arXiv subjects

Gabrielle O'Brien

Publications and source records attributed to Gabrielle O'Brien.

4 recordsLinked to original sources

arXiv2026

A survey of generative AI adoption and perceived productivity among scientists who program

Programming is essential to modern scientific research, yet most scientists report inadequate training for the software development their work demands. Generative AI tools capable of code generation may support scientific programmers, but user studies indicate risks of over-reliance, particularly among inexperienced users. We surveyed 868 scientists who program, examining adoption patterns, tool preferences, and factors associated with perceived productivity. Adoption is highest among students and less experienced programmers, with variation across fields. Scientific programmers overwhelmingly prefer general-purpose conversational interfaces like ChatGPT over developer-specific tools. Both inexperience and limited use of development practices (like testing, code review, and version control) are associated with greater perceived productivity -- but these factors interact, suggesting formal practices may partially compensate for inexperience. The strongest predictor of perceived productivity is the number of lines of generated code typically accepted at once. These findings suggest scientific programmers using generative AI may gauge productivity by code generation rather than validation.

cs.SE↗

arXiv2026

User Misconceptions of LLM-Based Conversational Programming Assistants

Programming assistants powered by large language models (LLMs) have become widely available, with conversational assistants like ChatGPT particularly accessible to novice programmers. However, varied tool capabilities and inconsistent availability of extensions (web search, code execution, retrieval-augmented generation) create opportunities for user misconceptions that may lead to over-reliance, unproductive practices, or insufficient quality control. We characterize misconceptions that users of conversational LLM-based assistants may have in programming contexts through a two-phase approach: first brainstorming and cataloging potential misconceptions, then conducting qualitative analysis of Python-programming conversations from the WildChat dataset. We find evidence that users have misplaced expectations about features like web access, code execution, and non-text outputs. We also note the potential for deeper conceptual issues around information requirements for debugging, validation, and optimization. Our findings reinforce the need for LLM-based tools to more clearly communicate their capabilities to users and empirically ground aspects that require clarification in programming contexts.

cs.HC↗

arXiv2025

Cellular Mechanisms of Phase Maintenance in a Pyloric Motif of a Central Pattern Generator

In many neural networks, patterns controlling rhythmic behaviors are maintained across a wide range of periods. In the crustacean pyloric central pattern generator (CPG), a constant bursting pattern is preserved over a three-to-fivefold range of periods. We idescribe how neuromodulation could adjust neuronal properties to preserve phase relations as the period changes. We developed a biophysical model implementing a reduced pyloric network motif, which has a bursting neuron and two follower neurons interconnected through inhibitory synaptic coupling. We described cellular mechanisms supporting phase maintenance and investigated possible coordination between these mechanisms in four dynamically distinct ensembles of a pyloric CPG producing a triphasic pattern. The coordinated variation of the voltages of half-activation for potassium (VK2) and hyperpolarization-activated (Vh) currents provides a family of three mechanisms for control of burst duration, interburst interval, and latency to spiking. The mechanisms are determined by the Cornerstone bifurcation, one of the Shilnikov blue sky catastrophe scenarios. In Mechanism 1, in a bursting neuron, the burst duration increases as VK2 nears a blue-sky catastrophe bifurcation, while the interburst interval grows as Vh approaches a saddle-node on an invariant circle bifurcation. In Mechanism 2, a silent neuron responds with a single burst to short input; the burst duration grows as VK2 approaches a saddle-node bifurcation for periodic orbits. In Mechanism 3, a spiking neuron responds with a pause to short input; the pause duration grows as Vh nears a saddle-node bifurcation for stationary states. In all three mechanisms, the measured quantities grow without bound as the bifurcation parameter nears its critical value, consistent with an inverse-square-root law.

q-bio.NC↗

arXiv2025

How Scientists Use Large Language Models to Program

Scientists across disciplines write code for critical activities like data collection and generation, statistical modeling, and visualization. As large language models that can generate code have become widely available, scientists may increasingly use these models during research software development. We investigate the characteristics of scientists who are early-adopters of code generating models and conduct interviews with scientists at a public, research-focused university. Through interviews and reviews of user interaction logs, we see that scientists often use code generating models as an information retrieval tool for navigating unfamiliar programming languages and libraries. We present findings about their verification strategies and discuss potential vulnerabilities that may emerge from code generation practices unknowingly influencing the parameters of scientific analyses.

cs.SE↗