Re-examination of the Role of Latent Variables in Sequence Modeling
Authors: Guokun Lai, Zihang Dai, Yiming Yang, Shinjae Yoo
NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | To better understand this discrepancy, we re-examine the roles of latent variables in stochastic recurrent models for speech density estimation. Our analysis reveals that under the restriction of fully factorized output distribution in previous evaluations, the stochastic variants were implicitly leveraging intrastep correlation but the deterministic recurrent baselines were prohibited to do so, resulting in an unfair comparison. To correct the unfairness, we remove such restriction in our re-examination, where all the models can explicitly leverage intra-step correlation with an auto-regressive structure. Over a diverse set of univariate and multivariate sequential data, including human speech, MIDI music, handwriting trajectory and frame-permuted speech, our results show that stochastic recurrent models fail to deliver the performance advantage claimed in previous work. |
| Researcher Affiliation | Academia | 1Carnegie Mellon University, 2Brookhaven National Laboratory |
| Pseudocode | No | The paper does not contain any pseudocode or algorithm blocks. |
| Open Source Code | Yes | For more implementation details, please refer to the Supplementary D as well as the source code1. 1https://github.com/zihangdai/reexamine-srnn |
| Open Datasets | Yes | we replicate the experiments under the setting introduced above and evaluate them on three speech datasets, namely TIMIT, VCTK, and Blizzard. [...] Besides speech sequences, we additionally consider three more types of multivariate sequences with different patterns of intra-step correlation, they are MIDI sound sequence data (including Muse and Nottingham datasets), handwriting trajectory data (IAM-On DB) and the Perm-TIMIT dataset. |
| Dataset Splits | No | The paper mentions 'training set' and 'test log-likelihood' but does not explicitly specify validation splits or proportions for any of the datasets. |
| Hardware Specification | No | The paper does not specify the hardware used for running experiments (e.g., specific GPU or CPU models). |
| Software Dependencies | No | The paper does not provide specific software dependencies with version numbers. |
| Experiment Setup | No | The paper states 'For more implementation details, please refer to the Supplementary D as well as the source code' but does not include specific hyperparameter values or training configurations in the main text. |