arXiv subjects

Zhengjie Miao

Publications and source records attributed to Zhengjie Miao.

12 recordsLinked to original sources

arXiv2026

Building SQL Equivalence Proofs from Query Rewrite Rules

Query rewriting and semantic evaluation of text-to-SQL systems both depend on deciding whether two SQL queries return the same result on every database satisfying the schema constraints. Decades of equivalence-preserving optimizer rules provide a natural basis for constructive proofs: two queries are equivalent when admitted rewrites transform them to a common logical expression. Building on this observation, \sys{} organizes query-rewriting knowledge as a standalone proof-producing system. It separates rule validation from proof discovery, explicitly aligns alternative query representations, and returns the resulting rewrite derivation. Across three corpus classes, EquiTrace proves 648/1,106 expected-equivalent text-to-SQL pairs (58.6%), 229/266 curated pairs (86.1%), and 97/240 LLM-generated rewrites whose results agree under benchmark execution (40.4%). On the query rewrite corpus, EquiTrace proves 97 pairs, compared with 35 for QED and 38 for SQLSolver. An analysis of 1,242 recorded proofs shows that coverage depends on a small number of transformations, while many frequently used rules are individually substitutable. The resulting derivations make each proof replayable and expose the specific transformations on which it depends.

cs.DB↗

arXiv2026

A Function-Space Approach to the Statistical Mechanics of Learning Dynamics

In the kernel regime, neural-network learning inherits its preferences from a frozen spectrum. During feature learning, this spectrum evolves, yet networks retain systematic biases toward simple, smooth directions. We develop a function-space statistical framework explaining the origin of these preferences, treating functions and their learning operators as macroscopic variables, with parameterization entering through the multiplicity of parameter configurations realizing each function. For mean-squared loss, error relaxes exactly under the evolving learning operator $M=JJ^\ast$. Training stochasticity induces a Gaussian weight over function-space states, while parameter multiplicity contributes an entropic operator $B$, defined by the curvature of its log multiplicity. A local Laplace expansion yields the fluctuation free energy $Φ_{\mathrm{fluc}}(M;B)=\frac{σ_ξ^2}{2}\log\det(M^{-1}+B)+\mathrm{const}$, analogous to an Occam factor. Under mild statistical conditions, this free energy is rotationally stationary exactly when $[M,B]=0$, is minimized by pairing large eigenvalues of $M$ with small eigenvalues of $B$, and generates a local restoring force against mismatch. Learning is therefore biased toward faster relaxation along entropically cheaper directions. This preference strengthens with training noise and vanishes in the deterministic limit, beyond gradient-flow accounts of operator alignment. For ReLU networks, we relate entropic curvature to the minimal rearrangement of activation boundaries required for a functional change and bound this structural cost by directional smoothness. Consequently, smooth directions are preferentially learned faster, in a data-adaptive manner, even as the learning operator evolves.

cs.AI↗

arXiv2026

Benchmarking the Full Pipeline of Materialized-View-Based Query Rewriting

Materialized views (MVs) accelerate OLAP and data-warehouse workloads by precomputing reusable subexpressions, but practical MV-based query acceleration is a multi-stage pipeline: candidate enumeration, view selection under storage budgets, and query rewriting inside the optimizer. Existing evaluations typically study only parts of this pipeline and within a single system, leaving end-to-end trade-offs and cross-system behavior unclear. In this paper, we benchmark MV-based query rewriting by jointly evaluating enumeration, selection, and rewriting with a modular evaluation framework and by using controlled ablations. We also introduce a cross-engine protocol allowing us to compare systems that expose only execution plans by contrasting native optimizer-level rewriting with portable SQL rewriting baselines when available. Across representative academic methods and modern open-source and commercial systems, we find strong interaction effects across stages and large variability in MV usage and realized savings. We identify recurring failure modes that explain performance regressions after rewriting. Our results highlight which pipeline stages most often limit performance and provide evidence to guide future MV enumeration, selection, and rewriting designs.

cs.DB↗

arXiv2026

SPA: A SQL-Plan-Aware Reinforcement Learning Framework for Query Rewriting with LLMs

SQL query rewriting is a well-established technique for improving database performance without schema or index changes, yet finding effective rewrites for modern analytical workloads remains difficult: rule-based methods are limited to predefined transformations, while LLM-based approaches often produce rewrites that are semantically valid but compile to equivalent physical plans or degrade runtime performance. We present SPA, a SQL-Plan-Aware reinforcement learning framework that trains LLMs to rewrite queries using physical execution feedback. SPA formulates rewriting as a policy optimization problem and extends GRPO with rewards spanning semantic equivalence, textual rewrite distance, physical-plan divergence, and runtime speedup. To handle reward sparsity across query difficulty, SPA introduces Probability-Gated Adaptive Reward Shaping, a query-level curriculum that unlocks higher-level rewards only once a rollout group achieves sufficient mastery of lower-level objectives, and further improves sample efficiency through on-policy self-improvement by recycling slowdown rewrites from the current policy as targeted training signals. On both IID and OOD workloads, SPA outperforms rule-based and strong LLM baselines in end-to-end runtime, substantially reduces harmful slowdown rewrites, and yields strong tail-latency gains.

cs.DB↗

arXiv2026

The Case for Text-to-SQL Friendly Logical Database Design

Logical database design has traditionally optimized database schemas, including tables, columns, keys, constraints, and views, for correctness, integrity, and human-written application queries. LLM-based Text-to-SQL changes the consumer: the schema is now often read as text by a language model, so design choices that preserve database semantics can still change SQL-generation accuracy. We argue that this creates a new design objective alongside the classical ones - LLM-friendly logical database design, the property that a schema is easy for a language model to map from natural language to correct SQL - and treat it as the optimization target of this paper. We instantiate this objective with three semantics-preserving schema transformations that re-purpose classical schema-design ideas: schema abstraction (+A: logical views that materialize recurring join paths), schema partitioning (+P: workload-aware logical partitions that prune irrelevant context), and schema renaming (+R: descriptive identifiers that improve downstream column linking and predicate construction). The three operators compose, and each preserves the underlying database semantics. When historical question-SQL pairs are available, they guide both partitioning and abstraction; in zero-shot settings, renaming applies directly, and abstraction falls back to an ad-hoc per-question variant. We evaluate the resulting schemas on BIRD-Union and Spider-Union across multiple Text-to-SQL pipelines and language model backbones, with gains of up to 4.2% in execution accuracy. The best transformation varies modestly across pipelines and models, with the full +A+P+R consistently improving; multiple operator combinations are competitive on each pipeline. These results show that LLM-friendly logical design is a practical and underexplored database-side optimization target, complementary to existing Text-to-SQL pipelines.

cs.DB↗

arXiv2025

Refining Labeling Functions with Limited Labeled Data

Programmatic weak supervision (PWS) significantly reduces human effort for labeling data by combining the outputs of user-provided labeling functions (LFs) on unlabeled datapoints. However, the quality of the generated labels depends directly on the accuracy of the LFs. In this work, we study the problem of fixing LFs based on a small set of labeled examples. Towards this goal, we develop novel techniques for repairing a set of LFs by minimally changing their results on the labeled examples such that the fixed LFs ensure that (i) there is sufficient evidence for the correct label of each labeled datapoint and (ii) the accuracy of each repaired LF is sufficiently high. We model LFs as conditional rules which enables us to refine them, i.e., to selectively change their output for some inputs. We demonstrate experimentally that our system improves the quality of LFs based on surprisingly small sets of labeled datapoints.

cs.LG↗

arXiv2025

CleanAgent: Automating Data Standardization with LLM-based Agents

Data standardization is a crucial part of the data science life cycle. While tools like Pandas offer robust functionalities, their complexity and the manual effort required for customizing code to diverse column types pose significant challenges. Although large language models (LLMs) like ChatGPT have shown promise in automating this process through natural language understanding and code generation, it still demands expert-level programming knowledge and continuous interaction for prompt refinement. To solve these challenges, our key idea is to propose a Python library with declarative, unified APIs for standardizing different column types, simplifying the LLM's code generation with concise API calls. We first propose Dataprep.Clean, a component of the Dataprep Python Library, significantly reduces the coding complexity by enabling the standardization of specific column types with a single line of code. Then, we introduce the CleanAgent framework integrating Dataprep.Clean and LLM-based agents to automate the data standardization process. With CleanAgent, data scientists only need to provide their requirements once, allowing for a hands-free process. To demonstrate the practical utility of CleanAgent, we developed a user-friendly web application, allowing users to interact with it using real-world datasets.

cs.LG↗

arXiv2025

LINEAGEX: A Column Lineage Extraction System for SQL

As enterprise data grows in size and complexity, column-level data lineage, which records the creation, transformation, and reference of each column in the warehouse, has been the key to effective data governance that assists tasks like data quality monitoring, storage refactoring, and workflow migration. Unfortunately, existing systems introduce overheads by integration with query execution or fail to achieve satisfying accuracy for column lineage. In this paper, we demonstrate LINEAGEX, a lightweight Python library that infers column level lineage from SQL queries and visualizes it through an interactive interface. LINEAGEX achieves high coverage and accuracy for column lineage extraction by intelligently traversing query parse trees and handling ambiguities. The demonstration walks through use cases of building lineage graphs and troubleshooting data quality issues. LINEAGEX is open sourced at https://github.com/sfu-db/lineagex and our video demonstration is at https://youtu.be/5LaBBDDitlw

cs.DB↗

arXiv2022

Understanding Queries by Conditional Instances

A powerful way to understand a complex query is by observing how it operates on data instances. However, specific database instances are not ideal for such observations: they often include large amounts of superfluous details that are not only irrelevant to understanding the query but also cause cognitive overload; and one specific database may not be enough. Given a relational query, is it possible to provide a simple and generic "representative" instance that (1) illustrates how the query can be satisfied, (2) summarizes all specific instances that would satisfy the query in the same way by abstracting away unnecessary details? Furthermore, is it possible to find a collection of such representative instances that together completely characterize all possible ways in which the query can be satisfied? This paper takes initial steps towards answering these questions. We design what these representative instances look like, define what they stand for, and formalize what it means for them to satisfy a query in "all possible ways." We argue that this problem is undecidable for general domain relational calculus queries, and develop practical algorithms for computing a minimum collection of such instances subject to other constraints. We evaluate the efficiency of our approach experimentally, and show its effectiveness in helping users debug relational queries through a user study.

cs.DB↗

arXiv2021

Putting Things into Context: Rich Explanations for Query Answers using Join Graphs (extended version)

In many data analysis applications, there is a need to explain why a surprising or interesting result was produced by a query. Previous approaches to explaining results have directly or indirectly used data provenance (input tuples contributing to the result(s) of interest), which is limited by the fact that relevant information for explaining an answer may not be fully contained in the provenance. We propose a new approach for explaining query results by augmenting provenance with information from other related tables in the database. We develop a suite of optimization techniques, and demonstrate experimentally using real datasets and through a user study that our approach produces meaningful results by efficiently navigating the large search space of possible explanations.

cs.DB↗

arXiv2020

Snippext: Semi-supervised Opinion Mining with Augmented Data

Online services are interested in solutions to opinion mining, which is the problem of extracting aspects, opinions, and sentiments from text. One method to mine opinions is to leverage the recent success of pre-trained language models which can be fine-tuned to obtain high-quality extractions from reviews. However, fine-tuning language models still requires a non-trivial amount of training data. In this paper, we study the problem of how to significantly reduce the amount of labeled training data required in fine-tuning language models for opinion mining. We describe Snippext, an opinion mining system developed over a language model that is fine-tuned through semi-supervised learning with augmented data. A novelty of Snippext is its clever use of a two-prong approach to achieve state-of-the-art (SOTA) performance with little labeled training data through: (1) data augmentation to automatically generate more labeled training data from existing ones, and (2) a semi-supervised learning technique to leverage the massive amount of unlabeled data in addition to the (limited amount of) labeled data. We show with extensive experiments that Snippext performs comparably and can even exceed previous SOTA results on several opinion mining tasks with only half the training data required. Furthermore, it achieves new SOTA results when all training data are leveraged. By comparison to a baseline pipeline, we found that Snippext extracts significantly more fine-grained opinions which enable new opportunities of downstream applications.

cs.CL↗

arXiv2019

Explaining Wrong Queries Using Small Examples

For testing the correctness of SQL queries, e.g., evaluating student submissions in a database course, a standard practice is to execute the query in question on some test database instance and compare its result with that of the correct query. Given two queries $Q_1$ and $Q_2$, we say that a database instance $D$ is a counterexample (for $Q_1$ and $Q_2$) if $Q_1(D)$ differs from $Q_2(D)$; such a counterexample can serve as an explanation of why $Q_1$ and $Q_2$ are not equivalent. While the test database instance may serve as a counterexample, it may be too large or complex to read and understand where the inequivalence comes from. Therefore, in this paper, given a known counterexample $D$ for $Q_1$ and $Q_2$, we aim to find the smallest counterexample $D' \subseteq D$ where $Q_1(D') \neq Q_2(D')$. The problem in general is NP-hard. We give a suite of algorithms for finding the smallest counterexample for different classes of queries, some more tractable than others. We also present an efficient provenance-based algorithm for SPJUD queries that uses a constraint solver, and extend it to more complex queries with aggregation, group-by, and nested queries. We perform extensive experiments indicating the effectiveness and scalability of our solution on student queries from an undergraduate database course and on queries from the TPC-H benchmark. We also report a user study from the course where we deployed our tool to help students with an assignment on relational algebra.

cs.DB↗