Skip to content
DraconiticResearch
Search Sources API
203,234 studies31 sourcesUpdated 42 min ago
PreprintCC BY-NC-NDChemRxiv

A Reported Featurizer Ranking Instability Is Noise Between Inseparable Arms

  • Phani Kumar Balagam1
  1. 1Independent researcher
Open at sourcePDFDOICite

Abstract

Representation comparisons are commonly published from one split at one seed on uncurated public sets. Repeating one, the count of endpoints whose best featurizer changes with the split read three, one, one and three under two choices, seeds and deduplication, and every reorder lay between arms whose intervals overlap. Only the fingerprint baseline separated anywhere, by being worse; duplicates placed up to 3.8% of random-split test rows in training. A four-step remedy is given.

Related

  1. Limits of geometric packing descriptors in molecular crystal energy ranking
    Pavlo Bondarenko
    ChemRxivPreprintLicence not statedGeneral chemistryChemRxiv
  2. SolviScope: An Interaction-Aware Graph Neural Network for Solvent Ranking of Small Organic Molecules in Pure and Multicomponent Solvent Systems
    Maansi Jain, Ayan Kashyap, Chanukya Nanduru +1
    ChemRxivPreprintLicence not statedGeneral chemistryChemRxiv
  3. Incorporating Polarization into ML/MM Endpoint Calculations for Protein–Ligand Affinity Ranking
    Leshen Zhang, Xujian Wang, Taoyu Niu +6
    ChemRxivPreprintLicence not statedGeneral chemistryChemRxiv
  4. Ranking versus Absolute Accuracy in Matter–Light Simulation Cost Prediction
    Rithvik Kothuri
    ChemRxivPreprintCC BYFull textGeneral chemistryChemRxiv
  5. Structure-Based Prediction and Ranking of PFAS Groundwater Mobility for EU Drinking Water Regulation.
    Sergio de-la-Huerta-Sainz, Ozge Ozkilinc, Valentín Diez-Cabanes +6
    Environmental science & technologyPeer-reviewedCC BYFull textEnvironmental chemistryEurope PMC

Cite

BibTeXRIS
BibTeX
@misc{balagam2026reported,
  title = {{A Reported Featurizer Ranking Instability Is Noise Between Inseparable Arms}},
  author = {Phani Kumar Balagam},
  howpublished = {ChemRxiv},
  year = {2026},
  month = sep,
  doi = {10.26434/chemrxiv.15009293/v1},
  url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15009293/v1},
  note = {CC BY-NC-ND},
}
RIS
TY  - UNPB
TI  - A Reported Featurizer Ranking Instability Is Noise Between Inseparable Arms
AU  - Phani Kumar Balagam
PY  - 2026
DA  - 2026/09/22
PB  - ChemRxiv
DO  - 10.26434/chemrxiv.15009293/v1
UR  - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15009293/v1
AB  - Representation comparisons are commonly published from one split at one seed on uncurated public sets. Repeating one, the count of endpoints whose best featurizer changes with the split read three, one, one and three under two choices, seeds and deduplication, and every reorder lay between arms whose intervals overlap. Only the fingerprint baseline separated anywhere, by being worse; duplicates placed up to 3.8% of random-split test rows in training. A four-step remedy is given.
ER  -