mlsanityData to 14 Oct 2026

InterFeat: A Pipeline for Finding Interesting Scientific Features

arXiv 2505.13534 ↗First posted 18 May 2025q-bio.QM cs.AI cs.CLcited by 0

Not on the map. Why?

Abstract (arXiv)

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

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 21 of its 88 references that are in the map. First posted on arXiv 18 May 2025.

Builds on

21 of 88 references.

References · 21

  1. 2024The Llama 3 Herd of Modelscited by 14,715
  2. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  3. 2017A Unified Approach to Interpreting Model Predictionscited by 6,555
  4. 2024Benchmarking Retrieval-Augmented Generation for Medicinecited by 331
  5. 2019ScispaCy: Fast and Robust Models for Biomedical Natural Language Processingcited by 244
  6. 2022Can large language models reason about medical questions?cited by 194
  7. 2023SciMON: Scientific Inspiration Machines Optimized for Noveltycited by 179
  8. 2024Reducing hallucination in structured outputs via Retrieval-Augmented Generationcited by 93
  9. 2024Medical Graph RAG: Towards Safe Medical Large Language Model via Graph Retrieval-Augmented Generationcited by 90
  10. 2024OpenScholar: Synthesizing Scientific Literature with Retrieval-augmented LMscited by 72
  11. 2024Hypothesis Generation with Large Language Modelscited by 71
  12. 2024Large Language Models as Biomedical Hypothesis Generators: A Comprehensive Evaluationcited by 34
9 more
  1. 2024LLM-Select: Feature Selection with Large Language Modelscited by 30
  2. 2024Superhuman performance of a large language model on the reasoning tasks of a physiciancited by 28
  3. 2022BioLORD: Learning Ontological Representations from Definitions (for Biomedical Concepts and their Textual Descriptions)cited by 21
  4. 2024Automating psychological hypothesis generation with AI: when large language models meet causal graphcited by 15
  5. 2024Do causal predictors generalize better to new domains?cited by 11
  6. 2023Whats next? Forecasting scientific research trendscited by 5
  7. 2018Accelerating Prototype-Based Drug Discovery using Conditional Diversity Networkscited by 4
  8. 2018Are Abstracts Enough for Hypothesis Generation?cited by 3
  9. 2024LLpowershap: Logistic Loss-based Automated Shapley Values Feature Selection Methodcited by 2

Built on by

Not cited yet.