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..
Solving Inverse Problems with FLAIR
Authors: Julius Erbach, Dominik Narnhofer, Andreas Dombos, Bernt Schiele, Jan Eric Lenssen, Konrad Schindler
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Results on standard imaging benchmarks demonstrate that FLAIR consistently outperforms existing diffusion and flow-based methods in terms of reconstruction quality and sample diversity. Source code is available at https://inverseflair.github.io/. [...] We evaluate the performance of FLAIR in a variety of inverse imaging tasks and compare it against several baselines, using the SD3 backbone without any fine-tuning. |
| Researcher Affiliation | Academia | 1ETH Zürich 2Max Planck Institute for Informatics, Saarland Informatics Campus |
| Pseudocode | Yes | Algorithm 1: The FLAIR solver for inverse imaging problems |
| Open Source Code | Yes | Source code is available at https://inverseflair.github.io/. |
| Open Datasets | Yes | Datasets. We utilize two high-resolution image datasets: FFHQ [23] and DIV2K [2]. [...] Data. We use the publicly available Flickr Faces High Quality dataset [24], which is realeased under the Creative Commons BY 2.0 License and the DIV2K dataset [3], which is released under a research only license. |
| Dataset Splits | Yes | FFHQ consists of 70k diverse face images at 1024 1024 resolution of which we take the first 1000 samples. [...] For FFHQ we use the first 1000 samples of the evaluation dataset and for DIV2K we use the 800 training samples. |
| Hardware Specification | Yes | All experiments were performed on an NVidia RTX 4090 GPU with 24GB of VRAM. |
| Software Dependencies | No | As flow matching model, we us Stable Diffusion 3.5-Medium, which has been released under the Stability Community License. The classifier-free guidance scale is set to 2 for all experiments. To minimize the regularization term, we use stochastic gradient descent with a learning rate of 1. [...] A different motion blur kernel is created for each sample using the Motion Blur package [8], available via github, with kernel size 61 and intensity 0.5. |
| Experiment Setup | Yes | To minimize the regularization term, we use stochastic gradient descent with a learning rate of 1. [...] Super-resolution. We employ bicubic downsampling as the forward operator, as implemented in [53]. The learning rate is set to 12 for 12 super-resolution and to 6 for 8 super-resolution. Motion Deblurring. A different motion blur kernel is created for each sample using the Motion Blur package [8], available via github, with kernel size 61 and intensity 0.5. The learning rate for our data term optimizer is set to 10 1. Inpainting. For inpainting on FFHQ we always use the same rectangular mask at a fixed position, chosen such that it roughly masks out the right side of the face (Figure 3). For DIV2k we also use a fixed mask for all samples, consisting of six randomly generated rectangles. |