Auto-regressive (AR) models have recently made notable progress in image generation, achieving performance comparable to diffusion-based approaches. However, their computational intensity and sequential nature impede on-device deployment, causing disruptive latency. We address this via a cloud-device collaboration framework \textbf{CIAR}, which utilizes on-device self-verification to handle two key properties of visual synthesis: \textit{the vast token vocabulary} required for high-fidelity images and \textit{inherent spatial redundancy} which leads to extreme predictability in homogeneous regions, while object boundaries exhibit high uncertainty. Uniform verification wastes resources on such redundant tokens. Our solution centers on an on-device token uncertainty quantifier, which adopts continuous probability intervals to accelerate processing and make it feasible for large visual vocabularies instead of conventional discrete solution sets. Additionally, we incorporate a Interval-enhanced decoding module to further speed up decoding while maintaining visual fidelity and semantic consistency via a distribution alignment training strategy. Extensive experiments demonstrate that CIAR achieves a 2.18× speed-up and reduces cloud requests by 70\%, while preserving image quality compared to existing methods.

Semi-structured documents integrate diverse interleaved data elements (e.g., tables, charts, hierarchical paragraphs) arranged in various and often irregular layouts. These documents are widely observed across domains and account for a large portion of real-world data. However, existing methods struggle to support natural language question answering over these documents due to three main technical challenges: (1) The elements extracted by techniques like OCR are often fragmented and stripped of their original semantic context, making them inadequate for analysis. (2) Existing approaches lack effective representations to capture hierarchical structures within documents (e.g., associating tables with nested chapter titles) and to preserve layout-specific distinctions (e.g., differentiating sidebars from main content). (3) Answering questions often requires retrieving and aligning relevant information scattered across multiple regions or pages, such as linking a descriptive paragraph to table cells located elsewhere in the document. To address these issues, we propose MoDora, an LLM-powered system for semi-structured document analysis. First, we adopt a local-alignment aggregation strategy to convert OCR-parsed elements into layout-aware components, and conduct type-specific information extraction for components with hierarchical titles or non-text elements. Second, we design the Component-Correlation Tree (CCTree) to hierarchically organize components, explicitly modeling inter-component relations and layout distinctions through a bottom-up cascade summarization process. Finally, we propose a question-type-aware retrieval strategy that supports (1) layout-based grid partitioning for location-based retrieval and (2) LLM-guided pruning for semantic-based retrieval. Experiments show MoDora outperforms baselines by 5.97

Road network data provides rich information about cities, but processing a large volume of worldwide OpenStreetMap (OSM) data is computationally intensive, and the resulting graphs are often difficult to unify for benchmarking downstream tasks. Existing graph learning benchmarks fail to capture the billion-scale and unique topological properties of real-world road networks, leading to a gap in our understanding of model scalability. To study and close this gap, we process OpenStreetMap data with distributed cloud computing using 5,000 cores and release OSM+, a structured worldwide 1-billion-vertex road network graph dataset designed for high accessibility and usability. OSM+ is open source and globally downloadable, and it provides an open-box graph structure together with an easy spatial query interface. We demonstrate the utility of OSM+ through three illustrative use cases: city boundary detection, traffic prediction, and traffic policy control. For traffic prediction, we construct a new 31-city benchmark by processing traffic data and combining it with OSM+, enabling broader spatial coverage and more comprehensive evaluation than previously frequently-used datasets, while scaling from hundreds of road network intersections to thousands. For traffic policy control, we release a new six-city dataset at a much larger scale, introducing challenges for thousand-scale multi-agent coordination. In addition, we provide comprehensive data processing tools that support integrating multimodal spatial-temporal data with OSM+ for geospatial foundation model training, thereby expediting the discovery of compelling scientific insights.

The operation and maintenance (O&M) of database systems is critical to ensuring system availability and performance, typically requiring expert experience (e.g., identifying metric-to-anomaly relations) for effective diagnosis and recovery. However, existing automatic database O&M methods, including commercial products, cannot effectively utilize expert experience. On the one hand, rule-based methods only support basic O&M tasks (e.g., metric-based anomaly detection), which are mostly numerical equations and cannot effectively incorporate literal O&M experience (e.g., troubleshooting guidance in manuals). On the other hand, LLM-based methods, which retrieve fragmented information (e.g., standard documents + RAG), often generate inaccurate or generic results. To address these limitations, we present DBAIOps, a novel hybrid database O&M system that combines reasoning LLMs with knowledge graphs to achieve DBA-style diagnosis. First, DBAIOps introduces a heterogeneous graph model for representing the diagnosis experience, and proposes a semi-automatic graph construction algorithm to build that graph from thousands of documents. Second, DBAIOps develops a collection of (800+) reusable anomaly models that identify both directly alerted metrics and implicitly correlated experience and metrics. Third, for each anomaly, DBAIOps proposes a two-stage graph evolution mechanism to explore relevant diagnosis paths and identify missing relations automatically. It then leverages a reasoning LLM (e.g., DeepSeek-R1) to infer root causes and generate clear diagnosis reports for both DBAs and common users. Our evaluation over four mainstream database systems (Oracle, MySQL, PostgreSQL, and DM8) demonstrates that DBAIOps outperforms state-of-the-art baselines, 34.85% and 47.22% higher in root cause and human evaluation accuracy, respectively.

With the rapid rise of intelligent data services, modern enterprises increasingly require efficient, multimodal, and cost-effective data analytics infrastructures. However, in ByteDance's production environments, existing systems fall short due to limitations such as I/O-inefficient multimodal storage, inflexible query optimization (e.g., failing to optimize multimodal access patterns), and performance degradation caused by resource disaggregation (e.g., loss of data locality in remote storage). To address these challenges, we introduce ByteHouse (https://bytehouse.cloud), a cloud-native data warehouse designed for real-time multimodal data analytics. The storage layer integrates a unified table engine that provides a two-tier logical abstraction and physically consistent layout, SSD-backed cluster-scale cache (CrossCache) that supports shared caching across compute nodes, and virtual file system (NexusFS) that enable efficient local access on compute nodes. The compute layer supports analytical, batch, and incremental execution modes, with tailored optimizations for hybrid queries (e.g., runtime filtering over tiered vector indexes). The control layer coordinates global metadata and transactions, and features an effective optimizer enhanced by historical execution traces and AI-assisted plan selection. Evaluations on internal and standard workloads show that ByteHouse achieves significant efficiency improvement over existing systems.

Machine learning models are widely integrated into modern mobile apps to analyze user behaviors and deliver personalized services. Ensuring low-latency on-device model execution is critical for maintaining high-quality user experiences. While prior research has primarily focused on accelerating model inference with given input features, we identify an overlooked bottleneck in real-world on-device model execution pipelines: extracting input features from raw application logs. In this work, we explore a new direction of feature extraction optimization by analyzing and eliminating redundant extraction operations across different model features and consecutive model inferences. We then introduce AutoFeature, an automated feature extraction engine designed to accelerate on-device feature extraction process without compromising model inference accuracy. AutoFeature comprises three core designs: (1) graph abstraction to formulate the extraction workflows of different input features as one directed acyclic graph, (2) graph optimization to identify and fuse redundant operation nodes across different features within the graph; (3) efficient caching to minimize operations on overlapping raw data between consecutive model inferences. We implement a system prototype of AutoFeature and integrate it into five industrial mobile services spanning search, video and e-commerce domains. Online evaluations show that AutoFeature reduces end-to-end on-device model execution latency by 1.33x-3.93x during daytime and 1.43x-4.53x at night.

Deploying Large Language Models (LLMs) on edge devices is increasingly critical due to network and privacy constraints. Key-Value (KV) cache reuse and elastic LLM inference are both essential for on-device LLM deployment as the available resources on edge devices are limited and vary over time. However, directly using the KV caches from a different precision model can cause performance degradation due to numerical deviation of the KV cache. To address this issue, we present ElastKV, a scheme that reuses precomputed KV caches generated by mixed-precision LLMs by selectively recomputing a small subset of KV values. ElastKV first filters tokens using a recomputation rule, then ranks them based on their estimated recomputation priority. Unlike prior methods, ElastKV estimates this priority by combining each token's KV deviation with its estimated influence on output tokens, which is derived from pairwise token influence estimation and influence paths. We evaluate ElastKV on four open-source LLMs with different quantization bit-widths over four datasets. Results show that ElastKV limits average accuracy loss to within 2% and provides an average recomputation speedup of nearly 3× compared to full recomputation and achieves a better speed-quality trade-off than all state-of-the-art KV cache reuse methods.

Workspace learning requires AI agents to identify, reason over, exploit, and update explicit and implicit dependencies among heterogeneous files in a worker's workspace, enabling them to complete both routine and advanced tasks effectively. Despite its importance, existing relevant benchmarks largely evaluate agents on pre-specified or synthesized files with limited real-world dependencies, leaving workspace-level evaluation underexplored. To this end, we introduce Workspace-Bench, a benchmark for evaluating AI agents on Workspace Learning involving Large-Scale File Dependencies. We construct realistic workspaces with 5 worker profiles, 74 file types, 20,476 files (up to 20GB) and curate 388 tasks, each with its own file dependency graph, evaluated across 7,399 total rubrics that require cross-file retrieval, contextual reasoning, and adaptive decision-making. We further provide Workspace-Bench-Lite, a 100-task subset that preserves the benchmark distribution while reducing evaluation costs by about 70

With the development of federated learning and mobile computing, federated bandit frameworks have been proposed to enable multiple clients to explore and exploit collaboratively under the guarantee of data privacy. However, none of them take the widespread budget constraints into account. In this work, we introduce the Federated linear Bandits framework with Knapsacks (FBwK), where M clients can pull K arms with linear rewards and costs to minimize the total regret under the coordination of a central server. For FBwK, we propose a new definition of OPT solution and a bandit algorithm, namely FedUCBwK. Specifically, we design a uniform policy parameter update threshold for each client, which balances regret, communication, and computation costs. In this process, the central server uses a factor on knapsack constraints to pace the budget consumption and preserve the clients’ privacy by transmitting model updates instead of raw data. We conduct a theoretical analysis and show that FedUCBwK achieves a sublinear regret with logarithmic communication and computation. Evaluation on real-world mobile application datasets demonstrates that FedUCBwK consistently outperforms existing methods across diverse tasks.

Considering limited on-device resources, current practices are attempting to deploy a system-level mixture-of-experts (MoE)-based foundation LLM on a mobile device to serve multiple apps and support mobile intelligence. However, mobile apps are hard to customize their services that require fine-tuning adapters associated with the LLM using private in-app data. The difficulty arises due to both the limited on-device resources and the restricted control that apps have over the foundation LLM. To address this issue, in this work, we propose LiteMoE, a novel proxy submodel tuning framework that supports mobile apps to efficiently fine-tune customized adapters on devices using proxy submodels. The key technique behind LiteMoE is a post-training submodel extraction method, whereby without additional re-training, we can identify and reserve critical experts, match and merge moderate experts, to extract a lightweight and effective proxy submodel from the foundation LLM for a specific app. To further enhance scalability and adaptability, LiteMoE incorporates adapter reuse and continuous tuning mechanisms to handle multi-task requirements and evolving user preferences. We implemented a prototype of LiteMoE and evaluated it over various MoE-based LLMs and mobile computing tasks. The results show that with LiteMoE, mobile apps are able to fine-tune customized adapters on resource-limited devices, achieving 12.7% accuracy improvement and 6.6 x memory reduction compared with operating the original foundation LLM.

This paper envisions a quantum database (Qute) that treats quantum computation as a first-class execution option. Unlike prior simulation-based methods that either run quantum algorithms on classical machines or adapt existing databases for quantum simulation, Qute instead (i) compiles an extended form of SQL into gate-efficient quantum circuits, (ii) employs a hybrid optimizer to dynamically select between quantum and classical execution plans, (iii) introduces selective quantum indexing, and (iv) designs fidelity-preserving storage to mitigate current qubit constraints. We also present a three-stage evolution roadmap toward quantum-native database. Finally, by deploying Qute on a real quantum processor (origin_wukong), we show that it outperforms a classical baseline at scale, and we release an open-source prototype at https://github.com/weAIDB/Qute.

Active Learning (AL) aims to improve model performance by selecting the most informative samples from a pool of unlabeled data under a limited annotation budget. However, most existing active learning methods are designed for centralized settings and cannot be directly applied to edge-cloud collaborative learning systems, which have been widely adopted for edge intelligence. In this work, we propose Edge-Cloud Collaborative Active Learning (ECAL), an efficient two-stage data selection framework. ECAL employs distributed edge nodes to collaboratively pre-filter data using coordination information generated by the cloud, thereby reducing data transmission overhead between edge nodes and the cloud. Furthermore, ECAL offloads part of the data selection workload from the cloud to edge nodes. Through extensive experiments, we demonstrate that ECAL significantly improves model performance under fixed annotation budgets compared with state-of-the-art non-collaborative baselines.

Memory for large language model (LLM) agents has rapidly evolved from simple retrieval-augmented mechanisms into a data management system that supports persistent information storage, retrieval, update, consolidation, and dynamic lifecycle governance throughout agent execution. Despite this evolution, existing evaluations still benchmark agent memory mainly through end-to-end task success metrics (e.g., F1, BLEU), while treating the underlying system as a monolithic black box. As a result, critical system-level concerns, including operational costs, architectural trade-offs across memory modules, and robustness under dynamic knowledge updates, remain insufficiently explored. In this paper, we present a systematic experimental study of agent memory from a data management perspective. We propose an analytical framework that decomposes agent memory into four core modules: memory representation and storage, extraction, retrieval and routing, and maintenance. Under this framework, we evaluate 12 representative memory systems and two reference baselines across five benchmark workloads spanning 11 datasets. Our extensive end-to-end evaluation shows that no single architecture dominates across all scenarios; instead, effectiveness depends heavily on how well the memory structure aligns with the workload bottleneck. Furthermore, through fine-grained ablation studies, we quantify their individual effects on representation fidelity, retrieval precision, update correctness, and long-horizon stability. Finally, we reveal cost-performance trade-offs under realistic workloads, showing localized maintenance is more cost-efficient than global reorganization. Based on these findings, we identify promising directions towards building truly agent-native memory systems. The code is publicly available at https://github.com/OpenDataBox/MemoryData.

This study shares lessons learned from a culturally responsive program designed to broaden computer science (CS) participation and build foundational AI literacy among minority high school students from low-income communities. The program emphasized collaborative, project-based app design as a strategy to foster CS interest and develop AI literacy skills. The program included two phases: a one-month summer academy and a semester-long fall project. In the summer academy, 42 students engaged in MIT App Inventor programming, Agile software development, and AI lessons covering machine learning concepts and AI ethics. In the fall, 26 students continued in bi-weekly online Zoom meetings, collaboratively designing personally meaningful mobile app prototypes, such as a hair-care recommendation app for Black women and a mental-health support app for teenagers. Evaluation methods included surveys, focus group interviews, and app design presentations. Results indicated participants' growth in problem-solving, time management, and collaboration, as well as an increased interest in CS career pathways previously not envisioned. However, students encountered several challenges, particularly in using and integrating AI. These included fear or hesitation in using AI, over-reliance on AI for research or solutions, and a tendency to focus on coding without understanding how AI works. This work highlights both the promise and the complexities of promoting AI literacy through mobile app design in pre-college settings. We emphasize the need for scaffolded instruction that supports students in understanding the fundamentals and applications of AI through concrete, youth-relevant activities, while reinforcing the critical role of human judgment when working with AI.

Users interacting with large language models (LLMs) under their real identifiers often unknowingly risk disclosing private information. Automatically notifying users whether their queries leak privacy and which phrases leak what private information has therefore become a practical need. Existing privacy detection methods, however, were designed for different objectives and application domains, typically tagging personally identifiable information (PII) in anonymous content, which is insufficient in real-name interaction scenarios with LLMs. In this work, to support the development and evaluation of privacy detection models for LLM interactions that are deployable on local user devices, we construct a large-scale multilingual dataset with 249K user queries and 154K annotated privacy phrases. In particular, we build an automated privacy annotation pipeline with strong LLMs to automatically extract privacy phrases from dialogue datasets and annotate leaked information. We also design evaluation metrics at the levels of privacy leakage, extracted privacy phrase, and privacy information. We further establish baseline methods using light-weight LLMs with both tuning-free and tuning-based methods, and report a comprehensive evaluation of their performance. Evaluation results demonstrate that the 1B model, fine-tuned with our dataset, outperforms the directly prompted 72B model. However, a gap remains between current performance and the requirements of real-world LLM applications, motivating future research into more effective local privacy detection methods grounded in our dataset(1).

Enterprises commonly deploy heterogeneous database systems, each of which owns a distinct SQL dialect with different syntax rules, built-in functions, and execution constraints. However, most existing NL2SQL methods assume a single dialect (e.g., SQLite) and struggle to produce queries that are both semantically correct and executable on target engines. Prompt-based approaches tightly couple intent reasoning with dialect syntax, rule-based translators often degrade native operators into generic constructs, and multi-dialect fine-tuning suffers from cross-dialect interference. In this paper, we present Dial, a knowledge-grounded framework for dialect-specific NL2SQL. Dial introduces: (1) a Dialect-Aware Logical Query Planning module that converts natural language into a dialect-aware logical query plan via operator-level intent decomposition and divergence-aware specification; (2) HINT-KB, a hierarchical intent-aware knowledge base that organizes dialect knowledge into (i) a canonical syntax reference, (ii) a declarative function repository, and (iii) a procedural constraint repository; and (3) an execution-driven debugging and semantic verification loop that separates syntactic recovery from logic auditing to prevent semantic drift. We construct DS-NL2SQL, a benchmark covering six major database systems with 2,218 dialect-specific test cases. Experimental results show that Dial consistently improves translation accuracy by 10.25

Functionality-correct repository setup aims to configure execution environments (e.g., dependencies, build scripts) to successfully execute a repository's documented features. It presents significant challenges due to diverse, repository-specific failures, including dependency incompatibilities, missing toolchains, incomplete installations, and verification-strategy mismatches. Existing LLM agents struggle to robustly resolve these issues, specifically failing to support (1) cross-repository experience transfer, (2) multi-step trial-and-repair under non-invertible state changes, and (3) robust verification of setup outcomes to distinguish setup-induced failures from repository bugs. To address this, we introduce SetupX, an experiential learning-based setup framework. First, we construct a Self-Evolving Experience Representation (XPU), a dual-modality knowledge unit encoding setup signals, textual guidance, executable actions to dynamically transfer verified environment fixes to unseen repositories. Second, we employ Experience-Augmented Speculative Execution backed by a LIFO Docker snapshot stack, enabling the agent to proactively trial fixes and safely roll back to known-good states. Third, we introduce a Prosecutor-Judge Verification Protocol that separates evidence collection from final judgment, enabling more reliable setup verification beyond superficial build-time metrics. Evaluation results on carefully-crafted benchmarks show SetupX achieves highest performance (e.g., 92

Automatically generating slide decks from source documents is an important application of large language models (LLMs). Existing benchmarks primarily assess slide completeness and technical depth, while overlooking the target audience as a critical real-world factor. For instance, specialists demand rigorous proofs, whereas decision-makers prioritize actionable conclusions. To bridge this gap, we introduce X+Slides, a benchmark specifically designed for audience-conditioned slide generation. Built on a diverse corpus spanning 113 topics and seven presentation scenes, X+Slides employs a dynamic evaluation framework constructed from 8,133 deduplicated, source-grounded probes. By assigning audience-specific utility weights to the same source-grounded probes, X+Slides reports four complementary metrics: Audience Coverage measures how much audience-essential information is conveyed, Domain-wise Coverage shows which information types are covered, Efficiency measures delivered utility per unit of attention cost, and Correctness verifies whether slide claims are supported by the source. Experiments on DeepPresenter, SlideTailor, and NotebookLM show that current systems can recover a substantial but still incomplete part of audience-essential information: at τ_A=0.7, DeepPresenter reaches a best Audience Coverage of 0.714, SlideTailor reaches 0.594, and the NotebookLM ablation reaches 0.853 while showing clear grounding differences. These results indicate that visual quality and broad topic coverage should not be treated as evidence support without source-grounded evaluation.

Modern content platforms offer paid promotion to mitigate cold start by allocating exposure via auctions. Our empirical analysis reveals a counterintuitive flaw in this paradigm: while promotion rescues low-to-medium quality content, it can harm high-quality content by forcing exposure to suboptimal audiences, polluting engagement signals and downgrading future recommendation. We recast content promotion as a dual-objective optimization that balances short-term value acquisition with long-term model improvement. To make this tractable at bid time in content promotion, we introduce a decomposable surrogate objective, gradient coverage, and establish its formal connection to Fisher Information and optimal experimental design. We design a two-stage auto-bidding algorithm based on Lagrange duality that dynamically paces budget through a shadow price and optimizes impression-level bids using per-impression marginal utilities. To address missing labels at bid time, we propose a confidence-gated gradient heuristic, paired with a zeroth-order variant for black-box models that reliably estimates learning signals in real time. We provide theoretical guarantees, proving monotone submodularity of the composite objective, sublinear regret in online auction, and budget feasibility. Extensive offline experiments on synthetic and real-world datasets validate the framework: it outperforms baselines, achieves superior final AUC/LogLoss, adheres closely to budget targets, and remains effective when gradients are approximated zeroth-order. These results show that strategic, information-aware promotion can improve long-term model performance and organic outcomes beyond naive impression-maximization strategies.