Agent-Based Unsupervised Grammar Induction
نویسنده
چکیده
In this paper, we describe an agent-based evolutionary computing approach to unsupervised grammar induction called grael (Grammar Evolution). Extending a general framework for data driven grammar optimization and induction, the evolutionary setup of grael can be used to automatically induce and optimize grammars from scratch on the basis of unstructured text. Agents are equipped with a very basic grammar induction module to bootstrap structure. Over an extended series of inter-agent interactions, the agents optimize their grammars, as the society attempts to converge towards an optimal grammar. We highlight two proof-of-the-principle experiments that show that grael is able to yield a reasonably performant phrase structuregrammar in an unsupervised manner. In contrast to the current state-of-the-art systems, grael does not require a large amount of data to induce a workable grammar.
منابع مشابه
Evolutionary Computing as a Tool for Grammar Development
In this paper, an agent-based evolutionary computing technique is introduced, that is geared towards the automatic induction and optimization of grammars for natural language (grael). We outline three instantiations of the grael-environment: thegrael-1 system uses large annotated corpora to bootstrap grammatical structure in a society of autonomous agents, that tries to optimally redistribute g...
متن کاملHistory Based Unsupervised Data Oriented Parsing
Grammar induction is a basic step in natural language processing. Based on the volume of information that is used by different methods, we can distinguish three types of grammar induction method: supervised, unsupervised, and semi-supervised. Supervised and semisupervised methods require large tree banks, which may not currently exist for many languages. Accordingly, many researchers have focus...
متن کاملCreating Features from a Learned Grammar in a Simulated Student
Understanding and developing intelligent agents that simulate human learning has been a long-standing goal in both artificial intelligence and cognitive science. Although learning agents are able to produce intelligent behavior with less human knowledge engineering than in the past, intelligent agent developers are still required to manually encode much prior domain knowledge. We recently propo...
متن کاملUnsupervised Discriminative Induction of Synchronous Grammar for Machine Translation
We present a global log-linear model for synchronous grammar induction, which is capable of incorporating arbitrary features. The parameters in the model are trained in an unsupervised fashion from parallel sentences without word alignments. To make parameter training tractable, we also propose a novel and efficient cube pruning based synchronous parsing algorithm. Using learned synchronous gra...
متن کامل