Neural Production Systems

Authors: Anirudh Goyal ALIAS PARTH GOYAL, Aniket Didolkar, Nan Rosemary Ke, Charles Blundell, Philippe Beaudoin, Nicolas Heess, Michael C. Mozer, Yoshua Bengio

NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In a series of experiments, we demonstrate that this architecture achieves a flexible, dynamic flow of control and serves to factorize entity-specific and rule-based information. This disentangling of knowledge achieves robust future-state prediction in rich visual environments, outperforming state-of-the-art methods using GNNs, and allows for the extrapolation from simple (few object) environments to more complex environments.
Researcher Affiliation Collaboration 1 Mila, University of Montreal, 2 Google Deepmind, 3 Waverly, 4 Google Research, Brain Team.
Pseudocode No The paper mentions 'detailed algorithms for the sequential and parallel regimes in Appendix' but does not show them in the provided text. No explicit 'Pseudocode' or 'Algorithm' block is visible in the provided content.
Open Source Code Yes We include the code for reproducing our experiments
Open Datasets Yes We validate our model on a colored version of this dataset. This is a next-step prediction task in which the model is tasked with predicting the final binary mask of each ball. We compare the following methods: (a) SCOFF (Goyal et al., 2020)
Dataset Splits Yes We consider two evaluation settings. (1) Test setting: The number of rollout timesteps is same as that seen during training (i.e. t = 15); (2) Transfer Setting: The number of rollout timesteps is higher than that seen during training (i.e. t = 30).
Hardware Specification No The paper does not provide specific hardware details (e.g., specific GPU/CPU models, processor types, or memory details) used for running its experiments. It only mentions 'Google cloud credits' which is not specific enough.
Software Dependencies No The paper does not provide specific ancillary software details with version numbers (e.g., library or solver names with version numbers like Python 3.8, CPLEX 12.4) needed to replicate the experiment.
Experiment Setup Yes Here we briefly outline the tasks considered and direct the reader to the Appendix for full details on each task and details on hyperparameter settings.