This paper presents a model showing that while individually beneficial, widespread adoption of generative AI can reduce social welfare by degrading future model performance through "model collapse," especially for high-incentive tasks. It highlights a critical social dilemma where short-term gains lead to long-term collective losses.

The paper introduces an ontology‑driven dataspace to formalize test annotations for cyber‑physical energy systems, enabling reproducible and FAIR‑compliant experimentation across laboratories.

The paper proposes the “innovation annulus” model to explain the persistent gap between open scholarly metadata and commercially refined data, offering a welfare‑based framework for governing its optimal width.

This paper introduces SAFER, a continual unlearning framework designed to address critical issues like knowledge erosion and forgetting reversal in repeated machine unlearning scenarios. It aims to maintain model utility and prevent the reactivation of forgotten information, making AI systems more reliable and privacy-compliant over their lifecycle.

This paper introduces UAF (Unified Audio Front-end LLM), a novel large language model that unifies critical audio front-end tasks like voice activity detection, speaker recognition, and automatic speech recognition into a single end-to-end generative framework. UAF aims to overcome the limitations of traditional cascaded pipelines and enhance full-duplex speech interaction by jointly modeling semantic content and interaction-level control signals.