AI 安全与对齐 · 全部实体（第 21 / 100 页）

本页收录第 2001–2100 条，共 10000 条 · 返回 AI 安全与对齐 首页

• Tracing AI Assistance and AI Agents in Survey Research
  Generative AI creates a measurement problem for survey research. When respondents can outsource summarization, explanation, or fact retrieval, open-ended answers may no longer reflect their own reasoning. We introduce an auditable toolkit that combines respons…
  crossref Valentina Gonzalez-Rostani, Shir Raviv 2026-04-28T12:56:13Z 置信度 0.70
• Predicting the success of transfer learning for genetic programming using DeepInsight feature space alignment
  In Transfer Learning (TL) a model that is trained on one problem is used to simplify the learning process on a second problem. TL has achieved impressive results for Deep Learning, but has been scarcely studied in genetic programming (GP). Moreover, predicting…
  crossref Leonardo Trujillo, Joel Nation, Luis Muñoz, Edgar Galván 2023-07-28T12:08:21Z 置信度 0.70
• A Global Peak Alignment Approach for AI-ready Geochemical Datasets: Standardizing Historical Petroleum GC-MS Data
  Summary Unlocking the potential of historical Gas Chromatography-Mass Spectrometry (GC-MS) datasets for AI-driven insights has remained a challenge due to decades of variability in data acquisition. This study presents a transformative Al-based methodology to …
  crossref M. Conceição, W. Bastos, C. De Oliveira, J. Guzzo 2025-02-26T06:06:17Z 置信度 0.70
• Assessing AI-Generated Questions’ Alignment with Cognitive Frameworks in Educational Assessment
  This study evaluates the integration of Bloom’s Taxonomy into OneClickQuiz, an Artificial Intelligence (AI) driven plugin for automating Multiple-Choice Question (MCQ) generation in Moodle. Bloom’s Taxonomy provides a structured framework for categorizing educ…
  crossref Antoun Yaacoub, Jérôme Da-Rugna, Zainab Assaghir 2025-07-10T01:35:13Z 置信度 0.70
• Best Practices for AI Storytelling
  In previous chapters, we’ve explored the possibilities and pitfalls of using generative AI in narrative game design by examining everything from coherence issues to ethical blind spots. But while the challenges are significant, so too is the opportunity. When …
  crossref Shesh Narayan Gupta 2026-03-31T22:21:23Z 置信度 0.70
• Design for <scp>AI</scp>
  This chapter explores AI-first design, a strategic approach to integrating AI into organizational processes and products. It challenges the misconception that AI-first design prioritizes AI over human labour, instead presenting it as a mindset for solving prob…
  crossref 2026-03-11T12:39:07Z 置信度 0.70
• Front Matter: Volume 14302
  crossref 2026-07-17T18:37:12Z 置信度 0.70
• Beyond AI automation: Redesigning organizations for human-AI synergy
  crossref Vlad Krotov 2025-07-02T01:50:22Z 置信度 0.70
• Pengaruh Beban Kerja dan Lingkungan Kerja terhadapKinerja Guru (Studi Kasus SDN Kecamatan Bandar Kabupaten Bener Meriah)
  ABSTRACT This study aims to analyze the influence of workload and work environment on teacher performance at a public elementary school in Bandar District, Bener Meriah Regency. The research method used was a quantitative approach with multiple linear regressi…
  crossref Maisyarah Maisyarah, Tri Purwani 2026-02-28T05:08:51Z 置信度 0.70
• Rapid and High-Precision Automatic Alignment and Splicing Techniques for Multi-Core Fibers
  We developed two multicore fiber (MCF) automatic alignment techniques by processing the fiber end-face imaging, achieving fast and high-precision MCF alignment with less than 20 seconds and 0.1°accuracy, enabling passive low-loss MCF splicing.
  crossref Weiao Liu, Zhiyong Zhao, Can Chen, Ming Tang 2026-07-15T11:05:25Z 置信度 0.70
• Architectures of Global AI Governance: From Technological Change to Human Choice by Matthijs M. Maas
  crossref Dogan Sengul 2026-03-09T09:56:34Z 置信度 0.70
• Integrating Trustworthiness Into the AI Lifecycle (Trustworthiness)
  With the increasing impact on society of intelligent technology, trust is becoming more and more crucial in public institutions, healthcare, education, and finance as new systems are built to make or support human decisions. People will need to have faith in t…
  crossref Rajesh Lingam 2026-02-20T15:27:43Z 置信度 0.70
• Governing the AI-mediated corporation: corporate governance and the limits of detached AI governance
  crossref Kei Nakagawa 2026-08-26T10:01:10Z 置信度 0.70
• AI-Enabled Data Stewardship Real-Time Alignment of Privacy and Storage Policies Across Global Systems using Deep CNN-RNN Techniques
  In today’s data-centric world, organizations have multi geographic regulatory regimes to manage and complex requirements to timespan privacy and storage policies like GDPR, HIPAA, and regional data protection laws. In this paper propose a new AI-based model fo…
  crossref Saisuman Singamsetty 2025-12-16T18:29:19Z 置信度 0.70
• Formation Mechanism of Continuance Intention to Use Generative AI: The Role of Perceived Competence, Perceived Warmth, and AI Dependency
  crossref Yuna Kim, Youngseok Son 2026-07-21T01:47:01Z 置信度 0.70
• Correction: AIgemony: power dynamics, dominant narratives, and colonisation
  crossref Nasser Bahrami 2026-03-04T01:36:48Z 置信度 0.70
• Hallucinating with AI: Distributed Delusions and “AI Psychosis”
  Abstract There is much discussion of the false outputs that generative AI systems such as ChatGPT, Claude, Gemini, DeepSeek, and Grok create. In popular terminology, these have been dubbed “AI hallucinations”. However, deeming these AI outputs “hallucinations”…
  crossref Lucy Osler 2026-02-11T11:14:20Z 置信度 0.70
• Facing the pain: ethical considerations of AI-based pain detection of farmed animals
  Abstract Automated pain detection (APD) is an emerging technology that runs on artificial intelligence (AI) (e.g., machine learning, computer vision, and deep learning) and is aimed at detecting pain in farmed animals. One type of APD analyses the facial expre…
  crossref Mark Ryan, Leonie Bossert 2026-01-13T17:35:43Z 置信度 0.70
• The Infrastructure of AI
  crossref 2026-05-13T20:24:56Z 置信度 0.70
• Maia-2: A Unified Model for Human-AI Alignment in Chess
  crossref Zhenwei Tang, Difan Jiao, Reid McIlroy-Young, Jon Kleinberg 等 2025-11-03T11:20:56Z 置信度 0.70
• Are conscious machines valuers?
  Abstract Recent proposals in neuroscience and philosophy suggest that coarse-grained computational functionalism may suffice for artificial consciousness. However, I argue that even if such accounts are right, there is no reason to assume that consciousness so…
  crossref Jan Henrik Wasserziehr 2026-03-08T19:19:23Z 置信度 0.70
• ACM TechBrief: Agentic AI
  crossref Rohith Nama, Larry Medsker, Brian Peretti 2026-06-09T20:26:03Z 置信度 0.70
• AI and Language in the Urban Context: Conversational Artificial Intelligence in Cities
  crossref Ratan Sarkar, Yeasmin Sultana, Namrata Gogoi 2026-01-06T14:05:10Z 置信度 0.70
• GoLove AI Review 2026 – Is It Scam or Legit? (Update 2026)
  crossref Mark Cotta 2026-03-05T12:21:32Z 置信度 0.70
• Fictional prototypes of AI–human coexistence and relationality
  This study examines how fictional portrayals of artificial intelligence (AI) shape cultural imagination and readiness for AI–human coexistence. Drawing on 42 works ranging from ancient mythology to contemporary film, anime, and games, it analyzes AI characters…
  crossref Junichi Hoshino 2026-01-14T10:57:08Z 置信度 0.70
• How AI is rewiring the human brain: the generational transformation of cognition and knowing
  Abstract This Open Forum paper examines how artificial intelligence (AI) is transforming not only what humans know but how knowledge itself is constructed, remembered and valued. It argues that AI has evolved from a tool of efficiency into an epistemic infrast…
  crossref Hans Westerbeek 2026-02-12T17:31:03Z 置信度 0.70
• Operationalising ethical AI governance: a triadic framework for the European AI act
  crossref Joffrey Baeyaert 2025-12-01T03:20:52Z 置信度 0.70
• The intelligence we need now
  crossref Douglas Schuler 2026-06-09T01:51:29Z 置信度 0.70
• Risk Management with Agentic AI Solutions in Procurement
  crossref Bernardo Nicoletti 2026-05-21T21:03:49Z 置信度 0.70
• Shaping the Future of Farming with Generative AI
  crossref Basudha Dewan 2026-02-13T10:14:20Z 置信度 0.70
• AI
  The integration of artificial intelligence (AI) into environmental innovation represents a pivotal advancement in addressing global sustainability challenges. The chapter contributes by presenting a comprehensive analysis of AI applications in environmental mo…
  crossref Hardeep Singh, Bikram Pal Singh 2026-07-01T12:37:20Z 置信度 0.70
• From Human to AI: How Delegation to AI Shifts Moral Responsibility Attribution
  As AI systems increasingly make moral decisions on behalf of humans, questions ariseabout how this delegation reshapes moral responsibility. Yet little is known aboutwhether people hold AI systems responsible for delegated moral decisions, and whetherthis diff…
  crossref Philippe Sloksnath 2026-06-05T06:53:36Z 置信度 0.70
• Dark Brand: Measuring AI-Assembled Brand Image Accuracy Across Chinese AI Platforms
  This paper introduces the concept of dark brand, defined as the brand image that AI systems automatically assemble from publicly available internet content and deliver to consumers without brand-owner participation or review, and presents the first large-scale…
  crossref Andy Yang 2026-07-09T03:51:13Z 置信度 0.70
• AI Governance
  AI Governance: The Executive Handbook for Safe and Responsible AI is a book researched and written by cyber experts Ariel Evans and Ajay Singh with contributions from leading members of Cyber Intelligence 4U advisory board, including top CISOs, CIOs, CTOs, DPO…
  crossref Ariel Evans, Ajay Singh 2026-07-13T14:11:32Z 置信度 0.70
• The AI License Model A Conceptual Framework for Distributed AI Productivity Rights
  This paper presents a conceptual framework for the democratic distribution of AI deployment capacity through a tradable licensing mechanism. The model addresses the structural asymmetry between concentrated productivity gains and distributed labor displacement…
  crossref Ivo Ferrari 2026-04-20T15:47:32Z 置信度 0.70
• Can AI Be Contextualized?
  Responsible Artificial Intelligence (RAI) principles are informed by a commitment to attending to local context when designing and deploying Artificial Intelligence (AI) systems. Yet, Artificial Intelligence (AI) systems are developed and deployed across distr…
  crossref Glen Berman, James Smithies 2026-07-24T00:05:33Z 置信度 0.70
• Creating Characters Using Components
  crossref John Reynolds 2026-07-09T20:46:49Z 置信度 0.70
• Leveraging Artificial Intelligence to Improve Perioperative Staffing Consistency: A Quality Improvement Initiative at a Large Academic Medical Center.
  A large academic medical center in the Pacific Northwest addressed perioperative staffing challenges by implementing a workflow with application of AI assistance to optimize daily assignments. The initiative integrated real-time and historical data to support …
  europepmc 2026 置信度 0.80
• Assessing Guideline Knowledge Alignment of Large Language Models in Ophthalmology: A Preclinical Benchmarking Study on KLEx Evidence-Based Guidelines.
  Purpose To evaluate the guideline knowledge alignment of two large language models (LLMs), GPT-5.5 Instant and DeepSeek-V4, and to determine their preclinical reliability as reference tools in refractive surgery. Methods Using the 38 evidence-based recommendat…
  europepmc 2026 置信度 0.80
• Erratum: Agentic AI as a coordination paradigm in digital health and agri-food systems.
  [This corrects the article DOI: 10.1016/j.patter.2026.101496.].
  europepmc 2026 置信度 0.80
• DARE: An Explainable AI-Visualization Framework for Ill-Defined Decision Making.
  Real-world decision making often unfolds in fluid, uncertain, and ill-defined contexts where objectives shift, data are incomplete, and non-quantifiable factors such as social values, ethics, and institutional constraints play critical roles. Conventional AI a…
  europepmc 2026 置信度 0.80
• Emotional support or problem-solving? A randomized controlled trial of function-specific AI Chatbots for digital stress in college students.
  Digital stress has emerged as a salient risk factor for the mental health of emerging adults, yet targeted interventions remain limited. Generative AI chatbots offer a scalable approach to delivering context-sensitive supportive interactions that users may per…
  europepmc 2026 置信度 0.80
• Development and psychometric validation of the AI chatbots acceptance and perception scale for higher education students.
  The rapid integration of artificial intelligence (AI) chatbots in higher education has transformed student learning experiences and institutional operations, yet comprehensive measurement tools for assessing student usage patterns remain limited. This study pr…
  europepmc 2026 置信度 0.80
• Is it safe to deploy AI in safety-critical systems?
  We consider the reality of deploying artificial intelligence (AI) in safety-critical systems, such as autonomous vehicles, medical diagnoses and weather forecasting. Our discussion is grounded in the mathematical nature of AI systems, including how an AI's mat…
  europepmc 2026 置信度 0.80
• A multi-agent vision-language debate framework for zero-shot crop disease diagnosis.
  Accurate crop disease diagnosis is critical for agricultural productivity and food security, yet existing deep learning systems often struggle to generalize across visually similar diseases and varying environmental conditions. Recent Vision-Language Models (V…
  europepmc 2026 置信度 0.80
• Discordance Between Textual Reasoning and Visual Interpretation in Large Language Models for Low Back Pain: Cross-Sectional Quantitative Evaluation and Exploratory Multimodal Stress Test.
  Background Large language models (LLMs) are rapidly evolving from text-based agents to multimodal systems capable of interpreting medical images. While their textual reasoning has improved, the safety implications of this shift remain underexplored, specifical…
  europepmc 2026 置信度 0.80
• Connecting Europe’s XR Ecosystem: XR5.0 and XR2Industry Join Forces
  This blog presents ongoing collaboration between the Horizon Europe projects XR5.0 and XR2Industry. The objective of this collaboration is to ensure interoperability and alignment of XR5.0 technologies with emerging European XR infrastructures, supporting a su…
  europepmc 2026 置信度 0.80
• Evaluating the cultural alignment of multilingual LLMs in typical Japanese workplace scenarios.
  While current evaluations of LLM cultural alignment predominantly rely on static benchmarks in Western contexts, their ability to navigate generative, high-context socio-pragmatic demands in non-Western environments remains critically underexplored. This study…
  europepmc 2026 置信度 0.80
• Algorithmic scholarship and academic evaluation: governance misalignment in the age of generative AI.
  Introduction Generative artificial intelligence (AI), particularly large language models, is rapidly becoming embedded in academic research and scholarly publishing. These systems assist with drafting, literature synthesis, and analytical writing, increasingly…
  europepmc 2026 置信度 0.80
• A projection-domain deep learning approach for respiratory motion correction of myocardial perfusion imaging using a multi-pinhole solid-state SPECT.
  Background Myocardial perfusion imagining (MPI) is a nuclear medicine technique used in the assessment of coronary artery disease. Differences in the perfusion of the myocardium at rest and stress and regions of perfusion defects can be indicators of various c…
  europepmc 2026 置信度 0.80
• Mapping the movie-watching brain with AI-derived semantics.
  Naturalistic paradigms offer a powerful tool to investigate human brain function, but it remains difficult to link rich, continuous movie content to distributed brain activity in an interpretable way. In this study, I use a multimodal large language model (Gem…
  europepmc 2026 置信度 0.80
• Model and Task-Aware Test-Time Scaling Strategies for Large Language and Vision-Language Models in Medicine: Evaluation Study.
  Background Test-time scaling has emerged as a promising method to enhance the reasoning capabilities of large language models (LLMs) and vision-language models (VLMs) during inference without additional training. While foundational studies established scaling …
  europepmc 2026 置信度 0.80
• Readying Military Medicine for AI-Enabled Warfare.
  Artificial intelligence (AI) is transforming the character of warfare through autonomous systems, real-time data analytics, and algorithmic decision-making, creating new operational, ethical, and clinical challenges for military medicine. While modern battlefi…
  europepmc 2026 置信度 0.80
• AI-enhanced framework for optimizing CRISPR-Cas gene editing in crop biotechnology addressing regulatory challenges and opportunities in global agricultural practices.
  Introduction The integration of CRISPR Cas genome editing with artificial intelligence (AI) offers significant potential for crop biotechnology by supporting more precise and adaptive strategies for trait improvement under complex agricultural and regulatory c…
  europepmc 2026 置信度 0.80
• The Lie Awareness Spectrum: How Large Language Models Recognize, Rationalize, and Refuse Deception
  Abstract As large language models (LLMs) become embedded in high-stakes decision-making, understanding their capacity for deception—and awareness of it—is critical. We conducted a systematic empirical study comparing GPT-5.4, Claude 4.6 Sonnet, and Gemini 3.1 …
  europepmc 2026 置信度 0.80
• AI and 3D printing-assisted surgical exposure of impacted canines and tooth extraction in complex cleft lip and palate orthodontic treatment.
  Impacted maxillary canines are common in patients with cleft lip and palate (CLP) due to altered anatomy and scar tissue formation, making surgical exposure challenging. This case report describes the use of artificial intelligence (AI)-assisted CBCT segmentat…
  europepmc 2026 置信度 0.80
• Reflections and Recommendations on AI Adoption Practice from a Mixed-Ability Research Group
  Generative AI tools have recently been rapidly adopted by academics in mixed-ability research teams for both personal and professional tasks. While previous work on adoption of AI-based workflows has focused on collaboration and productivity, the perceptions o…
  arxiv Shalini Madan, Sreelakshmi Surabiyil Bindu, Veronica Pimenova, Ellie Seehorn 等 2026-07-24T19:53:08Z 置信度 0.78 cs.HC
• Beyond the Desk: Barriers and Future Opportunities for AI to Assist Scientists in Embodied Physical Tasks
  More scientists are now using AI, but prior studies have examined only how they use it 'at the desk' for computer-based work. However, given that scientific work often happens 'beyond the desk' at lab and field sites, we conducted the first study of how scient…
  arxiv Irene Hou, Alexander Qin, Lauren Cheng, Philip J. Guo 2026-03-19T22:18:11Z 置信度 0.78 cs.HCcs.AI
• Embodied AI Agents for Team Collaboration in Co-located Blue-Collar Work
  Blue-collar work is often highly collaborative, embodied, and situated in shared physical environments, yet most research on collaborative AI has focused on white-collar work. This position paper explores how the embodied nature of AI agents can support team c…
  arxiv Kaisa Vaananen, Niels van Berkel, Donald McMillan, Thomas Olsson 2026-02-12T16:22:18Z 置信度 0.78 cs.HC
• Brain-LLM Alignment Tracks Training Data, Not Typology
  Brain-LLM alignment is well established in English, yet the brain's language network is neuroanatomically universal across languages. Does alignment also generalize cross-linguistically, and what governs the variation? We test this using fMRI data from 112 par…
  arxiv Dongxin Guo, Jikun Wu, Siu Ming Yiu 2026-05-21T20:56:51Z 置信度 0.78 cs.CLcs.AIq-bio.NC
• Intra-Fairness Dynamics: The Bias Spillover Effect in Targeted LLM Alignment
  Conventional large language model (LLM) fairness alignment largely focuses on mitigating bias along single sensitive attributes, overlooking fairness as an inherently multidimensional and context-specific value. This approach risks creating systems that achiev…
  arxiv Eva Paraschou, Line Harder Clemmensen, Sneha Das 2026-02-18T13:19:11Z 置信度 0.78 cs.LGcs.AI
• The Geometry of Refusal: Linear Instability in Safety-Aligned LLMs
  Modern Large Language Models (LLMs) rely on extensive safety alignment, yet the mechanistic basis of refusal remains opaque. In this work, we investigate whether safety compliance is a deep semantic decision or a manipulable linear feature. We introduce Contra…
  arxiv Shivam Ratnakar, Kartikeya Vats 2026-06-21T22:04:48Z 置信度 0.78 cs.CRcs.AIcs.LG
• Code Review is a Conversation: Toward Conversational AI Review Assistants
  AI-based code review tools increasingly promise to help developers inspect pull requests, identify defects, and improve code quality. Yet most current approaches frame code review as a one-shot commenting task: given a diff, the system produces warnings or sug…
  arxiv Rosalia Tufano 2026-07-24T08:44:05Z 置信度 0.78 cs.SE
• The xPU-athalon: Quantifying the Competition of AI Acceleration
  The push for greater efficiency in AI computation has given rise to an array of accelerator architectures that increasingly challenge the GPU's long-standing dominance. In this work, we provide a quantitative view of this evolving landscape of AI accelerators,…
  arxiv Alicia Golden, Carole-Jean Wu, Gu-Yeon Wei, David Brooks 2026-04-12T23:10:14Z 置信度 0.78 cs.AR
• Value-Action Alignment in Large Language Models under Privacy-Prosocial Conflict
  Large language models (LLMs) are increasingly used to simulate decision-making tasks involving personal data sharing, where privacy concerns and prosocial motivations can push choices in opposite directions. Existing evaluations often measure privacy-related a…
  arxiv Guanyu Chen, Chenxiao Yu, Xiyang Hu 2026-01-07T03:30:42Z 置信度 0.78 cs.CLcs.AIcs.HCcs.LG
• Bounded-Compute Multimodal Regression for Product-Rating Prediction
  Vision-language models (VLMs) are increasingly attractive for multimodal quality assessment, but their default reliance on autoregressive text generation and dynamic visual processing is poorly matched to scalar regression under strict latency budgets. We pres…
  arxiv William Leach, Ru He, Sizhuo Ma, Yizhen Jia 等 2026-05-26T22:27:58Z 置信度 0.78 cs.CV
• Offline Reinforcement Learning of High-Quality Behaviors Under Robust Style Alignment
  We study offline reinforcement learning of style-conditioned policies using explicit style supervision via subtrajectory labeling functions. In this setting, aligning style with high task performance is particularly challenging due to distribution shift and in…
  arxiv Mathieu Petitbois, Rémy Portelas, Sylvain Lamprier 2026-01-30T10:49:22Z 置信度 0.78 cs.LGcs.AIcs.RO
• The Specification as Quality Gate: Three Hypotheses on AI-Assisted Code Review
  The dominant industry response to AI-generated code quality problems is to deploy AI reviewers. This paper argues that this response is structurally circular when executable specifications are absent: without an external reference, both the generating agent an…
  arxiv Christo Zietsman 2026-03-26T11:59:05Z 置信度 0.78 cs.SE
• Natural Language Processing in Health Professions Education: A Scoping Review
  Natural language processing (NLP) and artificial intelligence (AI) are rapidly transforming health professions education, yet no scoping review has systematically mapped their applications across the full spectrum of health education contexts, including public…
  arxiv Javad Mohammad Alizadeh, Pegah Saebi, Mukesh Kumar Patel, Huanmei Wu 2026-05-15T21:17:17Z 置信度 0.78 cs.HC
• Lifting to tensors when compiling scientific computing workloads for AI Engines
  It has been demonstrated that specialised architectures, such as FPGAs and AMD's AI Engines (AIEs), have the potential to deliver energy and performance advantages for scientific computing. Given the integration of AIEs into AMD's CPUs, this is an interesting …
  arxiv Nick Brown, Gabriel Rodriguez-Canal 2026-05-05T09:40:01Z 置信度 0.78 cs.DC
• SETA: Statistical Fault Attribution for Compound AI Systems
  Modern AI systems increasingly comprise multiple interconnected neural networks to tackle complex inference tasks. Testing such systems for robustness and safety entails significant challenges. Current state-of-the-art robustness testing techniques, whether bl…
  arxiv Sayak Chowdhury, Meenakshi D'Souza 2026-01-27T08:21:28Z 置信度 0.78 cs.AIcs.LGcs.SE
• PLAF: Pixel-wise Language-Aligned Feature Extraction for Efficient 3D Scene Understanding
  Accurate open-vocabulary 3D scene understanding requires semantic representations that are both language-aligned and spatially precise at the pixel level, while remaining scalable when lifted to 3D space. However, existing representations struggle to jointly s…
  arxiv Junjie Wen, Junlin He, Fei Ma, Jinqiang Cui 2026-04-17T07:24:14Z 置信度 0.78 cs.CVcs.RO
• Context-Guided Semantic Alignment for Feature Fusion Networks
  Feature fusion networks are fundamental components in modern object detectors, aggregating multi-scale features to detect objects of varying sizes. However, directly fusing features from different pyramid levels often introduces semantic inconsistency, causing…
  arxiv Hyungseop Lee, Jiho Lee, Woochul Kang 2026-06-12T00:54:11Z 置信度 0.78 cs.CV
• Differentiable Optimization Layers for Guaranteed Fairness in Deep Learning
  Differentiable optimization layers are traditionally integrated in predict-then-optimize frameworks where a neural model estimates parameters that subsequently serve as fixed inputs to downstream decision-making optimization problems. In this work, we introduc…
  arxiv David Troxell, Noah Roemer, Guido Montúfar 2026-05-16T18:49:08Z 置信度 0.78 cs.LGstat.COstat.ML
• Decoding AI Authorship: Can LLMs Truly Mimic Human Style Across Literature and Politics?
  Amidst the rising capabilities of generative AI to mimic specific human styles, this study investigates the ability of state-of-the-art large language models (LLMs), including GPT-4o, Gemini 1.5 Pro, and Claude Sonnet 3.5, to emulate the authorial signatures o…
  arxiv Nasser A Alsadhan 2026-03-24T13:58:09Z 置信度 0.78 cs.CLcs.LG
• Relevance as a Vulnerability: How Web Retrieval Degrades Safety Alignment in LLM Agents
  AI agents augment large language models with external tools such as web retrieval, enabling grounded and up-to-date responses. However, incorporating external content into the generation pipeline can weaken the safety alignment mechanisms that govern model out…
  arxiv Aditya Nawal, Manit Baser, Mohan Gurusamy 2026-05-28T01:23:48Z 置信度 0.78 cs.CLcs.AIcs.CR
• The User Side of AI Model Lifecycles: Evidence from the Keep4o Movement
  AI model lifecycles are commonly understood as a series of technical and organizational processes. Yet once a model enters sustained use, subsequent changes can also affect established user practices and user value. Using the Keep4o movement around GPT-4o as a…
  arxiv Yiwen Wu 2026-08-17T13:37:01Z 置信度 0.78 cs.HCcs.CY
• Can Good Writing Be Generative? Expert-Level AI Writing Emerges through Fine-Tuning on High-Quality Books
  Creative writing has long been considered a uniquely human endeavor, requiring voice and style that machines could not replicate. This assumption is challenged by Generative AI that can emulate thousands of author styles in seconds with negligible marginal lab…
  arxiv Tuhin Chakrabarty, Paramveer S. Dhillon 2026-01-26T10:59:21Z 置信度 0.78 cs.AIcs.CLcs.HC
• Contrastive Reasoning Alignment: Reinforcement Learning from Hidden Representations
  We propose CRAFT, a red-teaming alignment framework that leverages model reasoning capabilities and hidden representations to improve robustness against jailbreak attacks. Unlike prior defenses that operate primarily at the output level, CRAFT aligns large rea…
  arxiv Haozheng Luo, Yimin Wang, Jiahao Yu, Binghui Wang 等 2026-03-18T03:00:42Z 置信度 0.78 cs.AIcs.CLcs.LG
• XSearch: Explainable Code Search via Concept-to-Code Alignment
  Semantic code search has been widely adopted in both academia and industry. These approaches embed natural-language queries and code snippets into a shared embedding space and retrieve results based on vector similarity. Despit strong performance on benchmark …
  arxiv Yiming Liu, Ruofan Liu, Yun Lin, Zicong Zhang 等 2026-05-15T15:17:43Z 置信度 0.78 cs.SEcs.AI
• Task Alignment: A Simple Proxy for Practical Model Merging Across Diverse Vision Tasks
  Efficiently merging several models fine-tuned for different tasks, but stemming from the same pretrained base model, is of great practical interest. Despite extensive prior work, most evaluations of model merging in computer vision are restricted to image clas…
  arxiv Pau de Jorge, César Roberto de Souza, Björn Michele, Mert Bülent Sarıyıldız 等 2026-04-14T16:24:04Z 置信度 0.78 cs.CV
• Consistency Is Not Coherence: Orientation Search for Certified Alignments Between 4D Defence Upper Ontologies
  We align three upper ontologies that sit under UK and NATO defence data infrastructure: the Information Exchange Standard (IES), the Higher Quality Data Model (HQDM) that underpins the National Digital Twin, and Basic Formal Ontology (BFO). No public alignment…
  arxiv Fabio Rovai 2026-08-22T10:55:39Z 置信度 0.78 cs.AIcs.DBcs.LO
• Overview of ESDD2: Environment-Aware Speech and Sound Deepfake Detection Challenge
  The Environment-Aware Speech and Sound Deepfake Detection Challenge (ESDD2), held in conjunction with ICME 2026, evaluated systems for five component-level audio spoofing detection, where speech and environmental sounds may be manipulated independently or join…
  arxiv Xueping Zhang, Han Yin, Yang Xiao, Lin Zhang 等 2026-06-09T12:42:14Z 置信度 0.78 cs.SD
• LLM-as-a-Judge in Healthcare: A Scoping Analysis of Applications, Methods, and Human Alignment
  Large language models (LLMs) are increasingly deployed across healthcare applications, including clinical documentation, diagnostic reasoning, medicine recommendation, and medical education. Their outputs are largely unstructured clinical text, which is diffic…
  arxiv Lingyao Li, Deyi Li, Chen Chen, Renkai Ma 等 2026-05-24T21:59:32Z 置信度 0.78 cs.CY
• Position: AI Governance Needs ISO-like Interoperability Protocols, Not Just Laws
  As Artificial Intelligence (AI) systems become deeply integrated into critical global infrastructure, the urgency for robust governance frameworks has intensified. However, current approaches, led by jurisdiction-specific laws, policies, and voluntary framewor…
  arxiv Azmine Toushik Wasi, Mst Rafia Islam, Mahfuz Ahmed Anik, Taki Hasan Rafi 等 2026-05-30T14:58:24Z 置信度 0.78 cs.AIcs.CYcs.GTcs.HC
• A Task-Level Evaluation of AI Agents in Open-Source Projects
  In this paper, we present a comparative study of five autonomous coding agents using AIDev-pop, which is a public dataset containing thousands of AI-generated pull requests (PRs) across popular open-source repositories. We evaluate agents' performance along th…
  arxiv Shojibur Rahman, Md Fazle Rabbi, Minhaz Zibran 2026-02-02T17:05:19Z 置信度 0.78 cs.SE
• Multi-Faceted Self-Consistent Preference Alignment for Query Rewriting in Conversational Search
  Conversational Query Rewriting (CQR) aims to rewrite ambiguous queries to achieve more efficient conversational search. Early studies have predominantly focused on the rewriting in isolation, ignoring the feedback from query rewrite, passage retrieval and resp…
  arxiv Zhiyu Cao, Peifeng Li, Qiaoming Zhu 2026-04-08T07:38:45Z 置信度 0.78 cs.CLcs.AI
• PointAlign: Feature-Level Alignment Regularization for 3D Vision-Language Models
  The development of 3D Vision-Language Models (VLMs), crucial for applications in robotics, autonomous driving, and augmented reality, is severely constrained by the scarcity of paired 3D-text data. Existing methods rely solely on next-token prediction loss, us…
  arxiv Yuanhao Su, Shaofeng Zhang, Xiaosong Jia, Qi Fan 2026-02-28T02:17:46Z 置信度 0.78 cs.CV
• The 1st Winner for 5th PVUW MeViS-Text Challenge: Strong MLLMs Meet SAM3 for Referring Video Object Segmentation
  This report presents our winning solution to the 5th PVUW MeViS-Text Challenge. The track studies referring video object segmentation under motion-centric language expressions, where the model must jointly understand appearance, temporal behavior, and object i…
  arxiv Xusheng He, Canyang Wu, Jinrong Zhang, Weili Guan 等 2026-04-01T02:42:30Z 置信度 0.78 cs.CV
• EgoAction: Egocentric Action Composition with Reliability-Aware Temporal Fusion for the EPIC-KITCHENS Action Detection Challenge at CVPR 2026
  The EPIC-KITCHENS-100 Action Detection challenge evaluates whether a model can localize the start and end of each action in long untrimmed egocentric videos and assign the corresponding verb--noun action label. In this report, we formulate our submission as Eg…
  arxiv Zhiheng Fu, Zixu Li, Zhiwei Chen, Fangxu Liu 等 2026-05-23T10:05:56Z 置信度 0.78 cs.CV
• How Closely Do LLM Reviews Align with Human Peer Review?
  Large language models (LLMs) are increasingly used to generate scientific reviews, yet existing evaluations rarely examine whether different providers align with both conference decisions and human reviewing priorities within the same controlled setting. We co…
  arxiv Abraham Camelo-Guerrero, Jairo Diaz-Rodriguez 2026-08-04T13:39:36Z 置信度 0.78 cs.CLcs.AI
• Rank-Aware Hyperbolic Alignment for Vision-Language Dataset Distillation
  Vision-language dataset distillation (VLDD) compresses a large image-text paired dataset into a small set of synthetic pairs that can efficiently train contrastive vision-language models under strict data and compute budgets. Most existing methods match expert…
  arxiv Jongoh Jeong, Sun-Kyung Lee, Kuk-Jin Yoon 2026-06-28T15:41:31Z 置信度 0.78 cs.CVcs.AI
• Technology-Enhanced Tabletop Exercises for Cybersecurity Education: Lessons Learned
  This innovative practice full paper examines the integration of technology-enhanced tabletop exercises (TTXs) into computing education, focusing on cybersecurity curricula. The motivation is to better prepare students for complex, collaborative problem solving…
  arxiv Jan Vykopal, Pavel Čeleda, Martin Horák, Valdemar Švábenský 2026-07-30T13:17:12Z 置信度 0.78 cs.CYcs.CR
• Human-AI Interaction Traces as Blackout Poetry: Reframing AI-Supported Writing as Found-Text Creativity
  LLMs offer new creative possibilities for writers but also raise concerns about authenticity and reader trust, particularly when AI involvement is disclosed. Prior research has largely framed this as an issue of transparency and provenance, emphasizing the dis…
  arxiv Syemin Park, Soobin Park, Youn-kyung Lim 2026-03-10T16:04:53Z 置信度 0.78 cs.HCcs.AI
• Saliency-R1: Enforcing Interpretable and Faithful Vision-language Reasoning via Saliency-map Alignment Reward
  Vision-language models (VLMs) have achieved remarkable success across diverse tasks. However, concerns about their trustworthiness persist, particularly regarding tendencies to lean more on textual cues than visual evidence and the risk of producing ungrounded…
  arxiv Shizhan Gong, Minda Hu, Qiyuan Zhang, Chen Ma 等 2026-04-06T07:51:59Z 置信度 0.78 cs.CV
• Exploring Normativity in Stable Diffusion: Insights for XAI in the Arts
  Generative text-to-image (T2I) systems are increasingly adopted in creative practice, yet their normative behaviors remain underexplored from the perspective of creative practitioners. In this workshop paper, we present a within-subject study with 14 creative …
  arxiv Michelle Dutoit, Baptiste Caramiaux 2026-08-27T11:27:03Z 置信度 0.78 cs.HC
• AgentReputation: A Decentralized Agentic AI Reputation Framework
  Decentralized, agentic AI marketplaces are rapidly emerging to support software engineering tasks such as debugging, patch generation, and security auditing, often operating without centralized oversight. However, existing reputation mechanisms fail in this se…
  arxiv Mohd Sameen Chishti, Damilare Peter Oyinloye, Jingyue Li 2026-04-30T12:33:39Z 置信度 0.78 cs.AI
• Beyond Accuracy: Measuring Bias Acknowledgment in Chain-of-Thought Reasoning for Responsible AI Evaluation
  Reasoning models are increasingly used in settings where the final answer is not the only object of review: educational tools may show students intermediate steps, decision-support systems may require human oversight, and audit workflows may inspect traces for…
  arxiv Xian Sun, Wei Gao, Yingshuo Wang, Lingdong Kong 等 2026-06-13T05:41:57Z 置信度 0.78 cs.LG
• Child Safety in Generative AI: An Expert-Guided and Incident-Grounded Evaluation Framework
  As generative AI is increasingly used by children and adolescents, there is a growing need for risk evaluation frameworks that account for child-specific harms. However, most existing safety evaluation frameworks focus on general user populations, often overlo…
  arxiv Haein Kong 2026-07-01T03:43:47Z 置信度 0.78 cs.HC
• Return-to-Go Is More Than a Number: Q-Guided Alignment for Return-Conditioned Supervised Learning
  Conditioned Sequence Models (CSMs) learn policies by treating return-to-go (RTG) as a control signal. However, existing CSMs often treat the RTGs as simple numerical inputs rather than aligning them with the performance of their policies. In this paper, we pro…
  arxiv Yuxiao Yang, Weitong Zhang 2026-05-27T19:24:35Z 置信度 0.78 cs.LGcs.AI