Raul Castro Fernandez

Boxin Zhao, Boxiang Lyu, Raul Castro Fernandez et al.

High-quality machine learning models are dependent on access to high-quality training data. When the data are not already available, it is tedious and costly to obtain them. Data markets help with identifying valuable training data: model consumers pay to train a model, the market uses that budget to identify data and train the model (the budget allocation problem), and finally the market compensates data providers according to their data contribution (revenue allocation problem). For example, a bank could pay the data market to access data from other financial institutions to train a fraud detection model. Compensating data contributors requires understanding data's contribution to the model; recent efforts to solve this revenue allocation problem based on the Shapley value are inefficient to lead to practical data markets. In this paper, we introduce a new algorithm to solve budget allocation and revenue allocation problems simultaneously in linear time. The new algorithm employs an adaptive sampling process that selects data from those providers who are contributing the most to the model. Better data means that the algorithm accesses those providers more often, and more frequent accesses corresponds to higher compensation. Furthermore, the algorithm can be deployed in both centralized and federated scenarios, boosting its applicability. We provide theoretical guarantees for the algorithm that show the budget is used efficiently and the properties of revenue allocation are similar to Shapley's. Finally, we conduct an empirical evaluation to show the performance of the algorithm in practical scenarios and when compared to other baselines. Overall, we believe that the new algorithm paves the way for the implementation of practical data markets.

Sainyam Galhotra, Yue Gong, Raul Castro Fernandez

Data is a central component of machine learning and causal inference tasks. The availability of large amounts of data from sources such as open data repositories, data lakes and data marketplaces creates an opportunity to augment data and boost those tasks' performance. However, augmentation techniques rely on a user manually discovering and shortlisting useful candidate augmentations. Existing solutions do not leverage the synergy between discovery and augmentation, thus under exploiting data. In this paper, we introduce METAM, a novel goal-oriented framework that queries the downstream task with a candidate dataset, forming a feedback loop that automatically steers the discovery and augmentation process. To select candidates efficiently, METAM leverages properties of the: i) data, ii) utility function, and iii) solution set size. We show METAM's theoretical guarantees and demonstrate those empirically on a broad set of tasks. All in all, we demonstrate the promise of goal-oriented data discovery to modern data science applications.

Minbiao Han, Jonathan Light, Steven Xia et al.

Organizations often lack sufficient data to effectively train machine learning (ML) models, while others possess valuable data that remains underutilized. Data markets promise to unlock substantial value by matching data suppliers with demand from ML consumers. However, market design involves addressing intricate challenges, including data pricing, fairness, robustness, and strategic behavior. In this paper, we propose a pragmatic data-augmented AutoML market that seamlessly integrates with existing cloud-based AutoML platforms such as Google's Vertex AI and Amazon's SageMaker. Unlike standard AutoML solutions, our design automatically augments buyer-submitted training data with valuable external datasets, pricing the resulting models based on their measurable performance improvements rather than computational costs as the status quo. Our key innovation is a pricing mechanism grounded in the instrumental value - the marginal model quality improvement - of externally sourced data. This approach bypasses direct dataset pricing complexities, mitigates strategic buyer behavior, and accommodates diverse buyer valuations through menu-based options. By integrating automated data and model discovery, our solution not only enhances ML outcomes but also establishes an economically sustainable framework for monetizing external data.

Zhiru Zhu, Raul Castro Fernandez

Differential privacy (DP) enables private data analysis. In a typical DP deployment, controllers manage individuals' sensitive data and are responsible for answering analysts' queries while protecting individuals' privacy. They do so by choosing the privacy parameter $ε$, which controls the degree of privacy for all individuals in all possible datasets. However, it is challenging for controllers to choose $ε$ because of the difficulty of interpreting the privacy implications of such a choice on the within-dataset individuals. To address this challenge, we first derive a relative disclosure risk indicator (RDR) that indicates the impact of choosing $ε$ on the within-dataset individuals' disclosure risk. We then design an algorithm to find $ε$ based on controllers' privacy preferences expressed as a function of the within-dataset individuals' RDRs, and an alternative algorithm that finds and releases $ε$ while satisfying DP. Lastly, we propose a solution that bounds the total privacy leakage when using the algorithm to answer multiple queries without requiring controllers to set the total privacy budget. We evaluate our contributions through an IRB-approved user study that shows the RDR is useful for helping controllers choose $ε$, and experimental evaluations showing our algorithms are efficient and scalable.

Muhammad Imam Luthfi Balaka, John Hillesland, Kemal Badur et al.

When faced with data problems, many data workers cannot articulate their information need precisely enough for software to help. Although LLMs interpret natural-language requests, they behave brittly when intent is under-specified, e.g., hallucinating fields, assuming join paths, or producing ungrounded answers. We present Pneuma-Seeker, a system built around a central idea: relational reification. Pneuma-Seeker represents a user's evolving information need as a relational schema: a concrete, analysis-ready data model shared between user and system. Rather than answering prompts directly, Pneuma-Seeker iteratively refines this schema, then discovers and prepares relevant sources to construct a relation and executable program that compute the answer. Pneuma-Seeker employs an LLM-powered agentic architecture with conductor-style planning and macro- and micro-level context management to operate effectively over heterogeneous relational corpora. We evaluate Pneuma-Seeker across multiple domains against state-of-the-art academic and industrial baselines, demonstrating higher answer accuracy. Deployment in a real organization highlights trust and inspectability as essential requirements for LLM-mediated data systems.

Muhammad Imam Luthfi Balaka, Raul Castro Fernandez

Data analysts working with relational data often start with vague or underspecified questions and refine them iteratively as they explore the data. To support this iterative process, we demonstrate Pneuma-Seeker, a system that reifies a user's information need as explicit, inspectable relational specifications, enabling iterative refinement of the information need, targeted data discovery, and provenance-aware execution. Through two real-world procurement use cases, we show how Pneuma-Seeker leverages LLMs as transparent, interactive analytical collaborators rather than opaque answer engines.

Jiayao Zhang, Yuran Bi, Mengye Cheng et al.

Data is the new oil of the 21st century. The growing trend of trading data for greater welfare has led to the emergence of data markets. A data market is any mechanism whereby the exchange of data products including datasets and data derivatives takes place as a result of data buyers and data sellers being in contact with one another, either directly or through mediating agents. It serves as a coordinating mechanism by which several functions, including the pricing and the distribution of data as the most important ones, interact to make the value of data fully exploited and enhanced. In this article, we present a comprehensive survey of this important and emerging direction from the aspects of data search, data productization, data transaction, data pricing, revenue allocation as well as privacy, security, and trust issues. We also investigate the government policies and industry status of data markets across different countries and different domains. Finally, we identify the unresolved challenges and discuss possible future directions for the development of data markets.

Yue Gong, Raul Castro Fernandez

As hypothesis generation becomes increasingly automated, a new bottleneck has emerged: hypothesis assessment. Modern systems can surface thousands of statistical relationships-correlations, trends, causal links-but offer little guidance on which ones are novel, non-trivial, or worthy of expert attention. In this work, we study the complementary problem to hypothesis generation: automatic hypothesis assessment. Specifically, we ask: given a large set of statistical relationships, can we automatically assess which ones are novel and worth further exploration? We focus on correlations as they are a common entry point in exploratory data analysis that often serve as the basis for forming deeper scientific or causal hypotheses. To support automatic assessment, we propose to leverage the vast knowledge encoded in LLMs' weights to derive a prior distribution over the correlation value of a variable pair. If an LLM's prior expects the correlation value observed, then such correlation is not surprising, and vice versa. We propose the Logit-based Calibrated Prior, an LLM-elicited correlation prior that transforms the model's raw output logits into a calibrated, continuous predictive distribution over correlation values. We evaluate the prior on a benchmark of 2,096 real-world variable pairs and it achieves a sign accuracy of 78.8%, a mean absolute error of 0.26, and 95% credible interval coverage of 89.2% in predicting Pearson correlation coefficient. It also outperforms a fine-tuned RoBERTa classifier in binary correlation prediction and achieves higher precision@K in hypothesis ranking. We further show that the prior generalizes to correlations not seen during LLM pretraining, reflecting context-sensitive reasoning rather than memorization.

Manuel Cebrian, Tomomi Kito, Raul Castro Fernandez

Large language models are proliferating, and so are the benchmarks that serve as their common yardsticks. We ask how the agglomeration patterns of these two layers compare: do they evolve in tandem or diverge? Drawing on two curated proxies for the ecosystem, the Stanford Foundation-Model Ecosystem Graph and the Evidently AI benchmark registry, we find complementary but contrasting dynamics. Model creation has broadened across countries and organizations and diversified in modality, licensing, and access. Benchmark influence, by contrast, displays centralizing patterns: in the inferred benchmark-author-institution network, the top 15% of nodes account for over 80% of high-betweenness paths, three countries produce 83% of benchmark outputs, and the global Gini for inferred benchmark authority reaches 0.89. An agent-based simulation highlights three mechanisms: higher entry of new benchmarks reduces concentration; rapid inflows can temporarily complicate coordination in evaluation; and stronger penalties against over-fitting have limited effect. Taken together, these results suggest that concentrated benchmark influence functions as coordination infrastructure that supports standardization, comparability, and reproducibility amid rising heterogeneity in model production, while also introducing trade-offs such as path dependence, selective visibility, and diminishing discriminative power as leaderboards saturate.

Aldan Creo, Raul Castro Fernandez, Manuel Cebrian

As large language models (LLMs) become increasingly deployed, understanding the complexity and evolution of jailbreaking strategies is critical for AI safety. We present a mass-scale empirical analysis of jailbreak complexity across over 2 million real-world conversations from diverse platforms, including dedicated jailbreaking communities and general-purpose chatbots. Using a range of complexity metrics spanning probabilistic measures, lexical diversity, compression ratios, and cognitive load indicators, we find that jailbreak attempts do not exhibit significantly higher complexity than normal conversations. This pattern holds consistently across specialized jailbreaking communities and general user populations, suggesting practical bounds on attack sophistication. Temporal analysis reveals that while user attack toxicity and complexity remains stable over time, assistant response toxicity has decreased, indicating improving safety mechanisms. The absence of power-law scaling in complexity distributions further points to natural limits on jailbreak development. Our findings challenge the prevailing narrative of an escalating arms race between attackers and defenders, instead suggesting that LLM safety evolution is bounded by human ingenuity constraints while defensive measures continue advancing. Our results highlight critical information hazards in academic jailbreak disclosure, as sophisticated attacks exceeding current complexity baselines could disrupt the observed equilibrium and enable widespread harm before defensive adaptation.

Nadiia Chepurko, Ryan Marcus, Emanuel Zgraggen et al.

Automatic machine learning (\AML) is a family of techniques to automate the process of training predictive models, aiming to both improve performance and make machine learning more accessible. While many recent works have focused on aspects of the machine learning pipeline like model selection, hyperparameter tuning, and feature selection, relatively few works have focused on automatic data augmentation. Automatic data augmentation involves finding new features relevant to the user's predictive task with minimal ``human-in-the-loop'' involvement. We present \system, an end-to-end system that takes as input a dataset and a data repository, and outputs an augmented data set such that training a predictive model on this augmented dataset results in improved performance. Our system has two distinct components: (1) a framework to search and join data with the input data, based on various attributes of the input, and (2) an efficient feature selection algorithm that prunes out noisy or irrelevant features from the resulting join. We perform an extensive empirical evaluation of different system components and benchmark our feature selection algorithm on real-world datasets.

Guillaume Leclerc, Manasi Vartak, Raul Castro Fernandez et al.

As neural networks become widely deployed in different applications and on different hardware, it has become increasingly important to optimize inference time and model size along with model accuracy. Most current techniques optimize model size, model accuracy and inference time in different stages, resulting in suboptimal results and computational inefficiency. In this work, we propose a new technique called Smallify that optimizes all three of these metrics at the same time. Specifically we present a new method to simultaneously optimize network size and model performance by neuron-level pruning during training. Neuron-level pruning not only produces much smaller networks but also produces dense weight matrices that are amenable to efficient inference. By applying our technique to convolutional as well as fully connected models, we show that Smallify can reduce network size by 35X with a 6X improvement in inference time with similar accuracy as models found by traditional training techniques.