Hierarchical frame induction via probabilistic models (B08)

项目B 08旨在使用概率框架建模和机器学习技术，对动词描述的事件进行无监督的框架归纳。术语“无监督”是指语义信息，即，系统将句法信息作为输入。结果不仅要识别谓词-论元关系，而且要通过层次有序的框架类型和语义角色来提供这些关系的丰富表征。在第三个CRC阶段，B 08将专注于a）动词框架之间的交替，B）超越动词级别的扩展，c）更广泛的机器学习技术选择，以及d）多语言框架归纳。

Project B08 aims at unsupervised induction of frames for events described by verbs, using techniques from probabilistic frame modeling and from machine learning. The term ”unsupervised” refers to semantic information, i.e., the system takes syntactic information as input. The outcome should not only identify predicate-argument relations but also provide a rich characterization of these relations via hierarchically ordered frame types and semantic roles. In the third CRC phase, B08 will focus on a) alternations between verb frames, b) extensions beyond sentence-level, c) a broader choice of machine learning techniques, and d) multi-lingual frame induction.