Improved Convergence of Score-Based Diffusion Models via Prediction-Correction
Authors: Francesco Pedrotti, Jan Maas, Marco Mondelli
TMLR 2024 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Figure 1: Simulation results for an asymmetric mixture of two gaussians (left), the two moons dataset (center) and the rescaled swiss roll (right). For fixed T1 and variable T2, we plot in blue the W2 distance between the perturbed measure p T1 and the output of (4.1), while we plot in orange the W2 distance between the true distribution p and the output of the algorithm pθ. As expected, both quickly decrease as T2 increases. |
| Researcher Affiliation | Academia | Francesco Pedrotti EMAIL Institute of Science and Technology Austria Jan Maas EMAIL Institute of Science and Technology Austria Marco Mondelli EMAIL Institute of Science and Technology Austria |
| Pseudocode | No | The paper describes the algorithm steps using mathematical equations for stochastic differential equations (SDEs) and their discretizations, but it does not present these steps in a structured pseudocode or algorithm block format with explicit labels like 'Algorithm' or 'Pseudocode'. |
| Open Source Code | No | The paper does not contain any explicit statements about releasing source code for the described methodology, nor does it provide links to a code repository. |
| Open Datasets | No | The paper mentions "an asymmetric mixture of two gaussians", "the two moons dataset", and "the rescaled swiss roll" in the context of simulation results (Figure 1). However, it does not provide concrete access information such as specific links, DOIs, repository names, or formal citations for these datasets. |
| Dataset Splits | No | The paper describes simulation results on several datasets but does not provide any specific information regarding training, validation, or test splits (e.g., percentages, sample counts, or predefined split references). |
| Hardware Specification | No | The paper does not provide specific hardware details such as GPU/CPU models, processor types, or memory amounts used for running its simulations or experiments. |
| Software Dependencies | No | The paper does not provide specific ancillary software details, such as library names with version numbers, needed to replicate the experiment. |
| Experiment Setup | No | The paper describes the theoretical framework and algorithm. While it discusses aspects like step sizes for discretization (e.g., "constant step size"), it does not provide concrete experimental setup details such as specific hyperparameter values, learning rates, batch sizes, or optimizer settings for any practical implementation. |