5 papers
Steering Pretrained Drafters during Speculative Decoding
Frédéric Berdoz, Peer Rheinboldt, Roger Wattenhofer
Speculative decoding accelerates language model inference by separating generation into fast drafting and parallel verification. Its main limitation is drafter-verifier misalignmen…
High-Fidelity Speech Enhancement via Discrete Audio Tokens
Luca A. Lanzendörfer, Frédéric Berdoz, Antonis Asonitis +1
Recent autoregressive transformer-based speech enhancement (SE) methods have shown promising results by leveraging advanced semantic understanding and contextual modeling of speech…
Text-to-Scene with Large Reasoning Models
Frédéric Berdoz, Luca A. Lanzendörfer, Nick Tuninga +1
Prompt-driven scene synthesis allows users to generate complete 3D environments from textual descriptions. Current text-to-scene methods often struggle with complex geometries and…
Recommender Systems for Democracy: Toward Adversarial Robustness in Voting Advice Applications
Frédéric Berdoz, Dustin Brunner, Yann Vonlanthen +1
Voting advice applications (VAAs) help millions of voters understand which political parties or candidates best align with their views. This paper explores the potential risks thes…
Can an AI Agent Safely Run a Government? Existence of Probably Approximately Aligned Policies
Frédéric Berdoz, Roger Wattenhofer
While autonomous agents often surpass humans in their ability to handle vast and complex data, their potential misalignment (i.e., lack of transparency regarding their true objecti…