Learning low-dimensional state embeddings and metastable clusters from time series data
Authors: Yifan Sun, Yaqi Duan, Hao Gong, Mengdi Wang
NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiment on a simulated dynamical system shows that the state clustering method indeed reveals metastable structures. We also experiment with time series generated by layers of a Deep-Q-Network when playing an Atari game. |
| Researcher Affiliation | Academia | Yifan Sun Carnegie Mellon University yifans@andrew.cmu.edu Yaqi Duan Princeton University yaqid@princeton.edu Hao Gong Princeton University hgong@princeton.edu Mengdi Wang Princeton University mengdiw@princeton.edu |
| Pseudocode | Yes | Algorithm 1: Reshaping the Kernel Mean Embedding. [...] Algorithm 2: Learning State Embedding [...] Algorithm 3: Learning metastable state clusters |
| Open Source Code | No | The paper does not provide any statement or link indicating that source code for the described methodology is publicly available. |
| Open Datasets | No | The paper describes generating data for simulated diffusion processes and using time series data generated by a Deep Q-Network, but it does not provide concrete access information (e.g., specific links, DOIs, or citations to public repositories) for these datasets as they were used in their experiments. It mentions using data 'generated by the last hidden layer of a trained DQN'. |
| Dataset Splits | No | The paper describes the data used (e.g., 'trajectories of length n = 10^6' for diffusion, 'raw data is a time series of length 47936' for DQN) but does not specify how this data was split into training, validation, or test sets for their analysis. It mentions 'sample transition pairs {(Xt, Xt+1)}' but not their partitioning for evaluation. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., GPU/CPU models, memory) used for running its experiments. It mentions training a Deep Q-Network, but not the hardware for their own analysis. |
| Software Dependencies | No | The paper mentions using Gaussian kernels and t-SNE for visualization, but it does not specify any software names with version numbers (e.g., Python, PyTorch, TensorFlow versions or specific library versions) that would be needed to reproduce the experiments. |
| Experiment Setup | Yes | We conduct the state embedding and clustering procedures with rank r = 4. Figure 5.1 (c) shows the clustering results for τ = 1 with a varying number of clusters. [...] The raw data is a time series of length 47936 and dimension 512, comprising 130 game trajectories. We apply the state embedding method by approximating the Gaussian kernel with 200 random Fourier features. |