Papers

Seven papers at NeurIPS. Seven reproductions at ReScience.

Original work across scientific ML, medical imaging, energy, chip design and computational creativity, and careful reproductions of other people's results, including the ones that didn't hold up.

NeurIPS 2026

Workshops and the Creative AI track · December

01
When Does a Fourier Neural Operator Fail? Spectral Diagnostics for Low-Rank Surrogates of Reaction–Diffusion
Scientific MLRPS
Accepted
02
A Structural Load Model Cannot Predict the Effect of a Conservation Appeal
Energy demandFMTS
Accepted
03
Cartographic Generalization Exposes Non-Uniform Structure Loss in Downsampled Dermoscopy Benchmarks
Medical imagingAIDaR
Accepted
04
Contaminated Ancestors: Seven Ways Our Own Evaluation Protocol Nearly Told Us the Wrong Thing
Evaluation methodsTAE
Accepted
05
Who's to Blame? Actual Causation and Congestion Pricing for Explaining Macro-Placement Failures
Chip designAI4ChipDesign
Accepted
06
Restored Lines: A Model Finds the Poet's Hardest Line Breaks and Declines to Write Them
Computational creativityCreative AI
Accepted
07
The Dead Ear: A Language Model's Inherited Accent Is an Archive, Not an Author
Computational creativityCreative AI
Accepted

ReScience C

Reproductions of published results

01
A Novel CNN Ablation Reveals Limited Fourier Inductive Bias in LiteFNO: A Reproducibility Study
Scientific ML[Re] Reproduced
Under review
02
Association of genomic subtypes of lower-grade gliomas with shape features automatically extracted by a deep learning algorithm
Medical imaging[¬Re] Did not reproduce
Under review
03
Diffusion-Based Hierarchical Multi-Label Object Detection to Analyze Panoramic Dental X-rays
Dental imaging[¬Re] Did not reproduce
Under review
04
Improved survival analysis by learning shared genomic information from pan-cancer data
Cancer genomics[Re] Reproduced
Under review
05
MedMNIST v2: Replicating the ResNet baselines and auditing what macro-AUC hides on DermaMNIST
Medical imaging[Re] Reproduced
Under review
06
Robust deep learning–based protein sequence design using ProteinMPNN
Protein design[Re] Reproduced
Under review
07
TB-Net: A Tailored, Self-Attention Deep Convolutional Neural Network Design for Detection of Tuberculosis Cases from Chest X-Ray Images
Medical imaging[Re] Reproduced
Under review

Want your name on next year's list?

Apply to the cohort

Seven papers at NeurIPS. Seven reproductions at ReScience.

When Does a Fourier Neural Operator Fail? Spectral Diagnostics for Low-Rank Surrogates of Reaction–Diffusion

A Structural Load Model Cannot Predict the Effect of a Conservation Appeal

Cartographic Generalization Exposes Non-Uniform Structure Loss in Downsampled Dermoscopy Benchmarks

Contaminated Ancestors: Seven Ways Our Own Evaluation Protocol Nearly Told Us the Wrong Thing

Who's to Blame? Actual Causation and Congestion Pricing for Explaining Macro-Placement Failures

Restored Lines: A Model Finds the Poet's Hardest Line Breaks and Declines to Write Them

The Dead Ear: A Language Model's Inherited Accent Is an Archive, Not an Author

A Novel CNN Ablation Reveals Limited Fourier Inductive Bias in LiteFNO: A Reproducibility Study

Association of genomic subtypes of lower-grade gliomas with shape features automatically extracted by a deep learning algorithm

Diffusion-Based Hierarchical Multi-Label Object Detection to Analyze Panoramic Dental X-rays

Improved survival analysis by learning shared genomic information from pan-cancer data

MedMNIST v2: Replicating the ResNet baselines and auditing what macro-AUC hides on DermaMNIST

Robust deep learning–based protein sequence design using ProteinMPNN

TB-Net: A Tailored, Self-Attention Deep Convolutional Neural Network Design for Detection of Tuberculosis Cases from Chest X-Ray Images

Want your name on next year's list?