Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
Image Super-Resolution with Guarantees via Conformalized Generative Models
Authors: Eduardo Adame, Daniel Csillag, Guilherme Tegoni Goedert
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We empirically evaluate these results and establish our method s solid performance. 4 Experiments We dedicate this section to the empirical evaluation of our method, demonstrating its effectiveness. |
| Researcher Affiliation | Academia | Eduardo Adame School of Applied Mathematics Getulio Vargas Foundation EMAIL Daniel Csillag School of Applied Mathematics Getulio Vargas Foundation EMAIL Guilherme Tegoni Goedert School of Applied Mathematics Getulio Vargas Foundation EMAIL |
| Pseudocode | Yes | The relevant pseudocodes can be found in the supplementary material. C Pseudocodes Algorithm 1 Conformal mask calibration (computation of tĪ±) with dynamic programming Algorithm 2 Conformal mask calibration (computation of tĪ±) with a brute force search |
| Open Source Code | Yes | For reproducibility, the source code is available in https://github.com/adamesalles/experimen ts-conformal-superres, as well as in the supplementary material. |
| Open Datasets | Yes | Data All experiments were conducted using the Liu4K dataset [Liu et al., 2020] |
| Dataset Splits | Yes | All experiments were conducted using the Liu4K dataset [Liu et al., 2020], which contains 1,600 high-resolution (4K) images in the training set and an additional 400 4K images in the validation set. [...] We use the training set for calibration procedures, and the test set for evaluation and metrics. |
| Hardware Specification | Yes | Compute Experiments were run on an Intel Xeon E5-2696 v2 processor (2.5GHz base, 3.6GHz boost, 18 threads available) with 40GB of RAM and an NVIDIA RTX 6000 Ada Generation 48GB GPU. |
| Software Dependencies | No | The paper mentions the use of 'Sin SR [Wang et al., 2023]' as a base model but does not specify version numbers for any ancillary software dependencies like Python, PyTorch, or CUDA. |
| Experiment Setup | No | The paper describes its method and calibration procedure, including configurable parameters like fidelity levels (Ī±) and choices of Dp, but does not provide specific hyperparameters (e.g., learning rate, batch size, epochs, optimizer settings) for training a model in its main text. |