arXiv subjects

Sharad Mehrotra

Publications and source records attributed to Sharad Mehrotra.

At least 19 recordsLinked to original sources

arXiv2026

ReBound: Reuse-Aware Privacy For Interactive Decision Support

Differentially private decision support frameworks answer complex aggregate threshold queries with formal bounds on false negative and false positive rates, but treat each query independently with no memory of past results. In practice, analysts work interactively, issuing sequences of related queries that refine bounds, adjust thresholds, or derive new functions from previous ones. We propose ReBound, a framework that reuses cached results from previous queries to answer new queries at reduced or zero additional privacy cost while maintaining formal utility guarantees. ReBound introduces a reuse framework for multiple refinement types, a cache graph structure for efficient lookup of reusable results, and a negotiation mechanism for when requested bounds cannot be met within budget.

cs.CR↗

arXiv2026

MADRAG: Multi-Agent Debate with Retrieval-Augmented Generation for Training-Free Analytic Essay Scoring

We present MADRAG, a training-free framework for analytic essay scoring that combines multi-agent reasoning with retrieval-augmented grounding. Unlike standard LLM-as-judge approaches, which are prone to bias and unstable scoring, MADRAG decomposes evaluation into an interactive process: an Advocate identifies strengths, a Skeptic critiques weaknesses, and a Judge aggregates their arguments into a final score. Crucially, the Judge is augmented with rubric-aligned exemplar retrieval, enabling calibration through comparison with scored examples. Our results show that MADRAG significantly outperforms prompt-based baselines while approaching the performance of supervised systems without requiring task-specific training. Ablation studies demonstrate that retrieval drives calibration gains, while debate improves reasoning on higher-level traits. Our findings highlight the complementary roles of structured interaction and external memory in reliable LLM-based evaluation.

cs.MA↗

arXiv2026

Robust and Explainable Divide-and-Conquer Learning for Intrusion Detection

Machine learning-based intrusion detection requires complex models to capture patterns in high-dimensional, noisy, and class-imbalanced raw network traffic, yet deploying such models remains impractical on resource-constrained devices with limited processing power and memory. In this paper, we present a correlation-aware divide-and-conquer learning technique that decomposes a complex learning problem into smaller, more manageable subproblems. This enables lightweight models as simple as decision trees to be trained on focused subtasks, yielding up to 43.3% higher local accuracy and up to 257 times reduction in model size on real-world network intrusion detection datasets, while also improving adversarial robustness and explainability.

cs.LG↗

arXiv2026

Don't Stir the Pot! Authorized Vector Data Retrieval via Access-Aware Indexing

Vector databases increasingly enforce role-based access control, where each top-k approximate nearest neighbor query must return only vectors the querying role is authorized to access. Two extremes bracket the design space. A single global index built over all vectors avoids duplication but wastes search effort on unauthorized vectors and degrades recall, while an oracle index, built with all authorized vectors to the query roles, searches only authorized vectors but duplicates every shared vector between roles or queries. We present Veda and its efficient variant EffVeda, two indexing strategies built on an access-aware lattice to address access control in vector databases. The methods first partitions the dataset into disjoint data blocks by role combination, then leverage the structure of the access-aware lattice to apply copy and merge operations to group co-accessed blocks under a user-specified storage budget. Large nodes in the lattice are then indexed with HNSW, while small nodes are retained for linear scan. To facilitate query processing on the lattice, our methods construct a query plan that selects the minimal set of nodes that covers all authorized data for each role. At query time, coordinated search first queries pure (authorized-only) nodes to populate a global top-k heap, then leverages the resulting distance bound of the k-th data in the heap to prune exploration on impure nodes. Evaluations show that our methods deliver higher throughput at high recall while closely tracking the storage budget.

cs.DB↗

arXiv2026

Inference-Aware & Privacy-Preserving Deletion in Databases

Deletion is a fundamental database operation, yet modern systems often fail to provide the privacy guarantee that users expect from it. A deleted value may disappear from query results and even from physical storage, yet remain inferable from dependencies, derived data, or traces exposed by the deletion event itself. Meaningful deletion, therefore, requires more than logical removal or physical erasure; it requires a privacy guarantee that limits what remains inferable after deletion. In this paper, we take an inference-centric view of deletion, focusing on two leakage channels: leakage from the post-deletion state and leakage from the deletion pattern itself. We use this lens to distinguish logical, physical, and semantic deletion, organize the design space of deletion operations, and highlight open research challenges for building deletion mechanisms with meaningful privacy guarantees in database systems.

cs.DB↗

arXiv2026

SPRINT: Semi-supervised Prototypical Representation for Few-Shot Class-Incremental Tabular Learning

Real-world systems must continuously adapt to novel concepts from limited data without forgetting previously acquired knowledge. While Few-Shot Class-Incremental Learning (FSCIL) is established in computer vision, its application to tabular domains remains largely unexplored. Unlike images, tabular streams (e.g., logs, sensors) offer abundant unlabeled data, a scarcity of expert annotations and negligible storage costs, features ignored by existing vision-based methods that rely on restrictive buffers. We introduce SPRINT, the first FSCIL framework tailored for tabular distributions. SPRINT introduces a mixed episodic training strategy that leverages confidence-based pseudo-labeling to enrich novel class representations and exploits low storage costs to retain base class history. Extensive evaluation across six diverse benchmarks spanning cybersecurity, healthcare, and ecological domains, demonstrates SPRINT's cross-domain robustness. It achieves a state-of-the-art average accuracy of 77.37% (5-shot), outperforming the strongest incremental baseline by 4.45%.

cs.CV↗

arXiv2025

Scalable Enforcement of Fine Grained Access Control Policies in Relational Database Management Systems

The proliferation of smart technologies and evolving privacy regulations such as the GDPR and CPRA has increased the need to manage fine-grained access control (FGAC) policies in database management systems (DBMSs). Existing approaches to enforcing FGAC policies do not scale to thousands of policies, leading to degraded query performance and reduced system effectiveness. We present Sieve, a middleware for relational DBMSs that combines query rewriting and caching to optimize FGAC policy enforcement. Sieve rewrites a query with guarded expressions that group and filter policies and can efficiently use indexes in the DBMS. It also integrates a caching mechanism with an effective replacement strategy and a refresh mechanism to adapt to dynamic workloads. Experiments on two DBMSs with real and synthetic datasets show that Sieve scales to large datasets and policy corpora, maintaining low query latency and system load and improving policy evaluation performance by between 2x and 10x on workloads with 200 to 1,200 policies. The caching extension further improves query performance by between 6 and 22 percent under dynamic workloads, especially with larger cache sizes. These results highlight Sieve's applicability for real-time access control in smart environments and its support for efficient, scalable management of user preferences and privacy policies.

cs.DB↗

arXiv2025

GenIE - Simulator-Driven Iterative Data Exploration for Scientific Discovery

Physics-based simulators play a critical role in scientific discovery and risk assessment, enabling what-if analyses for events like wildfires and hurricanes. Today, databases treat these simulators as external pre-processing steps. Analysts must manually run a simulation, export the results, and load them into a database before analysis can begin. This linear workflow is inefficient, incurs high latency, and hinders interactive exploration, especially when the analysis itself dictates the need for new or refined simulation data. We envision a new database paradigm, entitled GenIE, that seamlessly integrates multiple simulators into databases to enable dynamic orchestration of simulation workflows. By making the database "simulation-aware," GenIE can dynamically invoke simulators with appropriate parameters based on the user's query and analytical needs. This tight integration allows GenIE to avoid generating data irrelevant to the analysis, reuse previously generated data, and support iterative, incremental analysis where results are progressively refined at interactive speeds. We present our vision for GenIE, designed as an extension to PostgreSQL, and demonstrate its potential benefits through comprehensive use cases: wildfire smoke dispersion analysis using WRF-SFIRE and HYSPLIT, and hurricane hazard assessment integrating wind, surge, and flood models. Our preliminary experiments show how GenIE can transform these slow, static analyses into interactive explorations by intelligently managing the trade-off between simulation accuracy and runtime across multiple integrated simulators. We conclude by highlighting the challenges and opportunities ahead in realizing the full vision of GenIE as a cornerstone for next-generation scientific data analysis.

cs.DB↗

arXiv2025

NOMAD -- Navigating Optimal Model Application to Datastreams

NOMAD (Navigating Optimal Model Application for Datastreams) is an intelligent framework for data enrichment during ingestion that optimizes realtime multiclass classification by dynamically constructing model chains, i.e ,sequences of machine learning models with varying cost-quality tradeoffs, selected via a utilitybased criterion. Inspired by predicate ordering techniques from database query processing, NOMAD leverages cheaper models as initial filters, proceeding to more expensive models only when necessary, while guaranteeing classification quality remains comparable to a designated role model through a formal chain safety mechanism. It employs a dynamic belief update strategy to adapt model selection based on per event predictions and shifting data distributions, and extends to scenarios with dependent models such as earlyexit DNNs and stacking ensembles. Evaluation across multiple datasets demonstrates that NOMAD achieves significant computational savings compared to static and naive approaches while maintaining classification quality comparable to that achieved by the most accurate (and often the most expensive) model.

cs.DB↗

arXiv2025

Meaningful Data Erasure in the Presence of Dependencies

Data regulations like GDPR require systems to support data erasure but leave the definition of "erasure" open to interpretation. This ambiguity makes compliance challenging, especially in databases where data dependencies can lead to erased data being inferred from remaining data. We formally define a precise notion of data erasure that ensures any inference about deleted data, through dependencies, remains bounded to what could have been inferred before its insertion. We design erasure mechanisms that enforce this guarantee at minimal cost. Additionally, we explore strategies to balance cost and throughput, batch multiple erasures, and proactively compute data retention times when possible. We demonstrate the practicality and scalability of our algorithms using both real and synthetic datasets.

cs.DB↗

arXiv2025

DIM-SUM: Dynamic IMputation for Smart Utility Management

Time series imputation models have traditionally been developed using complete datasets with artificial masking patterns to simulate missing values. However, in real-world infrastructure monitoring, practitioners often encounter datasets where large amounts of data are missing and follow complex, heterogeneous patterns. We introduce DIM-SUM, a preprocessing framework for training robust imputation models that bridges the gap between artificially masked training data and real missing patterns. DIM-SUM combines pattern clustering and adaptive masking strategies with theoretical learning guarantees to handle diverse missing patterns actually observed in the data. Through extensive experiments on over 2 billion readings from California water districts, electricity datasets, and benchmarks, we demonstrate that DIM-SUM outperforms traditional methods by reaching similar accuracy with lower processing time and significantly less training data. When compared against a large pre-trained model, DIM-SUM averages 2x higher accuracy with significantly less inference time.

cs.LG↗

arXiv2025

In-context Clustering-based Entity Resolution with Large Language Models: A Design Space Exploration

Entity Resolution (ER) is a fundamental data quality improvement task that identifies and links records referring to the same real-world entity. Traditional ER approaches often rely on pairwise comparisons, which can be costly in terms of time and monetary resources, especially with large datasets. Recently, Large Language Models (LLMs) have shown promising results in ER tasks. However, existing methods typically focus on pairwise matching, missing the potential of LLMs to perform clustering directly in a more cost-effective and scalable manner. In this paper, we propose a novel in-context clustering approach for ER, where LLMs are used to cluster records directly, reducing both time complexity and monetary costs. We systematically investigate the design space for in-context clustering, analyzing the impact of factors such as set size, diversity, variation, and ordering of records on clustering performance. Based on these insights, we develop LLM-CER (LLM-powered Clustering-based ER), which achieves high-quality ER results while minimizing LLM API calls. Our approach addresses key challenges, including efficient cluster merging and LLM hallucination, providing a scalable and effective solution for ER. Extensive experiments on nine real-world datasets demonstrate that our method significantly improves result quality, achieving up to 150% higher accuracy, 10% increase in the F-measure, and reducing API calls by up to 5 times, while maintaining comparable monetary cost to the most cost-effective baseline.

cs.DB↗

arXiv2024

SES: Bridging the Gap Between Explainability and Prediction of Graph Neural Networks

Despite the Graph Neural Networks' (GNNs) proficiency in analyzing graph data, achieving high-accuracy and interpretable predictions remains challenging. Existing GNN interpreters typically provide post-hoc explanations disjointed from GNNs' predictions, resulting in misrepresentations. Self-explainable GNNs offer built-in explanations during the training process. However, they cannot exploit the explanatory outcomes to augment prediction performance, and they fail to provide high-quality explanations of node features and require additional processes to generate explainable subgraphs, which is costly. To address the aforementioned limitations, we propose a self-explained and self-supervised graph neural network (SES) to bridge the gap between explainability and prediction. SES comprises two processes: explainable training and enhanced predictive learning. During explainable training, SES employs a global mask generator co-trained with a graph encoder and directly produces crucial structure and feature masks, reducing time consumption and providing node feature and subgraph explanations. In the enhanced predictive learning phase, mask-based positive-negative pairs are constructed utilizing the explanations to compute a triplet loss and enhance the node representations by contrastive learning.

cs.LG↗

arXiv2024

Graph Structure Prompt Learning: A Novel Methodology to Improve Performance of Graph Neural Networks

Graph neural networks (GNNs) are widely applied in graph data modeling. However, existing GNNs are often trained in a task-driven manner that fails to fully capture the intrinsic nature of the graph structure, resulting in sub-optimal node and graph representations. To address this limitation, we propose a novel Graph structure Prompt Learning method (GPL) to enhance the training of GNNs, which is inspired by prompt mechanisms in natural language processing. GPL employs task-independent graph structure losses to encourage GNNs to learn intrinsic graph characteristics while simultaneously solving downstream tasks, producing higher-quality node and graph representations. In extensive experiments on eleven real-world datasets, after being trained by GPL, GNNs significantly outperform their original performance on node classification, graph classification, and edge prediction tasks (up to 10.28%, 16.5%, and 24.15%, respectively). By allowing GNNs to capture the inherent structural prompts of graphs in GPL, they can alleviate the issue of over-smooth and achieve new state-of-the-art performances, which introduces a novel and effective direction for GNN research with potential applications in various domains.

cs.LG↗

arXiv2024

ProBE: Proportioning Privacy Budget for Complex Exploratory Decision Support

This paper studies privacy in the context of complex decision support queries composed of multiple conditions on different aggregate statistics combined using disjunction and conjunction operators. Utility requirements for such queries necessitate the need for private mechanisms that guarantee a bound on the false negative and false positive errors. This paper formally defines complex decision support queries and their accuracy requirements, and provides algorithms that proportion the existing budget to optimally minimize privacy loss while supporting a bounded guarantee on the accuracy. Our experimental results on multiple real-life datasets show that our algorithms successfully maintain such utility guarantees, while also minimizing privacy loss.

cs.DB↗

arXiv2024

Hiding Access-pattern is Not Enough! Veil: A Storage and Communication Efficient Volume-Hiding Algorithm

This paper addresses volume leakage (i.e., leakage of the number of records in the answer set) when processing keyword queries in encrypted key-value (KV) datasets. Volume leakage, coupled with prior knowledge about data distribution and/or previously executed queries, can reveal both ciphertexts and current user queries. We develop a solution to prevent volume leakage, entitled Veil, that partitions the dataset by randomly mapping keys to a set of equi-sized buckets. Veil provides a tunable mechanism for data owners to explore a trade-off between storage and communication overheads. To make buckets indistinguishable to the adversary, Veil uses a novel padding strategy that allow buckets to overlap, reducing the need to add fake records. Both theoretical and experimental results show Veil to significantly outperform existing state-of-the-art.

cs.DB↗

arXiv2023

Preventing Inferences through Data Dependencies on Sensitive Data

Simply restricting the computation to non-sensitive part of the data may lead to inferences on sensitive data through data dependencies. Inference control from data dependencies has been studied in the prior work. However, existing solutions either detect and deny queries which may lead to leakage -- resulting in poor utility, or only protects against exact reconstruction of the sensitive data -- resulting in poor security. In this paper, we present a novel security model called full deniability. Under this stronger security model, any information inferred about sensitive data from non-sensitive data is considered as a leakage. We describe algorithms for efficiently implementing full deniability on a given database instance with a set of data dependencies and sensitive cells. Using experiments on two different datasets, we demonstrate that our approach protects against realistic adversaries while hiding only minimal number of additional non-sensitive cells and scales well with database size and sensitive data.

cs.DB↗

arXiv2023

Draft & Verify: Lossless Large Language Model Acceleration via Self-Speculative Decoding

We present a novel inference scheme, self-speculative decoding, for accelerating Large Language Models (LLMs) without the need for an auxiliary model. This approach is characterized by a two-stage process: drafting and verification. The drafting stage generates draft tokens at a slightly lower quality but more quickly, which is achieved by selectively skipping certain intermediate layers during drafting. Subsequently, the verification stage employs the original LLM to validate those draft output tokens in one forward pass. This process ensures the final output remains identical to that produced by the unaltered LLM. Moreover, the proposed method requires no additional neural network training and no extra memory footprint, making it a plug-and-play and cost-effective solution for inference acceleration. Benchmarks with LLaMA-2 and its variants demonstrated a speedup up to 1.99$\times$.

cs.CL↗