An Empirical Investigation of Pre-Trained Deep Learning Model Reuse in the Scientific Process

Numerous metascience studies and other initiatives have begun to monitor the prevalence of open science practices when it is more important to understand the 'downstream' effects or impacts of open science. PLOS and DataSeer have developed a new LLM-based indicator to measure an important effect of open science: the reuse of research data. Our results show a data reuse rate of 43%, which is higher than established bibliometric techniques. We show that data reuse can be measured at scale using LLMs and generative artificial intelligence. The positive effects of research data sharing and reuse may currently be underestimated.

Scaling laws describe how language model capabilities grow with compute and data, but say nothing about how long a model matters once released. We introduce time-to-peak and lifespan as measures of model obsolescence and use them to characterize the scientific adoption trajectories of 62 LLMs across more than 108k citing papers (2019-2025), separating active adoption from background citation to recover per-model trajectories that citation counts cannot resolve. We find that a model's longevity is shaped more by when it was released than by its characteristics: release year predicts time-to-peak and lifespan more strongly than architecture, openness, or scale. LLM adoption follows an inverted-U curve (rising after release, peaking, and then declining), but this pattern is rapidly compressing. Each successive release year is associated with a 27% shorter time-to-peak and a 23% shorter lifespan ($p < 0.001$), robust to minimum-age thresholds and controls for model size. These adoption-side dynamics are invisible to scaling laws and suggest that specialization on any single model may be a depreciating investment, with costs falling on reproducibility and migration.

Persistence is often regarded as a virtue in science. In this paper, however, we challenge this conventional view by highlighting its contextual nature, particularly how persistence can become a liability during paradigm shifts. We focus on the deep learning revolution catalyzed by AlexNet in 2012. Analyzing the 20-year career trajectories of more than 5,000 scientists active in top machine learning venues during the preceding decade, we examine how their research focus and output evolved. We first uncover a dynamic period in which leading venues increasingly prioritized cutting-edge deep learning developments, displacing traditional statistical learning methods. Scientists responded to these changes in markedly different ways: those who were previously successful or affiliated with established teams adapted more slowly. Such persistence is positively associated with productivity but, after 2012, negatively associated with scientific impact. Most researchers, and the largest share of the field's output, cluster in a band of moderate persistence, pointing to a trade-off between output and impact, as well as to institutional frictions that make larger departures costly. These conclusions are robust to alternative identification strategies and to competing explanations such as topic popularity premiums and survivorship bias. Taken together, our macro- and micro-level findings suggest that, in this case, a paradigm shift creates an opportunity structure by devaluing the very expertise that conferred incumbents' advantage in the first place.