mlsanityData to 13 Oct 2026

DiffKnock: Diffusion-based Knockoff Statistics for Neural Networks Inference

arXiv 2510.01418 ↗First posted 1 Oct 2025stat.ME stat.AP stat.MLcited by 0

Not on the map. Why?

Abstract (arXiv)

We introduce DiffKnock, a diffusion-based knockoff framework for high-dimensional feature selection with finite-sample false discovery rate (FDR) control. DiffKnock addresses two key limitations of existing knockoff methods: preserving complex feature dependencies and detecting non-linear associations. Our approach trains diffusion models to generate valid knockoffs and uses neural network--based gradient and filter statistics to construct antisymmetric feature importance measures. Through simulations, we showed that DiffKnock achieved higher power than autoencoder-based knockoffs while maintaining target FDR, indicating its superior performance in scenarios involving complex non-linear architectures. Applied to murine single-cell RNA-seq data of LPS-stimulated macrophages, DiffKnock identifies canonical NF-kappaB target genes (Ccl3, Hmox1) and regulators (Fosb, Pdgfb). These results highlight that, by combining the flexibility of deep generative models with rigorous statistical guarantees, DiffKnock is a powerful and reliable tool for analyzing single-cell RNA-seq data, as well as high-dimensional and structured data in other domains.

Key facts. No citing papers in the map yet (data through 13 Oct 2026). Builds on 15 of its 35 references that are in the map. First posted on arXiv 1 Oct 2025.

Builds on

15 of 35 references.

References · 15

  1. 2020Denoising Diffusion Probabilistic Modelscited by 20,952
  2. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,023
  3. 2021Improved Denoising Diffusion Probabilistic Modelscited by 3,751
  4. 2017Axiomatic Attribution for Deep Networkscited by 3,368
  5. 2013Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Mapscited by 2,360
  6. 2017Learning Important Features Through Propagating Activation Differencescited by 1,455
  7. 2017Generative Adversarial Networks: An Overviewcited by 589
  8. 2016Panning for Gold: Model-X Knockoffs for High-dimensional Controlled Variable Selectioncited by 120
  9. 2014Controlling the false discovery rate via knockoffscited by 97
  10. 2018IntroVAE: Introspective Variational Autoencoders for Photographic Image Synthesiscited by 85
  11. 2018Gaussian Process Prior Variational Autoencoderscited by 72
  12. 2018DeepPINK: reproducible feature selection in deep neural networkscited by 41
3 more
  1. 2023Catch Missing Details: Image Reconstruction with Frequency Augmented Variational Autoencodercited by 18
  2. 2021Preventing Oversmoothing in VAE via Generalized Variance Parameterizationcited by 15
  3. 2021PriorVAE: Encoding spatial priors with VAEs for small-area estimationcited by 11

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025LapDDPM: A Conditional Graph Diffusion Model for scRNA-seq Generation with Spectral Adversarial Perturbations
  2. 2026CRNDiff: Count-Native Diffusion Framework via Chemical Reaction Networks
  3. 2024scRDiT: Generating single-cell RNA-seq data by diffusion transformers and accelerating sampling
  4. 2026Improving scDiffusion with Sparsity-Biased Classifier-Free Guidance
  5. 2026ScDiVa: Masked Discrete Diffusion for Joint Modeling of Single-Cell Identity and Expression
  6. 2024White-Box Diffusion Transformer for single-cell RNA-seq generation
  7. 2025Inference-Time Alignment in Diffusion Models with Reward-Guided Generation: Tutorial and Review
  8. 2026CoDiffGRN: Rethinking Gene Regulatory Network Inference via the BEELINE-KGC Benchmark and Co-evolutionary Discrete Diffusion