Summary of Latent Space Energy-based Neural Odes, by Sheng Cheng et al.
Latent Space Energy-based Neural ODEs
by Sheng Cheng, Deqian Kong, Jianwen Xie, Kookjin Lee, Ying Nian Wu, Yezhou Yang
First submitted to arxiv on: 5 Sep 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Machine Learning (stat.ML)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper introduces a novel deep dynamical modeling approach for continuous-time sequences. The neural emission model generates each data point through a non-linear transformation of a latent state vector, which is implicitly defined by a neural ODE with an informative prior distribution parameterized by an Energy-based model (EBM). To disentangle dynamic states from underlying static factors, the framework represents time-invariant variables in the latent space. The model is trained using maximum likelihood estimation with Markov chain Monte Carlo (MCMC) in an end-to-end manner. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper creates a new way to understand and predict continuous-time sequences. It uses special neural networks that transform information into a hidden state, which then changes over time according to a set of rules defined by another neural network. The model also includes a prior understanding of what the initial state might be. This helps the model learn more about static factors that don’t change over time, and how they relate to dynamic states. The researchers tested their approach on different types of data, including videos and real-world sequences, and found it outperformed existing methods. |
Keywords
» Artificial intelligence » Energy based model » Latent space » Likelihood » Neural network
