InterFeat: A Pipeline for Finding Interesting Scientific Features

Dan Ofer, Michal Linial, Dafna Shahaf
5/18/2025
q-bio.QMcs.AIcs.CL

Official code: LinialLab/InterFeat (5 stars, last commit 2025-09-21).

Abstract

Finding interesting phenomena is the core of scientific discovery, but it is a manual, ill-defined concept. We present an integrative pipeline for automating the discovery of interesting simple hypotheses (feature-target relations with effect direction and a potential underlying mechanism) in structured biomedical data. The pipeline combines machine learning, knowledge graphs, literature search and Large Language Models. We formalize "interestingness" as a combination of novelty, utility and plausibility. On 8 major diseases from the UK Biobank, our pipeline consistently recovers risk factors years before their appearance in the literature. 40--53% of our top candidates were validated as interesting, compared to 0--7% for a SHAP-based baseline. Overall, 28% of 109 candidates were interesting to medical experts. The pipeline addresses the challenge of operationalizing "interestingness" scalably and for any target. We release data and code: https://github.com/LinialLab/InterFeat

View on arXivView PDF

Code Implementations(1)

LinialLab/InterFeatOfficial100%

Automatically uncover Interesting (novel, plausible, and useful) features in biomedical data by combining statistical filtering, literature mining, knowledge graphs and LLMs on UK BioBank tabular features

51Python, Jupyter NotebookSep 21, 20251 years ago

Similar Papers

Interestingness First Classifiers

80% match2025

Large Language Models as Biomedical Hypothesis Generators: A Comprehensive Evaluation

78% match1 repo152024

Enhancing Biomedical Knowledge Discovery for Diseases: An Open-Source Framework Applied on Rett Syndrome and Alzheimer's Disease

78% match1 repo42024

Explainable Biomedical Hypothesis Generation via Retrieval Augmented Generation enabled Large Language Models

78% match2024

Large Language Models are Zero Shot Hypothesis Proposers

77% match2023

An LLM-based Knowledge Synthesis and Scientific Reasoning Framework for Biomedical Discovery

77% match1 repo32024

Cite this paper

@article{ofer2025interfeat,
  title  = {InterFeat: A Pipeline for Finding Interesting Scientific Features},
  author = {Dan Ofer and Michal Linial and Dafna Shahaf},
  year   = {2025},
  eprint = {2505.13534},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2505.13534v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2505.13534.svg)](https://codeofpaper.com/paper/2505.13534)

Discussion