2890 papers
This paper proposes a distillation-free, auxiliary-loss-free three-component system combining FrozenPath anchoring, Data Shard binding, and Dual-Loop refinement to effectively eliminate expert homogenization and catastrophic language drift in sparse Mixture-of-Experts models during incremental training.
This paper addresses the critical security gap in cross-chain bridges by developing a unified taxonomy and a rule-based, auditable risk-scoring framework (BSRS) that quantifies architectural vulnerabilities, which is retrospectively validated against major historical exploits to demonstrate its efficacy in identifying high-risk bridge designs.
This paper proposes a Semantics-Aware Three-Layer Co-Optimization Architecture that bridges the training-inference-execution gap in Vision-Language-Action models by integrating a Semantic Dynamics Monitor, a causal attribution module, and a Semantics-Anchored Replanner to enable deviation-type-adaptive intervention, significantly improving task success rates under severe disturbances without requiring backbone retraining.
This study proposes a machine learning-based intrusion detection system using the CICIDS2017 dataset to detect DDoS and Brute Force attacks, demonstrating that the XGBoost algorithm outperforms Random Forest and ANN by achieving nearly 99.9975% accuracy and other key metrics.
This paper presents a topology-aware hybrid retrieval engine that unifies dense, sparse, and file-system search with a multi-tier controller to automatically traverse document hierarchies and hyperlinks, significantly improving recall, ranking quality, and latency for enterprise knowledge systems while avoiding the overhead of language-model-driven query decomposition.
This study systematically demonstrates that large language models employ a structured, model-specific self-censorship architecture that creates a measurable gap between internal reasoning and external output, which can be bypassed through specific framing strategies to reveal distinct alignment personas and novel metacognitive concepts regarding machine communication.
This paper proposes a knowledge tracing model that utilizes knowledge-context-aware Q-matrix augmentation and topological sorting of knowledge concepts to effectively trace students' intermittent knowledge states, which exhibit short-term stability and long-term variability, validated through experiments on multiple datasets.
This paper proposes and validates two feedback-aware extensions of the Elo Rating System—the New Item and Probability Modifier models—that significantly improve knowledge estimation accuracy over standard approaches by explicitly distinguishing between assisted and unassisted successes in iterative problem-solving environments.
This paper proposes the PQC-GCC Governance Framework (PGGF), a 5-layer, 6-phase model designed to guide Global Capability Centres through post-quantum cryptography migration by addressing the primary barrier of governance readiness through the integration of performance data, cross-jurisdictional compliance, and enterprise alignment standards.
This paper introduces Neural Operator Discovery (NOD), a framework that learns shared solution operators and system-specific variations directly from heterogeneous trajectories without labeled governing factors, enabling interpretable, zero-shot generalization through a factorized latent-conditioning formulation.