arXiv subjects

Renée J. Miller

Publications and source records attributed to Renée J. Miller.

At least 19 recordsLinked to original sources

arXiv2026

ModelLakeFishing: Efficient Retrieval over Million-Scale Model Lakes

Open model lakes may contain millions of reusable models, making it costly to identify suitable models for a new dataset. We present ModelLakeFishing, a model-retrieval framework for queries specifying a target dataset, prediction task, and evaluation metric. It consolidates metadata and historical evaluations into a model-dataset-task evidence graph, learns model and query embeddings with a relation-aware graph encoder, and indexes model embeddings using Hierarchical Navigable Small World (HNSW) search. At query time, HNSW retrieves 1,000 candidates without scoring every model, after which a training-side task prior reranks candidates for the requested metric and returns the top 10. We evaluate on a lake of 3,016,439 models and 247,803 observed model-dataset performance pairs using three root-aware splits that hold test performance edges out of representation learning and retrieval. ModelLakeFishing achieves a mean eligible-query gold@10 of 0.2968, recovering the observed-best model in the top 10 for 29.68% of eligible queries and retaining 93.47% of the gold@10 of an exhaustive baseline using the same scoring and reranking procedure. Given precomputed query embeddings, retrieval and reranking take 0.747 ms median and 1.102 ms at the 95th percentile. These results demonstrate efficient retrieval over million-model lakes from sparse relational evidence.

cs.IR↗

arXiv2026

Data Bias Mitigation under Coverage Constraints & The Price of Fairness

Machine learning models have been shown to exhibit discriminatory outcomes or degraded performance for individuals at the intersection of multiple sensitive attributes, such as race and gender. This stems in part from two interrelated challenges: the lack of principled measures for quantifying bias (potentially intersectional), and insufficient representation of intersectional subgroups in training data. We extend a recent bias mitigation framework to incorporate coverage constraints that enforce sufficient representation across groups, including intersectional subgroups. Since achieving exactly zero bias for all groups may not be data efficient (meaning it may require large amounts of data), our solution trades small approximation errors in bias for greater data efficiency while satisfying coverage constraints. We also formulate bias mitigation as an integer linear program that optimizes over all mitigation strategies, and characterize the price of fairness, the minimum data modification cost, as a function of fairness tolerance. This is essential both for legal compliance, where regulations may mandate specific fairness thresholds, and for data governance, enabling practitioners to make informed trade-offs between bias reduction and data modification (particularly, data purchasing) costs. We evaluate our techniques on publicly available datasets, demonstrating that bias mitigation via our framework preserves predictive accuracy across multiple classifiers, and that coverage constraints, while motivated by statistical considerations, are essential for preserving downstream ML performance.

cs.LG↗

arXiv2026

Diversed Model Discovery via Structured Table Discovery

Model cards describe model behavior through a mixture of textual descriptions and structured artifacts, including performance, configuration, and dataset tables. Existing model search systems rely predominantly on semantic similarity over text, which can produce homogeneous result sets and limit exploration of alternatives. We argue that model search is inherently comparative: users want models that are task-aligned yet differentiated in measurable ways. We hypothesize that this balance requires retrieval over condensed, high-quality evidence rather than verbose descriptions, and much of that evidence is concentrated in structured tables. We present StructuredSemanticSearch, a table-driven model search framework built on the ModelTables benchmark. Given a query, StructuredSemanticSearch combines a semantic baseline for task alignment with a structure-aware pipeline that discovers query-related model-card tables using table discovery operators such as unionability, joinability, and keyword search. Retrieved tables are mapped back to model cards under a controlled top-k budget, enabling fair comparison between text-based and table-based retrieval. Beyond retrieval, StructuredSemanticSearch adapts table integration to the model-table domain through orientation-aware integration, producing compact integrated views of tables from partially overlapping and sometimes transposed evidence tables. For evaluation, we introduce a nugget-based, auditable protocol that extracts compact evidence items from model cards, matches queries to condition- or intent-specific nuggets, and measures evidence coverage and diversity over retrieved model-card candidate sets. This protocol also provides a scalable path toward approximate, evidence-based labeling in dynamic model lakes. Experiments on 597 model-recommendation queries show improved nugget coverage for the structure-aware pipeline than semantic baseline

cs.IR↗

arXiv2026

Novel Table Search [Technical Report]

Avoiding redundancy in query results has been extensively studied in relational databases and information retrieval, yet its implications for data lakes remain largely unexplored. We bridge this gap by investigating how to discover unionable tables that contribute new information for a given query table in large-scale data lakes. We formally define Novel Table Search (NTS) as the problem of finding tables that are novel with respect to a given query table and identify two desirable properties that any scoring function for NTS should satisfy. We introduce a concrete scoring mechanism designed to maximize syntactic novelty, prove that it satisfies the proposed properties, and show that the associated optimization problem is NP-hard. To address this challenge, we develop an efficient approximation technique based on penalization, i.e., Attribute-Based Novel Table Search (ANTs). We propose three additional NTS variants to achieve syntactic novelty and introduce two evaluation metrics for syntactic novelty. Through extensive experiments, we demonstrate that ANTs outperforms other methods in capturing syntactic novelty across evaluation metrics and various benchmarks, while also achieving the lowest execution time.

cs.DB↗

arXiv2025

ModelTables: A Corpus of Tables about Models

We present ModelTables, a benchmark of tables in Model Lakes that captures the structured semantics of performance and configuration tables often overlooked by text only retrieval. The corpus is built from Hugging Face model cards, GitHub READMEs, and referenced papers, linking each table to its surrounding model and publication context. Compared with open data lake tables, model tables are smaller yet exhibit denser inter table relationships, reflecting tightly coupled model and benchmark evolution. The current release covers over 60K models and 90K tables. To evaluate model and table relatedness, we construct a multi source ground truth using three complementary signals: (1) paper citation links, (2) explicit model card links and inheritance, and (3) shared training datasets. We present one extensive empirical use case for the benchmark which is table search. We compare canonical Data Lake search operators (unionable, joinable, keyword) and Information Retrieval baselines (dense, sparse, hybrid retrieval) on this benchmark. Union based semantic table retrieval attains 54.8 % P@1 overall (54.6 % on citation, 31.3 % on inheritance, 30.6 % on shared dataset signals); table based dense retrieval reaches 66.5 % P@1, and metadata hybrid retrieval achieves 54.1 %. This evaluation indicates clear room for developing better table search methods. By releasing ModelTables and its creation protocol, we provide the first large scale benchmark of structured data describing AI model. Our use case of table discovery in Model Lakes, provides intuition and evidence for developing more accurate semantic retrieval, structured comparison, and principled organization of structured model knowledge. Source code, data, and other artifacts have been made available at https://github.com/RJMillerLab/ModelTables.

cs.DB↗

arXiv2025

Distribution-Aware Exploration for Adaptive HNSW Search

Hierarchical Navigable Small World (HNSW) is widely adopted for approximate nearest neighbor search (ANNS) for its ability to deliver high recall with low latency on large-scale, high-dimensional embeddings. The exploration factor, commonly referred to as ef, is a key parameter in HNSW-based vector search that balances accuracy and efficiency. However, existing systems typically rely on manually and statically configured ef values that are uniformly applied across all queries. This results in a distribution-agnostic configuration that fails to account for the non-uniform and skewed nature of real-world embedding data and query workloads. As a consequence, HNSW-based systems suffer from two key practical issues: (i) the absence of recall guarantees, and (ii) inefficient ANNS performance due to over- or under-searching. In this paper, we propose Adaptive-ef (Ada-ef), a data-driven, update-friendly, query-adaptive approach that dynamically configures ef for each query at runtime to approximately meet a declarative target recall with minimal computation. The core of our approach is a theoretically grounded statistical model that captures the similarity distribution between each query and the database vectors. Based on this foundation, we design a query scoring mechanism that distinguishes between queries requiring only small ef and those that need larger ef to meet a target recall, and accordingly assigns an appropriate ef to each query. Experimental results on real-world embeddings produced by state-of-the-art Transformer models from OpenAI and Cohere show that, compared with state-of-the-art learning-based adaptive approaches, our method achieves the target recall while avoiding both over- and under-searching, reducing online query latency by up to 4x, offline computation time by 50x, and offline memory usage by 100x.

cs.DB↗

arXiv2025

Diverse Unionable Tuple Search: Novelty-Driven Discovery in Data Lakes [Technical Report]

Unionable table search techniques input a query table from a user and search for data lake tables that can contribute additional rows to the query table. The definition of unionability is generally based on similarity measures which may include similarity between columns (e.g., value overlap or semantic similarity of the values in the columns) or tables (e.g., similarity of table embeddings). Due to this and the large redundancy in many data lakes (which can contain many copies and versions of the same table), the most unionable tables may be identical or nearly identical to the query table and may contain little new information. Hence, we introduce the problem of identifying unionable tuples from a data lake that are diverse with respect to the tuples already present in a query table. We perform an extensive experimental analysis of well-known diversity algorithms applied to this novel problem and identify a gap that we address with a novel, clustering-based tuple diversity algorithm called DUST. DUST uses a novel embedding model to represent unionable tuples that outperforms other tuple representation models by at least 15 % when representing unionable tuples. Using real data lake benchmarks, we show that our diversification algorithm is more than six times faster than the most efficient diversification baseline. We also show that it is more effective in diversifying unionable tuples than existing diversification algorithms.

cs.DB↗

arXiv2025

A Principled Approach for Data Bias Mitigation

The widespread use of machine learning and data-driven algorithms for decision making has been steadily increasing over many years. \emph{Bias} in the data can adversely affect this decision-making. We present a new mitigation strategy to address data bias. Our methods are explainable and come with mathematical guarantees of correctness. They can take advantage of new work on table discovery to find new tuples that can be added to a dataset to create real datasets that are unbiased or less biased. Our framework covers data with non-binary labels and with multiple sensitive attributes. Hence, we are able to measure and mitigate bias that does not appear over a single attribute (or feature), but only intersectionally, when considering a combination of attributes. We evaluate our techniques on publicly available datasets and provide a theoretical analysis of our results, highlighting novel insights into data bias.

cs.LG↗

arXiv2025

Model Lakes

Given a set of deep learning models, it can be hard to find models appropriate to a task, understand the models, and characterize how models are different one from another. Currently, practitioners rely on manually-written documentation to understand and choose models. However, not all models have complete and reliable documentation. As the number of models increases, the challenges of finding, differentiating, and understanding models become increasingly crucial. Inspired from research on data lakes, we introduce the concept of model lakes. We formalize key model lake tasks, including model attribution, versioning, search, and benchmarking, and discuss fundamental research challenges in the management of large models. We also explore what data management techniques can be brought to bear on the study of large model management.

cs.DB↗

arXiv2025

Fuzzy Integration of Data Lake Tables

Data integration is an important step in any data science pipeline where the objective is to unify the information available in different datasets for comprehensive analysis. Full Disjunction, which is an associative extension of the outer join operator, has been shown to be an effective operator for integrating datasets. It fully preserves and combines the available information. Existing Full Disjunction algorithms only consider the equi-join scenario where only tuples having the same value on joining columns are integrated. This, however, does not realistically represent an open data scenario, where datasets come from diverse sources with inconsistent values (e.g., synonyms, abbreviations, etc.) and with limited metadata. So, joining just on equal values severely limits the ability of Full Disjunction to fully combine datasets. Thus, in this work, we propose an extension of Full Disjunction to also account for "fuzzy" matches among tuples. We present a novel data-driven approach to enable the joining of approximate or fuzzy matches within Full Disjunction. Experimentally, we show that fuzzy Full Disjunction does not add significant time overhead over a state-of-the-art Full Disjunction implementation and also that it enhances the integration effectiveness.

cs.DB↗

arXiv2024

Blend: A Unified Data Discovery System

Most research on data discovery has so far focused on improving individual discovery operators such as join, correlation, or union discovery. However, in practice, a combination of these techniques and their corresponding indexes may be necessary to support arbitrary discovery tasks. We propose BLEND, a comprehensive data discovery system that supports existing operators and enables their flexible pipelining. BLEND is based on a set of lower-level operators that serve as fundamental building blocks for more complex and sophisticated user tasks. To reduce the execution runtime of discovery pipelines, we propose a unified index structure and a rule-based optimizer that rewrites SQL statements into low-level operators when possible. We show the superior flexibility and efficiency of our system compared to ad-hoc discovery pipelines and stand-alone solutions.

cs.DB↗

arXiv2024

Gen-T: Table Reclamation in Data Lakes

We introduce the problem of Table Reclamation. Given a Source Table and a large table repository, reclamation finds a set of tables that, when integrated, reproduce the source table as closely as possible. Unlike query discovery problems like Query-by-Example or by-Target, Table Reclamation focuses on reclaiming the data in the Source Table as fully as possible using real tables that may be incomplete or inconsistent. To do this, we define a new measure of table similarity, called error-aware instance similarity, to measure how close a reclaimed table is to a Source Table, a measure grounded in instance similarity used in data exchange. Our search covers not only SELECT-PROJECT- JOIN queries, but integration queries with unions, outerjoins, and the unary operators subsumption and complementation that have been shown to be important in data integration and fusion. Using reclamation, a data scientist can understand if any tables in a repository can be used to exactly reclaim a tuple in the Source. If not, one can understand if this is due to differences in values or to incompleteness in the data. Our solution, Gen-T, performs table discovery to retrieve a set of candidate tables from the table repository, filters these down to a set of originating tables, then integrates these tables to reclaim the Source as closely as possible. We show that our solution, while approximate, is accurate, efficient and scalable in the size of the table repository with experiments on real data lakes containing up to 15K tables, where the average number of tuples varies from small (web tables) to extremely large (open data tables) up to 1M tuples.

cs.DB↗

arXiv2023

Generative Benchmark Creation for Table Union Search

Data management has traditionally relied on synthetic data generators to generate structured benchmarks, like the TPC suite, where we can control important parameters like data size and its distribution precisely. These benchmarks were central to the success and adoption of database management systems. But more and more, data management problems are of a semantic nature. An important example is finding tables that can be unioned. While any two tables with the same cardinality can be unioned, table union search is the problem of finding tables whose union is semantically coherent. Semantic problems cannot be benchmarked using synthetic data. Our current methods for creating benchmarks involve the manual curation and labeling of real data. These methods are not robust or scalable and perhaps more importantly, it is not clear how robust the created benchmarks are. We propose to use generative AI models to create structured data benchmarks for table union search. We present a novel method for using generative models to create tables with specified properties. Using this method, we create a new benchmark containing pairs of tables that are both unionable and non-unionable but related. We thoroughly evaluate recent existing table union search methods over existing benchmarks and our new benchmark. We also present and evaluate a new table search methods based on recent large language models over all benchmarks. We show that the new benchmark is more challenging for all methods than hand-curated benchmarks, specifically, the top-performing method achieves a Mean Average Precision of around 60%, over 30% less than its performance on existing manually created benchmarks. We examine why this is the case and show that the new benchmark permits more detailed analysis of methods, including a study of both false positives and false negatives that were not possible with existing benchmarks.

cs.DB↗

arXiv2023

DIALITE: Discover, Align and Integrate Open Data Tables

We demonstrate a novel table discovery pipeline called DIALITE that allows users to discover, integrate and analyze open data tables. DIALITE has three main stages. First, it allows users to discover tables from open data platforms using state-of-the-art table discovery techniques. Second, DIALITE integrates the discovered tables to produce an integrated table. Finally, it allows users to analyze the integration result by applying different downstreaming tasks over it. Our pipeline is flexible such that the user can easily add and compare additional discovery and integration algorithms.

cs.DB↗

arXiv2023

Explaining Dataset Changes for Semantic Data Versioning with Explain-Da-V (Technical Report)

In multi-user environments in which data science and analysis is collaborative, multiple versions of the same datasets are generated. While managing and storing data versions has received some attention in the research literature, the semantic nature of such changes has remained under-explored. In this work, we introduce \texttt{Explain-Da-V}, a framework aiming to explain changes between two given dataset versions. \texttt{Explain-Da-V} generates \emph{explanations} that use \emph{data transformations} to explain changes. We further introduce a set of measures that evaluate the validity, generalizability, and explainability of these explanations. We empirically show, using an adapted existing benchmark and a newly created benchmark, that \texttt{Explain-Da-V} generates better explanations than existing data transformation synthesis methods.

cs.DB↗

arXiv2022

SANTOS: Relationship-based Semantic Table Union Search

Existing techniques for unionable table search define unionability using metadata (tables must have the same or similar schemas) or column-based metrics (for example, the values in a table should be drawn from the same domain). In this work, we introduce the use of semantic relationships between pairs of columns in a table to improve the accuracy of union search. Consequently, we introduce a new notion of unionability that considers relationships between columns, together with the semantics of columns, in a principled way. To do so, we present two new methods to discover semantic relationship between pairs of columns. The first uses an existing knowledge base (KB), the second (which we call a "synthesized KB") uses knowledge from the data lake itself. We adopt an existing Table Union Search benchmark and present new (open) benchmarks that represent small and large real data lakes. We show that our new unionability search algorithm, called SANTOS, outperforms a state-of-the-art union search that uses a wide variety of column-based semantics, including word embeddings and regular expressions. We show empirically that our synthesized KB improves the accuracy of union search by representing relationship semantics that may not be contained in an available KB. This result hints at a promising future of creating a synthesized KBs from data lakes with limited KB coverage and using them for union search.

cs.DB↗

arXiv2021

DomainNet: Homograph Detection for Data Lake Disambiguation

Modern data lakes are deeply heterogeneous in the vocabulary that is used to describe data. We study a problem of disambiguation in data lakes: how can we determine if a data value occurring more than once in the lake has different meanings and is therefore a homograph? While word and entity disambiguation have been well studied in computational linguistics, data management and data science, we show that data lakes provide a new opportunity for disambiguation of data values since they represent a massive network of interconnected values. We investigate to what extent this network can be used to disambiguate values. DomainNet uses network-centrality measures on a bipartite graph whose nodes represent values and attributes to determine, without supervision, if a value is a homograph. A thorough experimental evaluation demonstrates that state-of-the-art techniques in domain discovery cannot be re-purposed to compete with our method. Specifically, using a domain discovery method to identify homographs has a precision and a recall of 38% versus 69% with our method on a synthetic benchmark. By applying a network-centrality measure to our graph representation, DomainNet achieves a good separation between homographs and data values with a unique meaning. On a real data lake our top-200 precision is 89%.

cs.DB↗

arXiv2020

Knowledge Translation: Extended Technical Report

We introduce Kensho, a tool for generating mapping rules between two Knowledge Bases (KBs). To create the mapping rules, Kensho starts with a set of correspondences and enriches them with additional semantic information automatically identified from the structure and constraints of the KBs. Our approach works in two phases. In the first phase, semantic associations between resources of each KB are captured. In the second phase, mapping rules are generated by interpreting the correspondences in a way that respects the discovered semantic associations among elements of each KB. Kensho's mapping rules are expressed using SPARQL queries and can be used directly to exchange knowledge from source to target. Kensho is able to automatically rank the generated mapping rules using a set of heuristics. We present an experimental evaluation of Kensho and assess our mapping generation and ranking strategies using more than 50 synthesized and real world settings, chosen to showcase some of the most important applications of knowledge translation. In addition, we use three existing benchmarks to demonstrate Kensho's ability to deal with different mapping scenarios.

cs.DB↗