Rihan Hai

• Dr. rer. nat. in Computer Science
• Assistant professor at Delft University of Technology

Generating temporal data under constraints is critical for forecasting, imputation, and synthesis. These datasets often include auxiliary conditions that influence the values within the time series signal. Existing methods face three key challenges: (1) they fail to adapt to conditions at inference time; (2) they rely on sequential generation, whic...

Serving large language models (LLMs) often demands specialized hardware, dedicated frameworks, and substantial development efforts, which restrict their accessibility, especially for edge devices and organizations with limited technical resources. We propose a novel compiler that translates LLM inference graphs into SQL queries, enabling relational...

The machine learning (ML) training over disparate data sources traditionally involves materialization, which can impose substantial time and space overhead due to data movement and replication. Factorized learning, which leverages direct computation on disparate sources through linear algebra (LA) rewriting, has emerged as a viable alternative to i...

The rapid growth of large-scale machine learning (ML) models has led numerous commercial companies to utilize ML models for generating predictive results to help business decision-making. As two primary components in traditional predictive pipelines, data processing, and model predictions often operate in separate execution environments, leading to...

Machine learning (ML) training data is often scattered across disparate collections of datasets, called data silos . This fragmentation poses a major challenge for data-intensive ML applications: integrating and transforming data residing in different sources demand a lot of manual work and computational resources. With data privacy constraints,...

Quantum computing has emerged as a promising tool for transforming the landscape of computing technology. Recent efforts have applied quantum techniques to classical database challenges, such as query optimization, data integration, index selection, and transaction management. In this paper, we shift focus to a critical yet underexplored area: data...

The advent of Large Language Models (LLMs) provides an opportunity to change the way queries are processed, moving beyond the constraints of conventional SQL-based database systems. However, using an LLM to answer a prediction query is still challenging, since an external ML model has to be employed and inference has to be performed in order to pro...

Online Large Language Model (LLM) services such as ChatGPT and Claude 3 have transformed business operations and academic research by effortlessly enabling new opportunities. However, due to data-sharing restrictions, sectors such as healthcare and finance prefer to deploy local LLM applications using costly hardware resources. This scenario requir...

Vertical federated learning (VFL) is a promising area for time series forecasting in industrial applications, such as predictive maintenance and machine control. Critical challenges to address in manufacturing include data privacy and over-fitting on small and noisy datasets during both training and inference. Additionally, to increase industry ada...

Quantum computing has emerged as a transforma-tive tool for future data management. Classical problems in database domains, including query optimization, data integration , and transaction management, have recently been addressed using quantum computing techniques. This tutorial aims to establish the theoretical foundation essential for enhancing m...

Data lakes are becoming increasingly prevalent for big data management and data analytics. In contrast to traditional ‘schema-on-write’ approaches such as data warehouses, data lakes are repositories storing raw data in its original formats and providing a common access interface. Despite the strong interest raised from both academia and industry,...

The proliferation of pre-trained ML models in public Web-based model zoos facilitates the engineering of ML pipelines to address complex inference queries over datasets and streams of unstructured content. Constructing optimal plan for a query is hard, especially when constraints (e.g. accuracy or execution time) must be taken into consideration, a...

Machine learning (ML) researchers and practitioners are building repositories of pre-trained models, called model zoos. These model zoos contain metadata that detail various properties of the ML models and datasets, which are useful for reporting, auditing, reproducibility, and interpretability. Unfortunately, the existing metadata representations...

The rapid growth of large-scale machine learning (ML) models has led numerous commercial companies to utilize ML models for generating predictive results to help business decision-making. As two primary components in traditional predictive pipelines, data processing, and model predictions often operate in separate execution environments, leading to...

The increasing need for data trading has created a high demand for data marketplaces. These marketplaces require a set of value-added services, such as advanced search and discovery, that have been proposed in the database research community for years, but are yet to be put to practice. In this paper we propose to demonstrate the Topio Marketplace,...

Machine learning (ML) practitioners and organizations are building model repositories of pre-trained models, referred to as model zoos . These model zoos contain metadata describing the properties of the ML models and datasets. The metadata serves crucial roles for reporting, auditing, ensuring reproducibility, and enhancing interpretability. Des...

A digital twin (DT), originally defined as a virtual representation of a physical asset, system, or process, is a new concept in health care. A DT in health care is not a single technology but a domain-adapted multimodal modeling approach incorporating the acquisition, management, analysis, prediction, and interpretation of data, aiming to improve...

Machine learning (ML) practitioners and organizations are building model zoos of pre-trained models, containing metadata describing properties of the ML models and datasets that are useful for reporting, auditing, reproducibility, and interpretability purposes. The metatada is currently not standardised; its expressivity is limited; and there is no...

Virtually every sizable organization nowadays is building a form of a data lake. In theory, every department or team in the organization would enrich their datasets with metadata, and store them in a central data lake. Those datasets can then be combined in different ways and produce added value to the organization. In practice, though, the situati...

The data needed for machine learning (ML) model training and inference, can reside in different separate sites often termed data silos. For data-intensive ML applications, data silos present a major challenge: the integration and transformation of data, demand a lot of manual work and computational resources. Sometimes, data cannot leave the local...

UNSTRUCTURED A Digital Twin (DT), which is defined originally as a virtual representation of a physical asset, system or process, is a new concept in healthcare. DT in healthcare cannot be a single technology, but a domain adapted multi-modal modelling approach, which incorporates the acquisition, management, analysis, prediction, and interpretatio...

Although big data has been discussed for some years, it still has many research challenges, especially the variety of data. It poses a huge difficulty to efficiently integrate, access, and query the large volume of diverse data in information silos with the traditional 'schema-on-write' approaches such as data warehouses. Data lakes have been propo...

This chapter introduces the most important features of data lake systems, and from there it outlines an architecture for these systems. The vision for a data lake system is based on a generic and extensible architecture with a unified data model, facilitating the ingestion, storage and metadata management over heterogeneous data sources. The chapte...

Functional dependencies are important for the definition of constraints and relationships that have to be satisfied by every database instance. Relaxed functional dependencies (RFDs) can be used for data exploration and profiling in datasets with lower data quality. In this work, we present an approach for RFD discovery in heterogeneous data lakes....

In this work, we present a Metadata Framework in the direction of extending intelligence mechanisms from the Cloud to the Edge. To this end, we build on our previously introduced notion of Data Lagoons—the analogous to Data Lakes at the network edge—and we introduce a novel architecture and Metadata model for the efficient interaction between Data...

Schema mappings express the relationships between sources in data interoperability scenarios and can be expressed in various formalisms. Source-to-target tuple-generating dependencies (s-t tgds) can be easily used for data transformation or query rewriting tasks. Second-order tgds (SO tgds) are more expressive as they can also represent the composi...

JSON has become one of the most popular data formats. Yet studies on JSON data integration (DI) are scarce. In this work, we study one of the key DI tasks, nested mapping generation in the context of integrating heterogeneous JSON based data sources. We propose a novel mapping representation, namely bucket forest mappings that models the nested map...

The increasing popularity of NoSQL systems has lead to the model of polyglot persistence, in which several data management systems with different data models are used. Data lakes realize the polyglot persistence model by collecting data from various sources, by storing the data in its original structure, and by providing the datasets for querying a...

Medical engineering (ME) is an interdisciplinary domain with short innovation cycles. Usually, researchers from several fields cooperate in ME research projects. To support the identification of suitable partners for a project, we present an integrated approach for patent classification combining ideas from topic modeling, ontology modeling & match...

In addition to volume and velocity, Big data is also characterized by its variety. Variety in structure and semantics requires new integration approaches which can resolve the integration challenges also for large volumes of data. Data lakes should reduce the upfront integration costs and provide a more flexible way for data integration and analysi...

As the challenge of our time, Big Data still has many research hassles, especially the variety of data. The high diversity of data sources often results in information silos, a collection of non-integrated data management systems with heterogeneous schemas, query languages, and APIs. Data Lake systems have been proposed as a solution to this proble...

The heterogeneity of sources in Big Data systems requires new integration approaches which can handle the large volume of the data as well as its variety. Data lakes have been proposed to reduce the upfront integration costs and to provide more flexibility in integrating and analyzing information. In data lakes, data from the sources is copied in i...

Data Warehouses (DW) have to continuously adapt to evolving business requirements, which implies structure modification (schema changes) and data migration requirements in the system design. However, it is challenging for designers to control the performance and cost overhead of different schema change implementations. In this paper, we demonstrate...