Dissertation questions transfer learning generalization assumptions

A new dissertation explores the effectiveness of transfer learning in deep learning, questioning whether strong generalization on a source task guarantees transferability to a related target task. The research hypothesizes that this is not necessarily the case and employs adversarial attacks to find counterexamples. Empirical results indicate a significant relationship between source task generalization and transferability, though they do not definitively disprove generalization as a sufficient condition for successful transfer. AI

IMPACT This research could refine understanding of how knowledge is transferred between AI tasks, potentially leading to more efficient model training.

RANK_REASON The item is a dissertation submitted to arXiv, focusing on a research question within machine learning. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• alphaXiv
• arXiv
• arXivLabs
• CatalyzeX Code Finder for Papers
• Connected Papers
• CORE Recommender
• DagsHub
• deep learning
• Gotit.pub
• Hugging Face
• IArxiv Recommender
• Litmaps
• ScienceCast
• scite Smart Citations
• transfer learning

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →