Learning Semantic Script Knowledge with Event Embeddings

Authors: Ashutosh Modi; Ivan Titov

ICLR 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate our approach on crowdsourced data collected for script induction by Regneri et al. (2010)... In our experiments, we compared our event embedding model (EE) against three baseline systems (BL , MSA) and BSMSA is the system of Regneri et al. (2010). BS is a a hierarchical Bayesian system of Frermann et al. (2014). The results are presented in Table 1.
Researcher Affiliation Academia Ashutosh Modi ASHUTOSH@COLI.UNI-SB.DE Saarland University, Saarbr ucken, Germany Ivan Titov TITOV@UVA.NL University of Amsterdam, Amsterdam, the Netherlands
Pseudocode Yes Algorithm 1 Learning Algorithm
Open Source Code No No explicit statement or link to the open-source code for the described methodology is provided.
Open Datasets Yes We evaluate our approach on crowdsourced data collected for script induction by Regneri et al. (2010), though, in principle, the method is applicable in arguably more general setting of Chambers & Jurafsky (2008).
Dataset Splits Yes We used 4 held-out scenarios to choose model parameters, no scenario-specific tuning was performed, and the 10 test scripts were not used to perform model selection.
Hardware Specification No No specific hardware specifications are mentioned in the paper.
Software Dependencies No We initialize word representations using the SENNA embeddings (Collobert et al., 2011).
Experiment Setup Yes We use an online ranking algorithm based on the Perceptron Rank (PRank, (Crammer & Singer, 2001)), or, more accurately, its large-margin extension. One crucial difference though is that the error is computed not only with respect to w but also propagated back through the structure of the neural network. Additionally, we use a Gaussian prior on weights, regularizing both the embedding parameters and the vector w. We initialize word representations using the SENNA embeddings (Collobert et al., 2011). We used 4 held-out scenarios to choose model parameters, no scenario-specific tuning was performed, and the 10 test scripts were not used to perform model selection.