VLMs preserve linearly separable visual magnitudes and can compare them, yet collapse at symbolic mapping because visual and textual number spaces remain fractured and disjoint.
Title resolution pending
8 Pith papers cite this work. Polarity classification is still indexing.
8
Pith papers citing it
citation-role summary
background 2
citation-polarity summary
roles
background 2polarities
background 2representative citing papers
Identifies concrete attacks from a malicious Provider on SAGA and proposes SAGA-BFT, SAGA-MON, SAGA-AUD, and SAGA-HYB mitigations offering different security-performance trade-offs.
The web's anti-bot regime should be replaced by a framework that presumptively lets user-authorized AI agents act for their principals, requires platforms to disclose access policies, and permits agent blocking only when proportionate to concrete harms.
The 2025 AI Agent Index catalogs technical and safety details for 30 deployed AI agents and finds low developer transparency on safety, evaluations, and societal impacts.
MAGIQ introduces a post-quantum governance system for multi-agent AI that supports policy budgets, session enforcement, message attribution, and UC-based security proofs while comparing overhead to SAGA.
Proposes layered architecture and three mechanisms (semantic propagation, identity/reputation, semantic-gradient design) for distributed agent networks with prototype simulations.
The paper introduces a dual-layer AI identification framework that integrates cryptographic, blockchain, and zero-knowledge techniques with governance checkpoints to support lifecycle accountability in digital enterprises.
LLM agents enable universal interoperability by serving as automatic translators and adapters between proprietary digital services.
citing papers explorer
-
Dissociative Identity: Language Model Agents Lack Grounding for Reputation Mechanisms
VLMs preserve linearly separable visual magnitudes and can compare them, yet collapse at symbolic mapping because visual and textual number spaces remain fractured and disjoint.
-
Attacks and Mitigations for Distributed Governance of Agentic AI under Byzantine Adversaries
Identifies concrete attacks from a malicious Provider on SAGA and proposes SAGA-BFT, SAGA-MON, SAGA-AUD, and SAGA-HYB mitigations offering different security-performance trade-offs.
-
The Agentic Web Requires New Normative Infrastructure
The web's anti-bot regime should be replaced by a framework that presumptively lets user-authorized AI agents act for their principals, requires platforms to disclose access policies, and permits agent blocking only when proportionate to concrete harms.
-
The 2025 AI Agent Index: Documenting Technical and Safety Features of Deployed Agentic AI Systems
The 2025 AI Agent Index catalogs technical and safety details for 30 deployed AI agents and finds low developer transparency on safety, evaluations, and societal impacts.
-
MAGIQ: A Post-Quantum Multi-Agentic AI Governance System with Provable Security
MAGIQ introduces a post-quantum governance system for multi-agent AI that supports policy budgets, session enforcement, message attribution, and UC-based security proofs while comparing overhead to SAGA.
-
Distributed General-Purpose Agent Networks: Architecture, Key Mechanisms, and Prototypes
Proposes layered architecture and three mechanisms (semantic propagation, identity/reputation, semantic-gradient design) for distributed agent networks with prototype simulations.
-
AI Identification: An Integrated Framework for Sustainable Governance in Digital Enterprises
The paper introduces a dual-layer AI identification framework that integrates cryptographic, blockchain, and zero-knowledge techniques with governance checkpoints to support lifecycle accountability in digital enterprises.
-
LLM Agents Are the Antidote to Walled Gardens
LLM agents enable universal interoperability by serving as automatic translators and adapters between proprietary digital services.