Data Sets

• AO-CHILDES

  5M words of American-English child-directed speech ordered by the age of the child.

  [more info]
• SRL-CHILDES

  200K American-English utterances from the CHILDES database annotated with semantic role labels using the Propbank formalism.

  [more info]
• AO-CHILDES Relations

  Word pairs in AO-CHILDES organized by semantic-relation (e.g. the pair cat-animal is an instance of hypernymy).

  [more info]
• MissingAdjunct

  A corpus of pseudo-English sentences with experimenter-controlled statistics for evaluating the ability of distributional models to infer the most plausible adjunct of a phrase.

  [more info]