Hacking Task Confounder in Meta-Learning

arXiv:2312.05771v2 Announce Type: replace-cross
Abstract: Meta-learning enables rapid generalization to new tasks by learning knowledge from various tasks. It is intuitively assumed that as the training progresses, a model will acquire richer knowledge, leading to better generalization performance. However, our experiments reveal an unexpected result: there is negative knowledge transfer between tasks, affecting generalization performance. To explain this phenomenon, we conduct Structural Causal Models (SCMs) for causal analysis. Our investigation uncovers the presence of spurious correlations between task-specific causal factors …

abstract arxiv cs.lg hacking however knowledge meta meta-learning negative performance stat.ml tasks training transfer type will