Dual-Space Hierarchical Learning for Deepfake Detection

IEEE Signal Processing Letters · 2026-01-01

Most existing deepfake detection methods learn visual artifacts in Euclidean space, overlooking the intrinsic hierarchical relationships among forgery samples. In practice, manipulated images deviate progressively from real ones, forming a hierarchy that Euclidean geometry struggles to model, while hyperbolic space better captures such structures due to its exponential capacity. Motivated by this observation, we propose a dual-space hierarchical learning framework that jointly models Euclidean visual semantics and hyperbolic hierarchical representations. Specifically, Euclidean features are projected into hyperbolic space, where a Hyperbolic Hierarchy-Aware Attention (HHAA) module is introduced to capture hierarchical dependencies among samples. To effectively integrate complementary information from the two spaces, a Dual-Space Gated Fusion (DSGF) module is designed to adaptively inject hierarchical cues into Euclidean features. A joint supervision strategy is further applied to enhance discriminative representation learning. Extensive experiments on multiple deepfake detection benchmarks demonstrate that the proposed method improves detection performance and generalization ability.

Don't Be a Pot Stirrer! Authorized Vector Data Retrieval via Access-Aware Indexing

ArXiv.org · 2026-05-02

Vector databases increasingly enforce role-based access control, where each top-k approximate nearest neighbor query must return only vectors the querying role is authorized to access. Two extremes bracket the design space. A single global index built over all vectors avoids duplication but wastes search effort on unauthorized vectors and degrades recall, while an oracle index, built with all authorized vectors to the query roles, searches only authorized vectors but duplicates every shared vector between roles or queries. We present Veda and its efficient variant EffVeda, two indexing strategies built on an access-aware lattice to address access control in vector databases. The methods first partitions the dataset into disjoint data blocks by role combination, then leverage the structure of the access-aware lattice to apply copy and merge operations to group co-accessed blocks under a user-specified storage budget. Large nodes in the lattice are then indexed with HNSW, while small nodes are retained for linear scan. To facilitate query processing on the lattice, our methods construct a query plan that selects the minimal set of nodes that covers all authorized data for each role. At query time, coordinated search first queries pure (authorized-only) nodes to populate a global top-k heap, then leverages the resulting distance bound of the k-th data in the heap to prune exploration on impure nodes, avoiding the inflated search that independent per-index execution would require. Evaluations show that our methods deliver higher throughput at high recall while closely tracking the storage budget.

Don't Be a Pot Stirrer! Authorized Vector Data Retrieval via Access-Aware Indexing

arXiv (Cornell University) · 2026-05-02

Vector databases increasingly enforce role-based access control, where each top-k approximate nearest neighbor query must return only vectors the querying role is authorized to access. Two extremes bracket the design space. A single global index built over all vectors avoids duplication but wastes search effort on unauthorized vectors and degrades recall, while an oracle index, built with all authorized vectors to the query roles, searches only authorized vectors but duplicates every shared vector between roles or queries. We present Veda and its efficient variant EffVeda, two indexing strategies built on an access-aware lattice to address access control in vector databases. The methods first partitions the dataset into disjoint data blocks by role combination, then leverage the structure of the access-aware lattice to apply copy and merge operations to group co-accessed blocks under a user-specified storage budget. Large nodes in the lattice are then indexed with HNSW, while small nodes are retained for linear scan. To facilitate query processing on the lattice, our methods construct a query plan that selects the minimal set of nodes that covers all authorized data for each role. At query time, coordinated search first queries pure (authorized-only) nodes to populate a global top-k heap, then leverages the resulting distance bound of the k-th data in the heap to prune exploration on impure nodes, avoiding the inflated search that independent per-index execution would require. Evaluations show that our methods deliver higher throughput at high recall while closely tracking the storage budget.

Research on the Revitalization of World Cultural Heritage in the Grand Canal under the Perspective of Cultural and Creative Canal

Journal of Social Science Humanities and Literature · 2025-01-10

This thesis focuses on the revitalisation of the World Cultural Heritage of the Jiangsu section of the Grand Canal from the perspective of cultural creativity, discusses its theoretical basis, analyses in detail the current situation of cultural creativity, and then proposes a series of targeted revitalisation strategies. By exploring the cultural connotation, promoting the integration of innovative design and science and technology, as well as strengthening the training and cooperation of talents, it aims to provide useful reference and practical guidance for the protection and sustainable development of the World Cultural Heritage of the Jiangsu section of the Grand Canal, and to realise the synergy between cultural inheritance and economic and social development.

The Application of Reinforcement Learning in the Personalized Learning Path Planning of Bilingual Study Tour Products

2025-08-26

This paper proposes a method that combines the TransR algorithm, the Deep Q Network (DQN), and Proximal Policy Optimization. The reinforcement learning method of PPO-Deep Embedding for Knowledge Optimization (DEKO), is used for personalized learning path planning of bilingual study tour products. Firstly, the deep embedding of entities and relations in the knowledge graph is achieved through the TransR algorithm to capture the complex semantic associations of bilingual knowledge. DQN is used to handle the cognitive state of learners, dynamically update the correlation weights of cross-language knowledge points, and achieve cross-language knowledge transfer in the learning path. PPO optimizes the dynamic adjustment strategy of the learning path to ensure the adaptability and stability of the path. The experimental results show that DEKO performs well in the personalized learning path planning of bilingual study tour products. Compared with traditional methods, it can dynamically adjust the learning path according to the progress and preferences of learners, significantly improving the learning efficiency and the accuracy of path planning. This verifies the effectiveness and feasibility of DEKO in bilingual study tour products, providing new ideas and technical means for achieving more intelligent educational technologies.

Development of a TaqMan probe-based dual real time PCR assay for the identification of NADC34-like PRRSV

Veterinary and Animal Science · 2025-08-27

Porcine reproductive and respiratory syndrome virus (PRRSV) is an RNA virus that induces reproductive disorders in sows and respiratory diseases in growing pigs. Recently, the NADC34-like strain of PRRSV has become more prevalent, with outbreaks occurring across pig farms in China. However, a reliable diagnostic method for the clinical detection of this strain has been absent. This study developed a TaqMan probe-based dual real-time quantitative PCR assay targeting the M and GP5 genes to specifically identify the NADC34-like PRRSV strain. The assay exhibited high specificity, detecting exclusively the NADC34-like strain without cross-reactivity with other PRRSV strains. The detection limits for pMD-M and pMD-GP5 plasmids were 2.67 × 10² and 1.35 × 10¹ copies/μL, respectively, indicating high assay sensitivity. The assay also demonstrated excellent reproducibility, with coefficient of variation (CV) values for both recombinant plasmids below 2 %. Among 251 clinical samples, 27 tested positive for NADC34-like PRRSV. This study establishes an accurate, sensitive, and reliable TaqMan dual real-time PCR assay for detecting NADC34-like PRRSV, offering a valuable tool for clinical diagnostics and outbreak management in pig farms.

Bridging Today and the Future of Humanity: AI Safety in 2024 and Beyond

Qeios · 2025-02-25

The advancements in generative AI inevitably raise concerns about their risks and safety implications, which, in return, catalyzes significant progress in AI safety. However, as this field continues to evolve, a critical question arises: are our current efforts on AI safety aligned with the advancements of AI as well as the long-term goal of human civilization? This paper presents a blueprint for an advanced human society and leverages this vision to guide current AI safety efforts. It outlines a future where the _Internet of Everything_ becomes reality, and creates a roadmap of significant technological advancements towards this envisioned future. For each stage of the advancements, this paper forecasts potential AI safety issues that humanity may face. By projecting current efforts against this blueprint, this paper examines the alignment between the current efforts and the long-term needs, and highlights unique challenges and missions that demand increasing attention from AI safety practitioners in the 2020s. This vision paper aims to offer a broader perspective on AI safety, emphasizing that our current efforts should not only address immediate concerns but also anticipate potential risks in the expanding AI landscape, thereby promoting a safe and sustainable future of AI and human civilization.

Bridging the Safety Gap: A Guardrail Pipeline for Trustworthy LLM Inferences

ArXiv.org · 2025-02-12 · 1 citations

We present Wildflare GuardRail, a guardrail pipeline designed to enhance the safety and reliability of Large Language Model (LLM) inferences by systematically addressing risks across the entire processing workflow. Wildflare GuardRail integrates several core functional modules, including Safety Detector that identifies unsafe inputs and detects hallucinations in model outputs while generating root-cause explanations, Grounding that contextualizes user queries with information retrieved from vector databases, Customizer that adjusts outputs in real time using lightweight, rule-based wrappers, and Repairer that corrects erroneous LLM outputs using hallucination explanations provided by Safety Detector. Results show that our unsafe content detection model in Safety Detector achieves comparable performance with OpenAI API, though trained on a small dataset constructed with several public datasets. Meanwhile, the lightweight wrappers can address malicious URLs in model outputs in 1.06s per query with 100% accuracy without costly model calls. Moreover, the hallucination fixing model demonstrates effectiveness in reducing hallucinations with an accuracy of 80.7%.

FedSecurity: A Benchmark for Attacks and Defenses in Federated Learning and Federated LLMs

2024-08-24 · 27 citations

This paper introduces FedSecurity, an end-to-end benchmark that serves as a supplementary component of the FedML library for simulating adversarial attacks and corresponding defense mechanisms in Federated Learning (FL). FedSecurity eliminates the need for implementing the fundamental FL procedures, e.g., FL training and data loading, from scratch, thus enables users to focus on developing their own attack and defense strategies. It contains two key components, including FedAttacker that conducts a variety of attacks during FL training, and FedDefender that implements defensive mechanisms to counteract these attacks. FedSecurity has the following features: i) It offers extensive customization options to accommodate a broad range of machine learning models (e.g., Logistic Regression, ResNet, and GAN) and FL optimizers (e.g., FedAVG, FedOPT, and FedNOVA); ii) it enables exploring the effectiveness of attacks and defenses across different datasets and models; and iii) it supports flexible configuration and customization through a configuration file and some APIs. We further demonstrate FedSecurity's utility and adaptability through federated training of Large Language Models (LLMs) to showcase its potential on a wide range of complex applications.

LLM Multi-Agent Systems: Challenges and Open Problems

arXiv (Cornell University) · 2024-02-05 · 26 citations

This paper explores multi-agent systems and identify challenges that remain inadequately addressed. By leveraging the diverse capabilities and roles of individual agents, multi-agent systems can tackle complex tasks through agent collaboration. We discuss optimizing task allocation, fostering robust reasoning through iterative debates, managing complex and layered context information, and enhancing memory management to support the intricate interactions within multi-agent systems. We also explore potential applications of multi-agent systems in blockchain systems to shed light on their future development and application in real-world distributed systems.