arXiv subjects

Yeasir Rayhan

Publications and source records attributed to Yeasir Rayhan.

13 recordsLinked to original sources

arXiv2026

iFVS: Towards Instance-Optimized Filtered Vector Search

Filtered vector search (FVS) is increasingly important in modern AI + DB systems, where vector similarity search is combined with relational predicates. Quantization plays a vital role in these systems by enabling query processing over large vector datasets. However, lossy approaches, e.g., Product Quantization (PQ), incur a precision penalty during distance calculation, thereby negatively impacting the query recall performance. This problem becomes more challenging in FVS because the relevant vector space can change with the relational predicate and selectivity. Motivated by the success of instance-optimized database system components, we introduce iFVS, an Instance-Optimized Filtered Vector Search technique. Given a fixed, quantized vector dataset, and a representative workload of filtered vector queries, iFVS adopts a query-specific codebook generation approach for FVS that is instance-optimized towards a certain dataset and query workload. Instead of using a fixed codebook for all queries, iFVS conditions distance estimation on both the query vector and the filter predicate. This enables more accurate ranking over compressed vectors while preserving compact per-vector storage. Experiments show that iFVS improves the Queries Per Second (QPS)-recall tradeoff across several filter selectivity bins compared with fixed-codebook quantized FVS baselines.

cs.DB↗

arXiv2026

Virtual-Memory Assisted Buffer Management In Tiered Memory

Tiered memory architectures have gained significant traction in the database community in recent years. In these architectures, the on-chip DRAM of the host processor is typically referred to as local memory, and forms the primary tier. Additional byte-addressable, cache-coherent memory resources, collectively referred to as remote memory (RMem, for short), form one or more secondary tiers. RMem is slower than local DRAM but faster than disk, e.g., NUMA memory located on a remote socket, chiplet-attached memory, and memory attached via high-performance interconnect protocols, e.g., RDMA and CXL. In this paper, we discuss how traditional two-tier (DRAM-Disk) virtual-memory assisted Buffer Management techniques generalize to an $n$-tier setting (DRAM-RMem-Disk). We present vmcache$^n$, an $n$-tier virtual-memory-assisted buffer pool that leverages the virtual memory subsystem and operating system calls to migrate pages across memory tiers. In this setup, page migration can become a bottleneck. To address this limitation, we introduce the move_pages2 system call that provides vmcache$^n$ with fine-grained control over the page migration process. Experiments show that vmcache$^n$ can achieve up to 4$\times$ higher query throughput over vmcache for TPC-C workloads.

cs.DB↗

arXiv2026

An Empirical Survey and Benchmark of Learned Distance Indexes for Road Networks

The calculation of shortest-path distances in road networks is a core operation in navigation systems, location-based services, and spatial analytics. Although classical algorithms, e.g., Dijkstra's algorithm, provide exact answers, their latency is prohibitive for modern real-time, large-scale deployments. Over the past two decades, numerous distance indexes have been proposed to speed up query processing for shortest distance queries. More recently, with the advancement in machine learning (ML), researchers have designed and proposed ML-based distance indexes to answer approximate shortest path and distance queries efficiently. However, a comprehensive and systematic evaluation of these ML-based approaches is lacking. This paper presents the first empirical survey of ML-based distance indexes on road networks, evaluating them along four key dimensions: Training time, query latency, storage, and accuracy. Using seven real-world road networks and workload-driven query datasets derived from trajectory data, we benchmark ten representative ML techniques and compare them against strong classical non-ML baselines, highlighting key insights and practical trade-offs. We release a unified open-source codebase to support reproducibility and future research on learned distance indexes.

cs.LG↗

arXiv2026

Gen-DBA: Generative Database Agents

Leveraging Machine Learning to optimize database systems, referred to as Machine Learning for Databases (ML4DB, for short), dates back to the early 1990s, spanning indexing techniques, selectivity estimation, and query optimization. However, the idea has gained mainstream traction following the introduction of learned indexes in 2018, triggering a surge of research spanning learned indexes and cardinality estimators to learned query optimizers, storage layout design, resource management, and database tuning. The current ML4DB optimization landscape is dominated by narrow specialist ML models that are small and are trained on limited training data. Each specialist ML model targets a single database learning task on a fixed database engine, hardware platform, query workload, and optimization objective. As a result, they fall short in real-world settings, where these factors can vary significantly and evolve over time. This leads to an exponential number of ML models with limited portability and generalization capability, thus limiting the utility of existing ML4DB approaches. We address this limitation with Gen-DBA, a single general-purpose foundation model for optimizing databases with agentic capabilities. This paper presents the vision for Gen-DBA, provides a sketch design of how to realize it, and highlights several research challenges that need to be addressed to fully realize Gen-DBA.

cs.DB↗

arXiv2025

Lightning Prediction under Uncertainty: DeepLight with Hazy Loss

Lightning, a common feature of severe meteorological conditions, poses significant risks, from direct human injuries to substantial economic losses. These risks are further exacerbated by climate change. Early and accurate prediction of lightning would enable preventive measures to safeguard people, protect property, and minimize economic losses. In this paper, we present DeepLight, a novel deep learning architecture for predicting lightning occurrences. Existing prediction models face several critical limitations: i) they often struggle to capture the dynamic spatial context and the inherent randomness of lightning events, including whether lightning occurs and its variability in location and timing even under similar meteorological conditions; ii) they underutilize key observational data, such as radar reflectivity and cloud properties; and iii) they rely heavily on Numerical Weather Prediction (NWP) systems, which are both computationally expensive and highly sensitive to parameter settings. To overcome these challenges, DeepLight leverages multi-source meteorological data, including radar reflectivity, cloud properties, and historical lightning occurrences through a dual-encoder architecture. By employing multi-branch convolution techniques, it dynamically captures spatial correlations across varying extents. Furthermore, its novel Hazy Loss function explicitly addresses the spatio-temporal uncertainty of lightning by penalizing deviations based on proximity to true events, enabling the model to better learn patterns amidst randomness. Extensive experiments show that DeepLight improves the Equitable Threat Score (ETS) by 18\%--30\% over state-of-the-art methods, establishing it as a robust solution for lightning prediction.

cs.LG↗

arXiv2025

Exploring Next Token Prediction For Optimizing Databases

The Next Token Prediction paradigm (NTP, for short) lies at the forefront of modern large foundational models that are pre-trained on diverse and large datasets. These models generalize effectively, and have proven to be very successful in Natural Language Processing (NLP). Inspired by the generalization capabilities of Large Language Models (LLMs), we investigate whether the same NTP paradigm can be applied to DBMS design and optimization tasks. Adopting NTP directly for database optimization is non-trivial due to the fundamental differences between the domains. In this paper, we present a framework, termed Probe and Learn (PoLe), for applying NTP to optimize database systems. PoLe leverages Decision Transformers and hardware-generated tokens to effectively incorporate NTP into database systems. As a proof of concept, we demonstrate PoLe in the context of the index scheduling task over NUMA servers in main-memory database systems. Preliminary results for this scheduling task demonstrate that adopting NTP and PoLe can improve both performance and generalizability.

cs.DB↗

arXiv2025

Revisiting Page Migration for Main-Memory Database Systems

Modern hardware architectures, e.g., NUMA servers, chiplet processors, tiered and disaggregated memory systems have significantly improved the performance of Main-Memory Databases, and are poised to deliver further improvements in the future. However, realizing this potential depends on the database system's ability to efficiently migrate pages among different NUMA nodes, and/or memory chips as the workload evolves. Modern main memory databases offload the migration procedure to the operating system without accounting for the workload and its migration characteristics. In this paper, we propose a custom system call move_pages2 as an alternate to Linux's own move_pages system call. In contrast to the original move_pages, move_pages2 allows partial migration and exposes two configuration knobs, enabling a Main-Memory Database tailor the migration process to its specific requirements. Experiments on a main-memory B$^+$-Tree for a YCSB-like workload show that the proposed move_pages2 custom system call improves the B$^+$-Tree query throughput by up to 2.3$\times$, and migrates up to 2.6$\times$ more memory pages, outperforming the native Linux system call.

cs.DB↗

arXiv2024

P-MOSS: Scheduling Main-Memory Indexes Over NUMA Servers Using Next Token Prediction

Ever since the Dennard scaling broke down in the early 2000s and the frequency of the CPUs stalled, vendors have started to increase the core count in each CPU chip at the expense of introducing heterogeneity, thus ushering the era of NUMA and Chiplet processors. Since then, the heterogeneity in the design space of hardware has only increased to the point that DBMS performance may vary significantly up to an order of magnitude in modern servers. An important factor that affects performance includes the location of the logical cores where the DBMS queries execute, and the location where the data resides. This paper introduces P-MOSS, a learned spatial scheduling framework that schedules query execution to specific logical cores, and co-locates data on the corresponding NUMA node. For cross-hardware and workload adaptability, P-MOSS leverages core principles from Large Language Models, such as Next Token prediction, Generative Pre-training, and Fine-tuning. In the spirit of hardware-software synergy, P-MOSS guides its scheduling decision solely based on the low-level hardware statistics collected from the hardware Performance Monitoring Unit with the aid of a Decision Transformer. Experimental evaluation is performed in the context of the B$^+$-Tree index. Performance results demonstrate that P-MOSS offers an improvement of up to $6\times$ over traditional schedules in terms of query throughput.

cs.DB↗

arXiv2024

Deep Learning Based Crime Prediction Models: Experiments and Analysis

Crime prediction is a widely studied research problem due to its importance in ensuring safety of city dwellers. Starting from statistical and classical machine learning based crime prediction methods, in recent years researchers have focused on exploiting deep learning based models for crime prediction. Deep learning based crime prediction models use complex architectures to capture the latent features in the crime data, and outperform the statistical and classical machine learning based crime prediction methods. However, there is a significant research gap in existing research on the applicability of different models in different real-life scenarios as no longitudinal study exists comparing all these approaches in a unified setting. In this paper, we conduct a comprehensive experimental evaluation of all major state-of-the-art deep learning based crime prediction models. Our evaluation provides several key insights on the pros and cons of these models, which enables us to select the most suitable models for different application scenarios. Based on the findings, we further recommend certain design practices that should be taken into account while building future deep learning based crime prediction models.

cs.LG↗

arXiv2024

GTX: A Write-Optimized Latch-free Graph Data System with Transactional Support -- Extended Version

This paper introduces GTX, a standalone main-memory write-optimized graph data system that specializes in structural and graph property updates while enabling concurrent reads and graph analytics through ACID transactions. Recent graph systems target concurrent read and write support while guaranteeing transaction semantics. However, their performance suffers from updates with real-world temporal locality over the same vertices and edges due to vertex-centric lock contentions. GTX has an adaptive delta-chain locking protocol on top of a carefully designed latch-free graph storage. It eliminates vertex-level locking contention, and adapts to real-life workloads while maintaining sequential access to the graph's adjacency lists storage. GTX's transactions further support cache-friendly block level concurrency control, and cooperative group commit and garbage collection. This combination of features ensures high update throughput and provides low-latency graph analytics. Based on experimental evaluation, in addition to not sacrificing the performance of read-heavy analytical workloads, and having competitive performance similar to state-of-the-art systems, GTX has high read-write transaction throughput. For write-heavy transactional workloads, GTX achieves up to 11x better transaction throughput than the best-performing state-of-the-art system.

cs.DB↗

arXiv2023

SIMD-ified R-tree Query Processing and Optimization

The introduction of Single Instruction Multiple Data (SIMD) instructions in mainstream CPUs has enabled modern database engines to leverage data parallelism by performing more computation with a single instruction, resulting in a reduced number of instructions required to execute a query as well as the elimination of conditional branches. Though SIMD in the context of traditional database engines has been studied extensively, it has been overlooked in the context of spatial databases. In this paper, we investigate how spatial database engines can benefit from SIMD vectorization in the context of an R-tree spatial index. We present vectorized versions of the spatial range select, and spatial join operations over a vectorized R-tree index. For each of the operations, we investigate two storage layouts for an R-tree node to leverage SIMD instructions. We design vectorized algorithms for each of the spatial operations given each of the two data layouts. We show that the introduction of SIMD can improve the latency of the spatial query operators up to 9x. We introduce several optimizations over the vectorized implementation of these query operators, and study their effectiveness in query performance and various hardware performance counters under different scenarios.

cs.DB↗

arXiv2022

ILX: Intelligent "Location+X" Data Systems (Vision Paper)

Due to the ubiquity of mobile phones and location-detection devices, location data is being generated in very large volumes. Queries and operations that are performed on location data warrant the use of database systems. Despite that, location data is being supported in data systems as an afterthought. Typically, relational or NoSQL data systems that are mostly designed with non-location data in mind get extended with spatial or spatiotemporal indexes, some query operators, and higher level syntactic sugar in order to support location data. The ubiquity of location data and location data services call for systems that are solely designed and optimized for the efficient support of location data. This paper envisions designing intelligent location+X data systems, ILX for short, where location is treated as a first-class citizen type. ILX is tailored with location data as the main data type (location-first). Because location data is typically augmented with other data types X, e.g., graphs, text data, click streams, annotations, etc., ILX needs to be extensible to support other data types X along with location. This paper envisions the main features that ILX should support, and highlights research challenges in realizing and supporting ILX.

cs.DB↗

arXiv2020

AIST: An Interpretable Attention-based Deep Learning Model for Crime Prediction

Accuracy and interpretability are two essential properties for a crime prediction model. Because of the adverse effects that the crimes can have on human life, economy and safety, we need a model that can predict future occurrence of crime as accurately as possible so that early steps can be taken to avoid the crime. On the other hand, an interpretable model reveals the reason behind a model's prediction, ensures its transparency and allows us to plan the crime prevention steps accordingly. The key challenge in developing the model is to capture the non-linear spatial dependency and temporal patterns of a specific crime category while keeping the underlying structure of the model interpretable. In this paper, we develop AIST, an Attention-based Interpretable Spatio Temporal Network for crime prediction. AIST models the dynamic spatio-temporal correlations for a crime category based on past crime occurrences, external features (e.g., traffic flow and point of interest (POI) information) and recurring trends of crime. Extensive experiments show the superiority of our model in terms of both accuracy and interpretability using real datasets.

cs.LG↗