arXiv subjects

Hamidah Oderinwale

Publications and source records attributed to Hamidah Oderinwale.

8 recordsLinked to original sources

arXiv2026

Structuring license permissiveness from pairwise comparisons

Licenses are legal instruments that inventors rely upon to protect the technologies they build and regulate how they are used---however, the nature of their authorship and selection implies that how they are interpreted, chosen, and enforced is largely unstructured. In practice, this makes it difficult to compare licenses at scale---when is one license considered more permissive than the other, and when are their terms incomparable to each other? Currently, there is a growing list of licenses that are introduced and used, yet no systematic way to study their relationships. This matters for platforms such as Hugging Face, GitHub, and the Python Package Index, where developers publish or build upon technologies that each have their own licenses. Using large language models (LLMs), we introduce methods for comparing licenses at scale: first, in a pairwise fashion to construct and validate a partial ordering based on permissiveness; and by drawing on existing taxonomies of software licenses. Then, we try to recover the structure with the Bradley-Terry model to see if permissiveness can be judged more cheaply and observe a loss of $\sim$20\%---and classify this loss to feature coverage. The former coupled with model rationale allows us to trace restrictiveness, and the latter allows us to understand license selection as a combination of shared provisions.

cs.SE↗

arXiv2026

The Economics of AI Training Data: A Research Agenda

Despite data's central role in AI production, it remains the least understood input. As AI labs exhaust public data and turn to proprietary sources, with deals reaching hundreds of millions of dollars, research across computer science, economics, law, and policy has fragmented. We establish data economics as a coherent field through three contributions. First, we characterize data's distinctive properties -- nonrivalry, context dependence, and emergent rivalry through contamination -- and trace historical precedents for market formation in commodities such as oil and grain. Second, we present systematic documentation of AI training data deals from 2020 to 2025, revealing persistent market fragmentation, five distinct pricing mechanisms (from per-unit licensing to commissioning), and that most deals exclude original creators from compensation. Third, we propose a formal hierarchy of exchangeable data units (token, record, dataset, corpus, stream) and argue for data's explicit representation in production functions. Building on these foundations, we outline four open research problems foundational to data economics: measuring context-dependent value, balancing governance with privacy, estimating data's contribution to production, and designing mechanisms for heterogeneous, compositional goods.

cs.CY↗

arXiv2026

Agent trajectories as programs: fingerprinting and programming coding-agent behavior

Benchmark scores tell you what an agent got right; they do not tell you how it got there. In this work, we introduce methods for comparing agents procedurally in different contexts, where the model, tasks, and approaches vary. We compare ten agents and find that they are identifiable by their behavioral habits, which we define as fingerprints: a probe over these procedural signatures attributes an unseen trajectory to the correct agent at 85.7% accuracy, controlling for leakage across tasks. We develop procedural representations for agent problem-solving procedures with an emergent vocabulary induction technique that is meant to be maximally compressive to avoid surface-level variation while being expressive enough to unveil the quirks of the models' patterns. We apply our framework to the software engineering evaluation dataset SWE-Bench to study the structural distinctness of agent trajectories and find that behavior is most similar between models from similar release periods and those that are distilled from one another (e.g., a distilled student model and its teacher have a Jensen-Shannon divergence of 0.25, about half the distance between other model pairs). As more models saturate evaluations, we believe that it will be important to probe model behavior along more holistic dimensions than success rates alone. We introduce ProcGrep, a library for auditing and evaluating agents for how they approach tasks at a procedural level given their traces in a top-down fashion. We believe this work has a range of applications to help developers work with and program coding agents, such as task-aware model routing, agent monitoring, and finer-grained cost analysis.

cs.SE↗

arXiv2025

Anatomy of a Machine Learning Ecosystem: 2 Million Models on Hugging Face

Many have observed that the development and deployment of generative machine learning (ML) and artificial intelligence (AI) models follow a distinctive pattern in which pre-trained models are adapted and fine-tuned for specific downstream tasks. However, there is limited empirical work that examines the structure of these interactions. This paper analyzes 1.86 million models on Hugging Face, a leading peer production platform for model development. Our study of model family trees -- networks that connect fine-tuned models to their base or parent -- reveals sprawling fine-tuning lineages that vary widely in size and structure. Using an evolutionary biology lens to study ML models, we use model metadata and model cards to measure the genetic similarity and mutation of traits over model families. We find that models tend to exhibit a family resemblance, meaning their genetic markers and traits exhibit more overlap when they belong to the same model family. However, these similarities depart in certain ways from standard models of asexual reproduction, because mutations are fast and directed, such that two `sibling' models tend to exhibit more similarity than parent/child pairs. Further analysis of the directional drifts of these mutations reveals qualitative insights about the open machine learning ecosystem: Licenses counter-intuitively drift from restrictive, commercial licenses towards permissive or copyleft licenses, often in violation of upstream license's terms; models evolve from multi-lingual compatibility towards english-only compatibility; and model cards reduce in length and standardize by turning, more often, to templates and automatically generated text. Overall, this work takes a step toward an empirically grounded understanding of model fine-tuning and suggests that ecological models and methods can yield novel scientific insights.

cs.SI↗

arXiv2025

Procedural Knowledge Libraries: Towards Executable (Research) Memory

Procedural Knowledge Libraries (PKLs) are frameworks for capturing the full arc of scientific inquiry, not just its outcomes. Whereas traditional libraries store static end products, PKLs preserve the process that leads to those results, including hypotheses, failures, decisions, and iterations. By addressing the loss of tacit knowledge -- typically buried in notebooks, emails, or memory -- PKLs lay a foundation for reproducible, collaborative, and adaptive research. PKLs provide executable, version-controlled records that contextualize each step of a research process. For example, a researcher using Jupyter notebooks could share not just final outputs, but also the reasoning, discarded approaches, and intermediate analyses that informed them. This work proposes a framework for implementing PKLs within the Jupyter ecosystem, supported by a lens-based transformation model and procedural storage schema.

cs.DL↗

arXiv2024

Croissant: A Metadata Format for ML-Ready Datasets

Data is a critical resource for machine learning (ML), yet working with data remains a key friction point. This paper introduces Croissant, a metadata format for datasets that creates a shared representation across ML tools, frameworks, and platforms. Croissant makes datasets more discoverable, portable, and interoperable, thereby addressing significant challenges in ML data management. Croissant is already supported by several popular dataset repositories, spanning hundreds of thousands of datasets, enabling easy loading into the most commonly-used ML frameworks, regardless of where the data is stored. Our initial evaluation by human raters shows that Croissant metadata is readable, understandable, complete, yet concise.

cs.LG↗

arXiv2024

Future and AI-Ready Data Strategies: Response to DOC RFI on AI and Open Government Data Assets

The following is a response to the US Department of Commerce's Request for Information (RFI) regarding AI and Open Government Data Assets. First, we commend the Department for its initiative in seeking public insights on the organization and sharing of data. To facilitate scientific discovery and advance AI development, it is crucial for all data producers, including the Department of Commerce and other governmental entities, to prioritize the quality of their data corpora. Ensuring data is accessible, scalable, and secure is essential for harnessing its full potential. In our response, we outline best practices and key considerations for AI and the Department of Commerce's Open Government Data Assets.

cs.CY↗

arXiv2024

Consent in Crisis: The Rapid Decline of the AI Data Commons

General-purpose artificial intelligence (AI) systems are built on massive swathes of public web data, assembled into corpora such as C4, RefinedWeb, and Dolma. To our knowledge, we conduct the first, large-scale, longitudinal audit of the consent protocols for the web domains underlying AI training corpora. Our audit of 14,000 web domains provides an expansive view of crawlable web data and how codified data use preferences are changing over time. We observe a proliferation of AI-specific clauses to limit use, acute differences in restrictions on AI developers, as well as general inconsistencies between websites' expressed intentions in their Terms of Service and their robots.txt. We diagnose these as symptoms of ineffective web protocols, not designed to cope with the widespread re-purposing of the internet for AI. Our longitudinal analyses show that in a single year (2023-2024) there has been a rapid crescendo of data restrictions from web sources, rendering ~5%+ of all tokens in C4, or 28%+ of the most actively maintained, critical sources in C4, fully restricted from use. For Terms of Service crawling restrictions, a full 45% of C4 is now restricted. If respected or enforced, these restrictions are rapidly biasing the diversity, freshness, and scaling laws for general-purpose AI systems. We hope to illustrate the emerging crises in data consent, for both developers and creators. The foreclosure of much of the open web will impact not only commercial AI, but also non-commercial AI and academic research.

cs.CL↗