The Berkeley NLP Group

State-of-the-art Software

The Berkeley Crossword Solver combines state-of-the-art QA with constraint solving.
The Berkeley Neural Parser infers syntactic annotations using neural networks and self-attention. We have an online demo!
The Berkeley Parser Analyser performs error analysis on parse trees.
Ocular is a state-of-the-art historical document recognition system.
The Berkeley Aligner aligns the words in multilingual parallel texts.

Murphy is a library of learning algorithms for structured prediction.
The Berkeley Parser parses sentences using PCFGs. It also has a demo.
Puck is a lightning-fast version of the Berkeley Parser that uses GPUs.
Epic is a discriminative parser using many kinds of annotations.
The neural CRF parser effectively leverages distributed representations of words by scoring anchored rule productions with feedforward neural networks.
The Berkeley CCG to PST tool converts CCG Derivations into Penn Treebank style trees.
The Berkeley Entity Resolution System jointly models and predicts named entity chunks, coreference, and entity links.
The Berkeley Coreference Resolution System is a high-performing English coreference resolution system.
The Berkeley Coreference Analyser performs error analysis on coreference resolution output.
The Berkeley Document Summarizer is an extractive and compressive single-document summarization system learned on a large newswire corpus.
The Berkeley LM provides a library for storing large n-gram language models efficiently in memory.