The Role of $n$-gram Smoothing in the Age of Neural Networks

arXiv:2403.17240v1 Announce Type: new
Abstract: For nearly three decades, language models derived from the $n$-gram assumption held the state of the art on the task. The key to their success lay in the application of various smoothing techniques that served to combat overfitting. However, when neural language models toppled $n$-gram models as the best performers, $n$-gram smoothing techniques became less relevant. Indeed, it would hardly be an understatement to suggest that the line of inquiry into $n$-gram smoothing techniques became …

abstract age application art arxiv cs.cl however key language language models networks neural networks overfitting role state state of the art success the key type