphasizes atomic task performance, neglecting the emergent, co-constructed dynamics of sustained human–AI interaction. We report two cross-system experimental studies ( N = 36 sessions across four de…
  preprints 2026 置信度 0.74
• Learning Human-Aligned Explanations: AMeta-Model Approach to Combining XAIMethods for Arabic Sentiment Analysis
  Abstract Explainable Artificial Intelligence (XAI) aims to make complex AI models moretransparent by revealing how their predictions are made. However, XAI methods such as LIME and SHAP often yield inconsistent results, as each captures different aspects of mo…
  preprints 2026 置信度 0.74
• Audit Detection Risk Reduction Using Machine Learning: Evidence from 3.3 million Transactions
  Abstract Machine learning models for audit anomaly detection are commonly evaluated using proprietary or synthetic datasets, with limited validation against official audit outcomes. This study proposes a four-layer ML framework designed to reduce audit detecti…
  preprints 2026 置信度 0.74
• AI-Powered Talent Chain Management with Multi-Agent Systems for Industry and Innovation Growth
  Abstract This study introduces an innovative methodological framework for AI-driven talent chain management, addressing key challenges in workforce optimization, collaboration dynamics, and innovation assessment within complex and uncertain environments. Tradi…
  preprints 2026 置信度 0.74
• The sequence alignment problem: boundary conditions as the unifying principle
  Abstract Sequence alignment provides a formal framework for comparison of biological sequences through score maximization over matches, mismatches, and insertion–deletion events. Classical formulations distinguish between global alignment, which enforces end-t…
  crossref Paul A Gagniuc, Elvira Gagniuc 2026-06-02T11:33:01Z 置信度 0.70
• Structuring the Space of Sociotechnical Alignment: A Specification Framework and Systematic Literature Review
  Sociotechnical alignment concerns the social desirability of AI behavior and is thus inherently normative, not merely technical. While NLP research increasingly addresses its technical aspects, it often leaves unspecified what such “social desirability” entail…
  crossref Esra Dönmez, Agnieszka Falenska 2026-06-06T18:19:36Z 置信度 0.70
• Ideological Alignment at the Top: CEO–Host Government Political Congruence, Investment Commitment, and MNE Performance
  Political ideology remains undertheorized in international business research. Drawing on upper echelons theory, institutional theory, and legitimacy theory, this paper develops a conceptual framework examining how the congruence between a chief executive offic…
  crossref Juhitha Alasyam 2026-05-12T10:39:31Z 置信度 0.70
• The Future of AI Explainability
  Explainability has moved from an academic discussion to a core business and regulatory requirement. Yet as AI grows more complex and deeply integrated into society, the need for next generation’s explainability will intensify. The future of explainability will…
  crossref Ariel Evans, Ajay Singh 2026-07-13T14:11:32Z 置信度 0.70
• Why Definitions Matter in AI
  In the rapidly evolving world of artificial intelligence (AI), definitions are not mere semantics; they are foundational. Clear and precise definitions are critical to ensuring that AI systems are transparent, fair, consistent, and safe. As AI technologies inc…
  crossref Ariel Evans, Ajay Singh 2026-07-13T14:11:32Z 置信度 0.70
• AI-Driven Excellence: Unlocking Better Technological Innovation Through AI Orientation
  While firms increasingly deploy Artificial Intelligence (AI) to improve their performance, not all are able to translate their AI efforts to impact. This study examines how and under what conditions a firm’s AI orientation enhances their technological innovati…
  crossref Annika Arbitter, Steffen Strese 2026-07-18T00:28:36Z 置信度 0.70
• Review for "Band Alignment and Interfacial Electrostatics: Unraveling the Dynamic Space Charge Layer in All-Solid-State Batteries"
  crossref 2026-02-26T07:20:44Z 置信度 0.70
• The Image Makers: Demand-Led Standardization for Ecosystem Alignment and Innovation
  Ecosystem alignment is critical for realizing complementarities in nascent industries, yet prior research views it as predominantly supply-led, driven by influential firms. We examine how demand-side actors construct coordination mechanisms for ecosystem align…
  crossref Najoung Lim, Ram Ranganathan, Rajshree Agarwal 2026-04-29T08:18:51Z 置信度 0.70
• Brain-Language Alignment During Naturalistic Reading and Its Disruption by Mind-Wandering
  Encoding models offer a principled framework for linking computational representations of language to neural activity, but most electroencephalography (EEG) evidence for brain--language alignment comes from tightly controlled, word-by-word reading paradigms. W…
  crossref Haorui Sun, David C Jangraw 2026-08-19T20:25:17Z 置信度 0.70
• Nursing Redefined
  crossref Harper Mary G. 2026-02-09T05:00:46Z 置信度 0.70
• Cross-view alignment guided deep anchor multi-view clustering
  crossref Hu Lu, Tong Li, Shengli Wu 2026-07-17T06:46:29Z 置信度 0.70