Daniele Calandriello

机构：Institut national de recherche en sciences et technologies du numérique, Italian Institute of Technology

发表论文 63 篇 · 总被引 403 次 · h-index 12

代表论文

• A General Theoretical Paradigm to Understand Learning from Human Preferences (2023 · arXiv (Cornell University) · 被引 14)
• Improved large-scale graph learning through ridge spectral sparsification (2026 · arXiv (Cornell University) · 被引 9)
• Understanding the performance gap between online and offline alignment algorithms (2024 · HAL (Le Centre pour la Communication Scientifique Directe) · 被引 3)
• Pack only the essentials: Adaptive dictionary learning for kernel ridge regression (2026 · arXiv (Cornell University) · 被引 2)
• Human Alignment of Large Language Models through Online Preference Optimisation (2024 · arXiv (Cornell University) · 被引 2)
• Multi-turn Reinforcement Learning with Preference Human Feedback (2024 · 被引 2)