732 citations
- Microsoft (United States)US95 papers
- Carnegie Mellon UniversityUS40 papers
- University of WashingtonUS38 papers
- Massachusetts Institute of TechnologyUS36 papers
- Stanford UniversityUS30 papers
- University of CambridgeGB30 papers
- University of Science and Technology of ChinaCN30 papers
- Peking UniversityCN28 papers
- Tsinghua UniversityCN24 papers
- University College LondonGB24 papers
- University of California, BerkeleyUS24 papers
- Microsoft Research Asia (China)CN18 papers
4 papers · 2 filters
Multi-Head Adapter Routing for Cross-Task Generalization
Lucas Caccia, Edoardo Ponti, Zhan Su +3
Parameter-efficient fine-tuning (PEFT) for cross-task generalization consists in pre-training adapters on a multi-task training set before few-shot adaptation to test tasks. Polytr…
Augmenting Interpretable Models with LLMs during Training
Chandan Singh, Armin Askari, Rich Caruana +1
Recent large language models (LLMs) have demonstrated remarkable prediction performance for a growing array of tasks. However, their proliferation into high-stakes domains (e.g. me…
Social Network Structure Shapes Innovation: Experience-sharing in RL with SAPIENS
Eleni Nisioti, Mateo Mahaut, Pierre-Yves Oudeyer +2
Human culture relies on innovation: our ability to continuously explore how existing elements can be combined to create new ones. Innovation is not solitary, it relies on collectiv…
Language and Culture Internalisation for Human-Like Autotelic AI
Cédric Colas, Tristan Karch, Clément Moulin-Frier +1
Building autonomous agents able to grow open-ended repertoires of skills across their lives is a fundamental goal of artificial intelligence (AI). A promising developmental approac…