Searcharxiv⌕ Search

arXiv subjects

Yiru Chen

Publications and source records attributed to Yiru Chen.

14 recordsLinked to original sources
arXiv2026

TACO: Task-Aware Column Description Generation Using LLMs

Generating accurate and informative column descriptions (e.g. "membership status of customers" for the column name "cust_mem") is essential for a wide range of downstream NLP tasks on tabular data, including NL2SQL, table question answering, and entity linking. This problem arises in enterprises, domain sciences, government data portals, and so on. Despite its importance, most real-world datasets suffer from missing or cryptic documentation, often due to abbreviated column names or domain-specific jargon. Existing approaches largely rely on single-prompt large language models (LLMs), which struggle with three key issues: (i) inconsistent or incorrect handling of abbreviations, (ii) hallucinated or incomplete descriptions, and (iii) redundancy or vagueness that hinders downstream performance. We present TACO, a task-aware framework for automatic column description generation using LLMs. TACO introduces a three-step pipeline: (1) abbreviation expansion, which standardizes column names; (2) description generation, which produces initial semantic descriptions enriched with synonyms and search-oriented keywords; and (3) description revision, which refines these outputs using simulated downstream tasks. In addition, we investigate human-in-the-loop extensions and release new evaluation datasets for entity linking and schema enrichment. Extensive experiments across public and proprietary datasets show that TACO consistently outperforms existing methods, improving downstream task performance by up to 32%.

cs.CL↗
arXiv2026

ISEE: Interactive Semantic Enrichment for Database Fields

LLM-based agents are increasingly being deployed for data-related tasks, including data sense-making, exploration, and retrieval. However, their performance heavily depends on the clarity and completeness of data semantics. In practice, many field descriptions remain ambiguous or incomplete, as much of the essential context (e.g., the meaning of a customized field) originates from users' domain knowledge and is rarely documented publicly. This gap restricts the agents' task performance in downstream tasks, such as entity-linking. To bridge this gap, we introduce a novel and comprehensive Interactive SEmantic Enrichment system (ISEE). Given a data field description, ISEE measures its quality through a scoring system, gathers domain knowledge, and collaboratively enriches the semantics with users. Through a user study, automated user simulation, quantitative evaluation, and case study, we demonstrate that ISEE significantly reduces cognitive load, improves description quality, and enhances downstream task performance.

cs.AI↗
arXiv2025

A Decade of Systems for Human Data Interaction

Human-data interaction (HDI) presents fundamentally different challenges from traditional data management. HDI systems must meet latency, correctness, and consistency needs that stem from usability rather than query semantics; failing to meet these expectations breaks the user experience. Moreover, interfaces and systems are tightly coupled; neither can easily be optimized in isolation, and effective solutions demand their co-design. This dependence also presents a research opportunity: rather than adapt systems to interface demands, systems innovations and database theory can also inspire new interaction and visualization designs. We survey a decade of our lab's work that embraces this coupling and argue that HDI systems are the foundation for reliable, interactive, AI-driven applications.

cs.DB↗
arXiv2025

Adobe Summit Concierge Evaluation with Human in the Loop

Generative AI assistants offer significant potential to enhance productivity, streamline information access, and improve user experience in enterprise contexts. In this work, we present Summit Concierge, a domain-specific AI assistant developed for Adobe Summit. The assistant handles a wide range of event-related queries and operates under real-world constraints such as data sparsity, quality assurance, and rapid deployment. To address these challenges, we adopt a human-in-the-loop development workflow that combines prompt engineering, retrieval grounding, and lightweight human validation. We describe the system architecture, development process, and real-world deployment outcomes. Our experience shows that agile, feedback-driven development enables scalable and reliable AI assistants, even in cold-start scenarios.

cs.AI↗
arXiv2023

Improving Vision-and-Language Reasoning via Spatial Relations Modeling

Visual commonsense reasoning (VCR) is a challenging multi-modal task, which requires high-level cognition and commonsense reasoning ability about the real world. In recent years, large-scale pre-training approaches have been developed and promoted the state-of-the-art performance of VCR. However, the existing approaches almost employ the BERT-like objectives to learn multi-modal representations. These objectives motivated from the text-domain are insufficient for the excavation on the complex scenario of visual modality. Most importantly, the spatial distribution of the visual objects is basically neglected. To address the above issue, we propose to construct the spatial relation graph based on the given visual scenario. Further, we design two pre-training tasks named object position regression (OPR) and spatial relation classification (SRC) to learn to reconstruct the spatial relation graph respectively. Quantitative analysis suggests that the proposed method can guide the representations to maintain more spatial context and facilitate the attention on the essential visual regions for reasoning. We achieve the state-of-the-art results on VCR and two other vision-and-language reasoning tasks VQA, and NLVR.

cs.CV↗
arXiv2023

DIG: The Data Interface Grammar

Building interactive data interfaces is hard because the design of an interface depends on the data processing needs for the underlying analysis task, yet we do not have a good representation for analysis tasks. To fill this gap, this paper advocates for a Data Interface Grammar (DIG) as an intermediate representation of analysis tasks. We show that DIG is compatible with existing data engineering practices, compact to represent any analysis, simple to translate into an interface design, and amenable to offline analysis. We further illustrate the potential benefits of this abstraction, such as automatic interface generation, automatic interface backend optimization, tutorial generation, and workload generation.

cs.DB↗
arXiv2022

TSEXPLAIN: Explaining Aggregated Time Series by Surfacing Evolving Contributors

Aggregated time series are generated effortlessly everywhere, e.g., "total confirmed covid-19 cases since 2019" and "total liquor sales over time." Understanding "how" and "why" these key performance indicators (KPI) evolve over time is critical to making data-informed decisions. Existing explanation engines focus on explaining one aggregated value or the difference between two relations. However, this falls short of explaining KPIs' continuous changes over time. Motivated by this, we propose TSEXPLAIN, a system that explains aggregated time series by surfacing the underlying evolving top contributors. Under the hood, we leverage prior works on two-relations diff as a building block and formulate a K-Segmentation problem to segment the time series such that each segment after segmentation shares consistent explanations, i.e., contributors. To quantify consistency in each segment, we propose a novel within-segment variance design that is explanation-aware; to derive the optimal K-Segmentation scheme, we develop an efficient dynamic programming algorithm. Experiments on synthetic and real-world datasets show that our explanation-aware segmentation can effectively identify evolving explanations for aggregated time series and outperform explanation-agnostic segmentation. Further, we proposed an optimal selection strategy of K and several optimizations to speed up TSEXPLAIN for interactive user experience, achieving up to 13X efficiency improvement.

cs.DB↗
arXiv2022

NL2INTERFACE: Interactive Visualization Interface Generation from Natural Language Queries

We develop NL2INTERFACE to explore the potential of generating usable interactive multi-visualization interfaces from natural language queries. With NL2INTERFACE, users can directly write natural language queries to automatically generate a fully interactive multi-visualization interface without any extra effort of learning a tool or programming language. Further, users can interact with the interfaces to easily transform the data and quickly see the results in the visualizations.

cs.HC↗
arXiv2021

PI2: End-to-end Interactive Visualization Interface Generation from Queries

Interactive visual analysis interfaces are critical in nearly every data task. However, creating new interfaces is deeply challenging, as it requires the developer to understand the queries needed to express the desired analysis task, design the appropriate interface to express those queries for the task, and implement the interface using a combination of visualization, browser, server, and database technologies. Although prior work generates a set of interactive widgets that can express an input query log, this paper presents PI2, the first system to generate fully functional visual analysis interfaces from an example sequence of analysis queries. PI2 analyzes queries syntactically and represents a set of queries using a novel Difftree structure that encodes systematic variations between query abstract syntax trees. PI2 then maps each Difftree to a visualization that renders its results, the variations in each Difftree to interactions, and generates a good layout for the interface. We show that PI2 can express data-oriented interactions in existing visualization interaction taxonomies, reproduce or improve several real-world visual analysis interfaces, generate interfaces in 2-19s (median 6s), and scale linearly with the number of queries.

cs.DB↗
arXiv2021

HiT: Hierarchical Transformer with Momentum Contrast for Video-Text Retrieval

Video-Text Retrieval has been a hot research topic with the growth of multimedia data on the internet. Transformer for video-text learning has attracted increasing attention due to its promising performance. However, existing cross-modal transformer approaches typically suffer from two major limitations: 1) Exploitation of the transformer architecture where different layers have different feature characteristics is limited; 2) End-to-end training mechanism limits negative sample interactions in a mini-batch. In this paper, we propose a novel approach named Hierarchical Transformer (HiT) for video-text retrieval. HiT performs Hierarchical Cross-modal Contrastive Matching in both feature-level and semantic-level, achieving multi-view and comprehensive retrieval results. Moreover, inspired by MoCo, we propose Momentum Cross-modal Contrast for cross-modal learning to enable large-scale negative sample interactions on-the-fly, which contributes to the generation of more precise and discriminative representations. Experimental results on the three major Video-Text Retrieval benchmark datasets demonstrate the advantages of our method.

cs.CV↗
arXiv2020

Monte Carlo Tree Search for Generating Interactive Data Analysis Interfaces

Interactive tools like user interfaces help democratize data access for end-users by hiding underlying programming details and exposing the necessary widget interface to users. Since customized interfaces are costly to build, automated interface generation is desirable. SQL is the dominant way to analyze data and there already exists logs to analyze data. Previous work proposed a syntactic approach to analyze structural changes in SQL query logs and automatically generates a set of widgets to express the changes. However, they do not consider layout usability and the sequential order of queries in the log. We propose to adopt Monte Carlo Tree Search(MCTS) to search for the optimal interface that accounts for hierarchical layout as well as the usability in terms of how easy to express the query log.

cs.DB↗
arXiv2019

DeepBase: Deep Inspection of Neural Networks

Although deep learning models perform remarkably well across a range of tasks such as language translation and object recognition, it remains unclear what high-level logic, if any, they follow. Understanding this logic may lead to more transparency, better model design, and faster experimentation. Recent machine learning research has leveraged statistical methods to identify hidden units that behave (e.g., activate) similarly to human understandable logic, but those analyses require considerable manual effort. Our insight is that many of those studies follow a common analysis pattern, which we term Deep Neural Inspection. There is opportunity to provide a declarative abstraction to easily express, execute, and optimize them. This paper describes DeepBase, a system to inspect neural network behaviors through a unified interface. We model logic with user-provided hypothesis functions that annotate the data with high-level labels (e.g., part-of-speech tags, image captions). DeepBase lets users quickly identify individual or groups of units that have strong statistical dependencies with desired hypotheses. We discuss how DeepBase can express existing analyses, propose a set of simple and effective optimizations to speed up a standard Python implementation by up to 72x, and reproduce recent studies from the NLP literature.

cs.DB↗
arXiv2018

PSFGAN: a generative adversarial network system for separating quasar point sources and host galaxy light

The study of unobscured active galactic nuclei (AGN) and quasars depends on the reliable decomposition of the light from the AGN point source and the extended host galaxy light. The problem is typically approached using parametric fitting routines using separate models for the host galaxy and the point spread function (PSF). We present a new approach using a Generative Adversarial Network (GAN) trained on galaxy images. We test the method using Sloan Digital Sky Survey (SDSS) r-band images with artificial AGN point sources added which are then removed using the GAN and with parametric methods using GALFIT. When the AGN point source PS is more than twice as bright as the host galaxy, we find that our method, PSFGAN, can recover PS and host galaxy magnitudes with smaller systematic error and a lower average scatter ($49\%$). PSFGAN is more tolerant to poor knowledge of the PSF than parametric methods. Our tests show that PSFGAN is robust against a broadening in the PSF width of $\pm 50\%$ if it is trained on multiple PSF's. We demonstrate that while a matched training set does improve performance, we can still subtract point sources using a PSFGAN trained on non-astronomical images. While initial training is computationally expensive, evaluating PSFGAN on data is more than $40$ times faster than GALFIT fitting two components. Finally, PSFGAN it is more robust and easy to use than parametric methods as it requires no input parameters.

astro-ph.GA↗