[Endorsement Request] cs.LG — When does a predictive system learn it's the one predicting? (40 experiments, 12 falsified hypotheses)

Research
1

Hi everyone,

I’m an independent researcher looking for an arXiv endorsement for cs.LG.

My paper asks: under what conditions does a minimal neural system (192-dim GRU, <100K params) learn to distinguish self-caused from world-caused changes? We propose agency gain — the predictive gap between a self-aware and a self-blind predictor sharing the same hidden state — as a directly trainable, density-free alternative to empowerment.

Key results from 40 controlled experiments:

  • A 93.7% prediction gap survives removal of all auxiliary components

  • Awareness must consolidate before intention (simultaneous learning fails in all configurations)

  • 12 falsified hypotheses mapping the boundary between “systems that predict” and “systems that know they are the ones predicting”

  • Forward-sampled action selection succeeds; two gradient-based alternatives degenerate

Preprint: https://doi.org/10.5281/zenodo.20523162

If you’re qualified to endorse for cs.LG and find this relevant, my endorsement code is 7XJ7XN.

https://arxiv.org/auth/endorse?x=7XJ7XN

Endorsement confirms the work is legitimate scholarly material, not an endorsement of conclusions.

Happy to answer any questions about the paper. Thanks!

2

“This request has been fulfilled. Could a moderator please close or delete this topic? Thanks!”

3

I am a researcher and academic staff member at Lucian Blaga University of Sibiu, Romania, preparing my first arXiv submission in cs.LG (Machine Learning).

Manuscript:
Leakage-Aware Benchmarking of Machine Learning for Food Spectroscopy: How Validation Design Changes Reported Predictive Performance

The study investigates how validation design affects reported ML performance in food spectroscopy, comparing random splitting, group-aware validation, batch-aware validation, and external/domain-shift validation.

I am looking for someone who is currently eligible to endorse submissions in cs.LG and is willing to inspect the manuscript before deciding.

To avoid asking anyone to follow an unfamiliar authentication link, here is only the official arXiv endorsement code:

Subject: cs.LG
Endorsement Code: 3MLSXP

The code can be entered directly after logging into arXiv through the official endorsement interface.

I can also provide my institutional profile, ORCID, publication record, or any additional information required.

Thank you for considering the request.

Ketney Otto