Smoothing the Geometry of Probabilistic Box Embeddings

Authors: Xiang Li, Luke Vilnis, Dongxu Zhang, Michael Boratko, Andrew McCallum

ICLR 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental 5 EXPERIMENTS We perform experiments on the Word Net hypernym prediction task in order to evaluate the performance of these improvements in practice.
Researcher Affiliation Academia Xiang Li , Luke Vilnis , Dongxu Zhang, Michael Boratko & Andrew Mc Callum College of Information and Computer Sciences University of Massachusetts Amherst
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks clearly labeled as 'Pseudocode' or 'Algorithm'.
Open Source Code Yes Detailed hyperparameter settings and code to reproduce experiments can be found at https://github.com/Lorraine333/smoothed_box_embedding.
Open Datasets Yes We perform experiments on the Word Net hypernym prediction task... We apply our method to a market-basket task constructed using the Movie Lens dataset. Here, the task is to predict users preference for movie A given that they liked movie B. We first collect all pairs of user-movie ratings higher than 4 points (strong preference) from the Movie Lens-20M dataset.
Dataset Splits Yes We used the same train/dev/test split as in Vendrov et al. (2016). ... The training data contains 1,176 positive examples, and the dev and test sets contain 209 positive examples.
Hardware Specification No The paper does not provide specific hardware details such as GPU/CPU models, memory specifications, or cloud instance types used for running its experiments.
Software Dependencies No The paper does not explicitly list specific software dependencies with their version numbers, only generally stating that code is available to reproduce experiments.
Experiment Setup No The paper mentions that 'Detailed hyperparameter settings and code to reproduce experiments can be found at https://github.com/Lorraine333/smoothed_box_embedding', but it does not provide specific hyperparameter values or concrete system-level training configurations directly in the main text.