On Path Integration of Grid Cells: Group Representation and Isotropic Scaling
Authors: Ruiqi Gao, Jianwen Xie, Xue-Xin Wei, Song-Chun Zhu, Ying Nian Wu
NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Finally, with our optimization-based approach, we manage to learn hexagon grid patterns that share similar properties of the grid cells in the rodent brain. The learned model is capable of accurate long distance path integration. Code is available at https://github.com/ruiqigao/grid-cell-path. We conduct numerical experiments to learn the representations as described in Section 5. |
| Researcher Affiliation | Collaboration | Ruiqi Gao1 ruiqigao@ucla.edu Jianwen Xie2 jianwen@ucla.edu Xue-Xin Wei3 weixx@utexas.edu Song-Chun Zhu1,4,5 sczhu@stat.ucla.edu Ying Nian Wu1 ywu@stat.ucla.edu 1Department of Statistics, UCLA 2Cognitive Computing Lab, Baidu Research 3Department of Neuroscience, UT Austin 4Department of Computer Science, UCLA 5Beijing Institute for General Artificial Intelligence (BIGAI) |
| Pseudocode | No | The paper does not include any explicitly labeled 'Pseudocode' or 'Algorithm' blocks, nor does it present structured steps formatted like code. |
| Open Source Code | Yes | Code is available at https://github.com/ruiqigao/grid-cell-path. |
| Open Datasets | No | The paper uses a square environment with size 1m × 1m, which is discretized into a 40 × 40 lattice. For A(x,x'), we use a Gaussian adjacency kernel with σ = 0.07. This input data is generated/defined within the paper's experimental setup and is not drawn from a public, pre-existing dataset with an associated link, DOI, or citation. |
| Dataset Splits | No | The paper describes the environment setup and training process but does not specify explicit training, validation, and test splits for a dataset. The evaluation is focused on the characteristics of the learned patterns and path integration performance. |
| Hardware Specification | No | The paper does not specify the exact hardware (e.g., GPU/CPU models, memory) used for running its experiments within the main text. |
| Software Dependencies | No | The paper mentions using the 'Adam [25] optimizer' but does not specify version numbers for any software dependencies, programming languages, or libraries used for implementation. |
| Experiment Setup | Yes | Specifically, we use a square environment with size 1m × 1m, which is discretized into a 40 × 40 lattice. For direction, we discretize the circle [0,2π] into 144 directions and use nearest neighbor linear interpolations for values in between. The displacement Δr are sampled within a small range, i.e., Δr is smaller than 3 grids on the lattice. For A(x,x'), we use a Gaussian adjacency kernel with σ = 0.07. v(x) is of d = 192 dimensions, which is partitioned into K = 16 modules, each of which has 12 cells. |