Model Crisis

Preprints that amplify unresolved contradictions, exposing structural weaknesses in prevailing theories and signaling the breakdown of explanatory power.

Generative Knowledge Production Pipeline Driven by Academic Influencers

URL https://arxiv.org/pdf/2505.24681.pdf Stage Model Drift / Model Crisis Paradigm framing The preprint focuses on the paradigm of academic knowledge production, specifically addressing the integration of generative AI, particularly ChatGPT, into research workflows and publication processes. It examines the dynamic interplay between established academic integrity norms and the transformative potential of AI-driven tools, highlighting the role […]

Generative Knowledge Production Pipeline Driven by Academic Influencers Read More »

Dust Budget Crisis in Little Red Dots

URL https://arxiv.org/pdf/2505.22600.pdf Stage Model Drift / Model Crisis Paradigm framing The preprint challenges the prevailing paradigm that Little Red Dots (LRDs) are heavily dust-reddened Active Galactic Nuclei (AGNs). The current paradigm, based on their red spectra, suggests high extinction levels (Av ~ 3). Highlights The authors present evidence suggesting a “dust budget crisis,” where the

Dust Budget Crisis in Little Red Dots Read More »

The Perspectivist Paradigm Shift: Assumptions and Challenges of Capturing Human Labels

URL Stage https://arxiv.org/abs/2405.05860 Model Drift/Model Crisis Paradigm framing The preprint focuses on the paradigm shift in data labeling from the traditional “ground truth” paradigm to the “perspectivist” paradigm, particularly within the field of Natural Language Processing. Highlights The preprint argues for a “perspectivist turn” in machine learning, challenging the established paradigm that treats annotator disagreement

The Perspectivist Paradigm Shift: Assumptions and Challenges of Capturing Human Labels Read More »