Evaluating Prerequisite Qualities for Learning End-to-end Dialog Systems

Authors: [data] Jesse Dodge, Andreea Gane, Xiang Zhang, Antoine Bordes, Sumit Chopra, Alexander Miller, Arthur Szlam, Jason Weston

ICLR 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We present results of various models on these tasks, and evaluate their performance.
Researcher Affiliation Industry Jesse Dodge , Andreea Gane , Xiang Zhang , Antoine Bordes, Sumit Chopra, Alexander H. Miller, Arthur Szlam & Jason Weston Facebook AI Research 770 Broadway New York, USA {jessedodge,agane,xiangz,abordes,spchopra,ahm,aszlam,jase}@fb.com
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks.
Open Source Code No The paper mentions 'We used the code available at: https://github.com/facebook/SCRNNs' (Footnote 9), which is for a baseline LSTM model, not the open-source code for the paper's main methodology or models.
Open Datasets Yes We provide a dataset covering 75k movie entities and with 3.5M training examples. (...) The datasets are available at: http://fb.ai/babi.
Dataset Splits Yes We split the questions into training, development and test sets with 96k, 10k and 10k examples, respectively.
Hardware Specification No The paper does not specify any particular hardware used for running experiments, such as GPU/CPU models or cloud instance types.
Software Dependencies No The paper states 'All models are implemented in the Torch library (see torch.ch)' but does not provide a specific version number for Torch or any other software dependencies.
Experiment Setup Yes Hyperparameters of all learning models have been set using grid search on the validation set. The main hyperparameters are embedding dimension d, learning rate λ, number of dictionaries w, number of hops K for Mem NNs and unfolding depth blen for LSTMs.