Reaction
The tension between elegant theories and messy observations
Peter Norvig argues that probabilistic language models can capture usage patterns that simpler grammatical rules miss.
TLDR
One post recalls a professor calling something a “fudge factor” that makes the maths fit experimental results. Another connects that anecdote to Peter Norvig’s essay on statistical learning, arguing that elegant theories shouldn’t leave out messy phenomena. In the essay, Norvig makes the case for probabilistic models that account for how people actually use language.
Combined views
195
1 Source, first seen ago