Counterfactual Dynamics Forecasting – a New Setting of Quantitative Reasoning

Authors: Yanzhu Liu, Ying Sun, Joo-Hwee Lim

AAAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The experimental results on two dynamical systems demonstrate the effectiveness of the proposed method.
Researcher Affiliation Academia Institute for Infocomm Research (I2R) & Centre for Frontier AI Research (CFAR), A*STAR, Singapore
Pseudocode Yes Algorithm 1: Counterfactual Dynamics Forecasting
Open Source Code Yes Code and data are available on https://github.com/yanzhuliu/cf lnn.
Open Datasets Yes Code and data are available on https://github.com/yanzhuliu/cf lnn.
Dataset Splits Yes For each dataset, 800, 100 and 100 pairs of trajectories are used as the training, validation and testing set respectively following (Zhong, Dey, and Chakraborty 2021).
Hardware Specification No The paper does not mention any specific hardware used for running the experiments, such as GPU models, CPU types, or cloud computing instances with their specifications.
Software Dependencies No The paper mentions software components like "odeint is the integration solver (Chen et al. 2018)" and states "All network backbones are implemented in the same way as (Zhong, Dey, and Chakraborty 2021)", but it does not provide specific version numbers for any of these software dependencies.
Experiment Setup No The paper describes the loss function used (mean absolute error in Eq. 5) but does not provide specific hyperparameters such as learning rate, batch size, number of epochs, or optimizer settings, nor other system-level training configurations in the main text.