Interaction Networks for Learning about Objects, Relations and Physics
Authors: Peter Battaglia, Razvan Pascanu, Matthew Lai, Danilo Jimenez Rezende, koray kavukcuoglu
NeurIPS 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluate its ability to reason about several challenging physical domains: n-body problems, rigid-body collision, and non-rigid dynamics. Our results show it can be trained to accurately simulate the physical trajectories of dozens of objects over thousands of time steps, estimate abstract quantities such as energy, and generalize automatically to systems with different numbers and configurations of objects and relations. |
| Researcher Affiliation | Academia | Anonymous Author(s) Affiliation Address email |
| Pseudocode | No | The paper describes the model mathematically and in prose but does not provide structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not include an explicit statement about releasing open-source code or provide a link to a code repository. |
| Open Datasets | No | Each of the training, validation, test data sets were generated by simulating 2000 scenes over 1000 time steps, and randomly sampling 1 million, 200k, and 200k one-step input/target pairs, respectively. The paper describes generating its own dataset and does not provide concrete access information or cite a publicly available dataset. |
| Dataset Splits | Yes | Each of the training, validation, test data sets were generated by simulating 2000 scenes over 1000 time steps, and randomly sampling 1 million, 200k, and 200k one-step input/target pairs, respectively. The model was trained for 2000 epochs, randomly shuffling the data indices between each. We used mini-batches of 100, and balanced their data distributions so the targets had similar per-element statistics. The performance reported in the Results was measured on held-out test data. ...with a waterfall schedule that began with a learning rate of 0.001 and down-scaled the learning rate by 0.8 each time the validation error, estimated over a window of 40 epochs, stopped decreasing. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware (e.g., CPU, GPU models, memory) used to run the experiments. |
| Software Dependencies | No | The paper mentions 'standard deep neural network building blocks, multilayer perceptrons (MLP), matrix operations, etc.' and 'Adam [13]' as the optimizer, but does not specify any software libraries or their version numbers (e.g., PyTorch, TensorFlow, etc.) used for implementation. |
| Experiment Setup | Yes | The f R MLP had four, 150-length hidden layers, and output length DE = 50. The f O MLP had one, 100-length hidden layer, and output length DP = 2, which targeted the x, y-velocity. We optimized the parameters using Adam [13], with a waterfall schedule that began with a learning rate of 0.001 and down-scaled the learning rate by 0.8 each time the validation error, estimated over a window of 40 epochs, stopped decreasing. We used mini-batches of 100... The model was trained for 2000 epochs... Two forms of L2 regularization were explored... |