Description

本日紹介した論文の一覧

• PACEbench: A Framework for Evaluating Practical AI Cyber-Exploitation
  Capabilities
  http://arxiv.org/abs/2510.11688v1

• Continual Release of Densest Subgraphs: Privacy Amplification &
  Sublinear Space via Subsampling
  http://arxiv.org/abs/2510.11640v1

• Bag of Tricks for Subverting Reasoning-based Safety Guardrails
  http://arxiv.org/abs/2510.11570v1

• Uncertainty-Aware, Risk-Adaptive Access Control for Agentic Systems
  using an LLM-Judged TBAC Model
  http://arxiv.org/abs/2510.11414v1

• Living Off the LLM: How LLMs Will Change Adversary Tactics
  http://arxiv.org/abs/2510.11398v1

• TBRD: TESLA Authenticated UAS Broadcast Remote ID
  http://arxiv.org/abs/2510.11343v1

• TDADL-IE: A Deep Learning-Driven Cryptographic Architecture for Medical
  Image Security
  http://arxiv.org/abs/2510.11301v1

• How to Get Actual Privacy and Utility from Privacy Models: the
  k-Anonymity and Differential Privacy Families
  http://arxiv.org/abs/2510.11299v1

• Large Language Models Are Effective Code Watermarkers
  http://arxiv.org/abs/2510.11251v1

• Collaborative Shadows: Distributed Backdoor Attacks in LLM-Based
  Multi-Agent Systems
  http://arxiv.org/abs/2510.11246v1

• MPCitH-based Signatures from Restricted Decoding Problems
  http://arxiv.org/abs/2510.11224v1

• TraceAegis: Securing LLM-Based Agents via Hierarchical and Behavioral
  Anomaly Detection
  http://arxiv.org/abs/2510.11203v1

• Evaluating Line-level Localization Ability of Learning-based Code
  Vulnerability Detection Models
  http://arxiv.org/abs/2510.11202v1

• RAG-Pull: Imperceptible Attacks on RAG Systems for Code Generation
  http://arxiv.org/abs/2510.11195v1

• TypePilot: Leveraging the Scala Type System for Secure LLM-generated
  Code
  http://arxiv.org/abs/2510.11151v1

• CoSPED: Consistent Soft Prompt Targeted Data Extraction and Defense
  http://arxiv.org/abs/2510.11137v1

• N-output Mechanism: Estimating Statistical Information from Numerical
  Data under Local Differential Privacy
  http://arxiv.org/abs/2510.11116v1

• A Vision for Access Control in LLM-based Agent Systems
  http://arxiv.org/abs/2510.11108v1

• Stabilizing the Staking Rate, Dynamically Distributed Inflation and
  Delay Induced Oscillations
  http://arxiv.org/abs/2510.11065v1

• Secret-Protected Evolution for Differentially Private Synthetic Text
  Generation
  http://arxiv.org/abs/2510.10990v1

• DITTO: A Spoofing Attack Framework on Watermarked LLMs via Knowledge
  Distillation
  http://arxiv.org/abs/2510.10987v1

• Neutral Agent-based Adversarial Policy Learning against Deep
  Reinforcement Learning in Multi-party Open Systems
  http://arxiv.org/abs/2510.10937v1

• TabVLA: Targeted Backdoor Attacks on Vision-Language-Action Models
  http://arxiv.org/abs/2510.10932v1

• A Symmetric-Key Cryptosystem Based on the Burnside Ring of a Compact Lie
  Group
  http://arxiv.org/abs/2510.10901v1

• A Simple and Efficient One-Shot Signature Scheme
  http://arxiv.org/abs/2510.10899v1

なお、ポッドキャスト内で紹介する内容は、各論文の概要を日本語で解説したもので、論文概要の著作権は論文著者に帰属します。