Learning interaction rules from multi-animal trajectories via augmented behavioral models
Authors: Keisuke Fujii, Naoya Takeishi, Kazushi Tsutsui, Emyo Fujioka, Nozomi Nishiumi, Ryoya Tanaka, Mika Fukushiro, Kaoru Ide, Hiroyoshi Kohno, Ken Yoda, Susumu Takahashi, Shizuko Hiryu, Yoshinobu Kawahara
NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In experiments using synthetic datasets, our method achieved better performance than various baselines. We then analyzed multi-animal datasets of mice, flies, birds, and bats, which verified our method and obtained novel biological insights. |
| Researcher Affiliation | Academia | Keisuke Fujii Nagoya University RIKEN Center for Advanced Intelligence Project JST PRESTO; Naoya Takeishi University of Applied Sciences and Arts Western Switzerland RIKEN Center for Advanced Intelligence Project; Kazushi Tsutsui Nagoya University Emyo Fujioka Doshisha University Nozomi Nishiumi National Institute for Basic Biology; Ryoya Tanaka Nagoya University; Mika Fukushiro Doshisha University Kaoru Ide Doshisha University Hiroyoshi Kohno Tokai University Ken Yoda Nagoya University; Susumu Takahashi Doshisha University Shizuko Hiryu Doshisha University Yoshinobu Kawahara Kyushu University RIKEN Center for Advanced Intelligence Project |
| Pseudocode | No | The paper does not contain any clearly labeled pseudocode or algorithm blocks. |
| Open Source Code | Yes | Our code is available at https://github.com/keisuke198619/ABM. |
| Open Datasets | No | The paper mentions synthetic datasets (Kuramoto model, boid model) and multi-animal datasets (mice, birds, bats, flies) but does not provide specific links, DOIs, or formal citations for public access to these datasets beyond mentioning their names. For instance, the boid model simulation data is described as |
| Dataset Splits | Yes | The hyperparameters of the models were determined by validation datasets in the synthetic data experiments (for the details, see Appendices E and G). Each method was trained only on one sequence according to most neural GC frameworks [73, 31, 45]. |
| Hardware Specification | No | The paper mentions computational resources (e.g., in Appendix D) but does not specify the exact hardware (e.g., GPU/CPU models, memory) used for running the experiments. It states: "The common training details, (binary) inference methods, computational resources, and the amount of computation are described in Appendix D." |
| Software Dependencies | No | The paper does not provide specific version numbers for any software dependencies. While it mentions Adam optimizer [32], it doesn't specify Python or library versions. |
| Experiment Setup | Yes | The hyperparameters of the models were determined by validation datasets in the synthetic data experiments (for the details, see Appendices E and G). The common training details, (binary) inference methods, computational resources, and the amount of computation are described in Appendix D. |