Combining Supervised and Unsupervised Enembles for Knowledge Base Population
نویسندگان
چکیده
We propose an algorithm that combines supervised and unsupervised methods to ensemble multiple systems for two popular Knowledge Base Population (KBP) tasks, Cold Start Slot Filling (CSSF) and Tri-lingual Entity Discovery and Linking (TEDL). We demonstrate that it outperforms the best system for both tasks in the 2015 competition, several ensembling baselines, as well as a state-of-the-art stacking approach. The success of our technique on two different and challenging problems demonstrates the power and generality of our combined approach to ensembling.
منابع مشابه
Supervised and Unsupervised Ensembling for Knowledge Base Population
We present results on combining supervised and unsupervised methods to ensemble multiple systems for two popular Knowledge Base Population (KBP) tasks, Cold Start Slot Filling (CSSF) and Tri-lingual Entity Discovery and Linking (TEDL). We demonstrate that our combined system along with auxiliary features outperforms the best performing system for both tasks in the 2015 competition, several ense...
متن کاملStacked Ensembles of Information Extractors for Knowledge-Base Population by Combining Supervised and Unsupervised Approaches
The UTAustin team participated in two main tasks this year the Cold Start Slot Filling (CSSF) task and the Slot-Filler Validation/Ensembling task, which was divided into the filtering and ensembling subtasks. Our system uses stacking to ensemble multiple systems for the KBP slot filling task, as described in our ACL 2015 paper. We expand the stacking approach by allowing the classifier to also ...
متن کاملUsing a weakly supervised approach and lexical patterns for the KBP slot filling task
We present in this article the system we developed for participating to the slot filling task in the Knowledge Base Population (KBP) track of the 2011 Text Analysis Conference (TAC). This system is based on a weakly supervised approach and lexical patterns. In this participation, we tested more specifically the integration of an additional unsupervised relation identification component dedicate...
متن کاملBasis Technology at TAC 2012 Entity Linking
Basis Technology participated in the TAC Entity-Link task of the Knowledge Base Population track at TAC 2012. We developed a supervised learning approach which produces a single model that is capable of operating across all languages and entity types. Our features are based on the outputs of other models, many of which are unsupervised.
متن کاملCombining Semi-Unsupervised Acquisition of Corpora and Supervised Learning of Textual Entailment Rules
As many NLP tasks, textual entailment recognition (RTE) requires large semantic knowledge bases. Unsupervised or semi-supervised knowledge learning methods are very attractive ways for acquiring these knowledge bases from raw texts. Many of these approaches use the Distributional Hypothesis. This can be used to detect equivalence or relatedness between words (e.g., dog and cat), word sequences,...
متن کامل