arXiv subjects

Hubert Naacke

Publications and source records attributed to Hubert Naacke.

15 recordsLinked to original sources

arXiv2025

Exploring Multi-Table Retrieval Through Iterative Search

Open-domain question answering over datalakes requires retrieving and composing information from multiple tables, a challenging subtask that demands semantic relevance and structural coherence (e.g., joinability). While exact optimization methods like Mixed-Integer Programming (MIP) can ensure coherence, their computational complexity is often prohibitive. Conversely, simpler greedy heuristics that optimize for query coverage alone often fail to find these coherent, joinable sets. This paper frames multi-table retrieval as an iterative search process, arguing this approach offers advantages in scalability, interpretability, and flexibility. We propose a general framework and a concrete instantiation: a fast, effective Greedy Join-Aware Retrieval algorithm that holistically balances relevance, coverage, and joinability. Experiments across 5 NL2SQL benchmarks demonstrate that our iterative method achieves competitive retrieval performance compared to the MIP-based approach while being 4-400x faster depending on the benchmark and search space settings. This work highlights the potential of iterative heuristics for practical, scalable, and composition-aware retrieval.

cs.IR↗

arXiv2025

Something's Fishy In The Data Lake: A Critical Re-evaluation of Table Union Search Benchmarks

Recent table representation learning and data discovery methods tackle table union search (TUS) within data lakes, which involves identifying tables that can be unioned with a given query table to enrich its content. These methods are commonly evaluated using benchmarks that aim to assess semantic understanding in real-world TUS tasks. However, our analysis of prominent TUS benchmarks reveals several limitations that allow simple baselines to perform surprisingly well, often outperforming more sophisticated approaches. This suggests that current benchmark scores are heavily influenced by dataset-specific characteristics and fail to effectively isolate the gains from semantic understanding. To address this, we propose essential criteria for future benchmarks to enable a more realistic and reliable evaluation of progress in semantic table union search.

cs.IR↗

arXiv2024

TISIS : Trajectory Indexing for SImilarity Search

Social media platforms enable users to share diverse types of information, including geolocation data that captures their movement patterns. Such geolocation data can be leveraged to reconstruct the trajectory of a user's visited Points of Interest (POIs). A key requirement in numerous applications is the ability to measure the similarity between such trajectories, as this facilitates the retrieval of trajectories that are similar to a given reference trajectory. This is the main focus of our work. Existing methods predominantly rely on applying a similarity function to each candidate trajectory to identify those that are sufficiently similar. However, this approach becomes computationally expensive when dealing with large-scale datasets. To mitigate this challenge, we propose TISIS, an efficient method that uses trajectory indexing to quickly find similar trajectories that share common POIs in the same order. Furthermore, to account for scenarios where POIs in trajectories may not exactly match but are contextually similar, we introduce TISIS*, a variant of TISIS that incorporates POI embeddings. This extension allows for more comprehensive retrieval of similar trajectories by considering semantic similarities between POIs, beyond mere exact matches. Extensive experimental evaluations demonstrate that the proposed approach significantly outperforms a baseline method based on the well-known Longest Common SubSequence (LCSS) algorithm, yielding substantial performance improvements across various real-world datasets.

cs.DB↗

arXiv2024

Leiden-Fusion Partitioning Method for Effective Distributed Training of Graph Embeddings

In the area of large-scale training of graph embeddings, effective training frameworks and partitioning methods are critical for handling large networks. However, they face two major challenges: 1) existing synchronized distributed frameworks require continuous communication to access information from other machines, and 2) the inability of current partitioning methods to ensure that subgraphs remain connected components without isolated nodes, which is essential for effective training of GNNs since training relies on information aggregation from neighboring nodes. To address these issues, we introduce a novel partitioning method, named Leiden-Fusion, designed for large-scale training of graphs with minimal communication. Our method extends the Leiden community detection algorithm with a greedy algorithm that merges the smallest communities with highly connected neighboring communities. Our method guarantees that, for an initially connected graph, each partition is a densely connected subgraph with no isolated nodes. After obtaining the partitions, we train a GNN for each partition independently, and finally integrate all embeddings for node classification tasks, which significantly reduces the need for network communication and enhances the efficiency of distributed graph training. We demonstrate the effectiveness of our method through extensive evaluations on several benchmark datasets, achieving high efficiency while preserving the quality of the graph embeddings for node classification tasks.

cs.LG↗

arXiv2023

ANTM: An Aligned Neural Topic Model for Exploring Evolving Topics

This paper presents an algorithmic family of dynamic topic models called Aligned Neural Topic Models (ANTM), which combine novel data mining algorithms to provide a modular framework for discovering evolving topics. ANTM maintains the temporal continuity of evolving topics by extracting time-aware features from documents using advanced pre-trained Large Language Models (LLMs) and employing an overlapping sliding window algorithm for sequential document clustering. This overlapping sliding window algorithm identifies a different number of topics within each time frame and aligns semantically similar document clusters across time periods. This process captures emerging and fading trends across different periods and allows for a more interpretable representation of evolving topics. Experiments on four distinct datasets show that ANTM outperforms probabilistic dynamic topic models in terms of topic coherence and diversity metrics. Moreover, it improves the scalability and flexibility of dynamic topic models by being accessible and adaptable to different types of algorithms. Additionally, a Python package is developed for researchers and scientists who wish to study the trends and evolving patterns of topics in large-scale textual data.

cs.IR↗

arXiv2023

ATEM: A Topic Evolution Model for the Detection of Emerging Topics in Scientific Archives

This paper presents ATEM, a novel framework for studying topic evolution in scientific archives. ATEM is based on dynamic topic modeling and dynamic graph embedding techniques that explore the dynamics of content and citations of documents within a scientific corpus. ATEM explores a new notion of contextual emergence for the discovery of emerging interdisciplinary research topics based on the dynamics of citation links in topic clusters. Our experiments show that ATEM can efficiently detect emerging cross-disciplinary topics within the DBLP archive of over five million computer science articles.

cs.IR↗

arXiv2023

Contextualized Topic Coherence Metrics

The recent explosion in work on neural topic modeling has been criticized for optimizing automated topic evaluation metrics at the expense of actual meaningful topic identification. But human annotation remains expensive and time-consuming. We propose LLM-based methods inspired by standard human topic evaluations, in a family of metrics called Contextualized Topic Coherence (CTC). We evaluate both a fully automated version as well as a semi-automated CTC that allows human-centered evaluation of coherence while maintaining the efficiency of automated methods. We evaluate CTC relative to five other metrics on six topic models and find that it outperforms automated topic coherence methods, works well on short documents, and is not susceptible to meaningless but high-scoring topics.

cs.CL↗

arXiv2020

A Malaria Control Model Using Mobility Data: An Early Explanation of Kedougou's Case in Senegal

Studies in malaria control cover many areas such as medicine, sociology, biology, mathematic, physic, computer science and so forth. Researches in the realm of mathematic are conducted to predict the occurrence of the disease and to support the eradication process. Basically, the modeling methodology is predominantly deterministic and differential equation based while selecting clinical and biological features that seem to be important. Yet, if the individual characteristics matter when modeling the disease, the overall estimation of the malaria is not done based on the health status of each individual but in a non-specified percentage of the global population. The goal of this paper is to propose a model that relies on a daily evolution of the individual's state, which depends on their mobility and the characteristics of the area they visit. Thus, the mobility data of a single person moving from one area to another, gathered thanks to mobile networks, is the essential building block to predict the outcome of the disease. We implement our solution and demonstrate its effectiveness through empirical experiments. The results show how promising the model is in providing possible insights into the failure of the disease control in the Kedougou region.

q-bio.PE↗

arXiv2018

BigSR: an empirical study of real-time expressive RDF stream reasoning on modern Big Data platforms

The trade-off between language expressiveness and system scalability (E&S) is a well-known problem in RDF stream reasoning. Higher expressiveness supports more complex reasoning logic, however, it may also hinder system scalability. Current research mainly focuses on logical frameworks suitable for stream reasoning as well as the implementation and the evaluation of prototype systems. These systems are normally developed in a centralized setting which suffer from inherent limited scalability, while an in-depth study of applying distributed solutions to cover E&S is still missing. In this paper, we aim to explore the feasibility of applying modern distributed computing frameworks to meet E&S all together. To do so, we first propose BigSR, a technical demonstrator that supports a positive fragment of the LARS framework. For the sake of generality and to cover a wide variety of use cases, BigSR relies on the two main execution models adopted by major distributed execution frameworks: Bulk Synchronous Processing (BSP) and Record-at-A-Time (RAT). Accordingly, we implement BigSR on top of Apache Spark Streaming (BSP model) and Apache Flink (RAT model). In order to conclude on the impacts of BSP and RAT on E&S, we analyze the ability of the two models to support distributed stream reasoning and identify several types of use cases characterized by their levels of support. This classification allows for quantifying the E&S trade-off by assessing the scalability of each type of use case \wrt its level of expressiveness. Then, we conduct a series of experiments with 15 queries from 4 different datasets. Our experiments show that BigSR over both BSP and RAT generally scales up to high throughput beyond million-triples per second (with or without recursion), and RAT attains sub-millisecond delay for stateless query operators.

cs.AI↗

arXiv2017

Strider-lsa: Massive RDF Stream Reasoning in the Cloud

Reasoning over semantically annotated data is an emerging trend in stream processing aiming to produce sound and complete answers to a set of continuous queries. It usually comes at the cost of finding a trade-off between data throughput and the cost of expressive inferences. Strider-lsa proposes such a trade-off and combines a scalable RDF stream processing engine with an efficient reasoning system. The main reasoning tasks are based on a query rewriting approach for SPARQL that benefits from an intelligent encoding of RDFS+ (RDFS + owl:sameAs) ontology elements. Strider-lsa runs in production at a major international water management company to detect anomalies from sensor streams. The system is evaluated along different dimensions and over multiple datasets to emphasize its performance.

cs.DB↗

arXiv2016

SPARQL query processing with Apache Spark

The number of linked data sources and the size of the linked open data graph keep growing every day. As a consequence, semantic RDF services are more and more confronted to various "big data" problems. Query processing is one of them and needs to be efficiently addressed with executions over scalable, highly available and fault tolerant frameworks. Data management systems requiring these properties are rarely built from scratch but are rather designed on top of an existing cluster computing engine. In this work, we consider the processing of SPARQL queries with Apache Spark. We propose and compare five different query processing approaches based on different join execution models and Spark components. A detailed experimentation, on real-world and synthetic data sets, emphasizes that two approaches tailored for the RDF data model outperform the other ones on all major query shapes, i.e., star, snowflake, chain and hybrid.

cs.DB↗

arXiv2015

LiteMat: a scalable, cost-efficient inference encoding scheme for large RDF graphs

The number of linked data sources and the size of the linked open data graph keep growing every day. As a consequence, semantic RDF services are more and more confronted with various "big data" problems. Query processing in the presence of inferences is one them. For instance, to complete the answer set of SPARQL queries, RDF database systems evaluate semantic RDFS relationships (subPropertyOf, subClassOf) through time-consuming query rewriting algorithms or space-consuming data materialization solutions. To reduce the memory footprint and ease the exchange of large datasets, these systems generally apply a dictionary approach for compressing triple data sizes by replacing resource identifiers (IRIs), blank nodes and literals with integer values. In this article, we present a structured resource identification scheme using a clever encoding of concepts and property hierarchies for efficiently evaluating the main common RDFS entailment rules while minimizing triple materialization and query rewriting. We will show how this encoding can be computed by a scalable parallel algorithm and directly be implemented over the Apache Spark framework. The efficiency of our encoding scheme is emphasized by an evaluation conducted over both synthetic and real world datasets.

cs.DB↗

arXiv2015

On the Evaluation of RDF Distribution Algorithms Implemented over Apache Spark

Querying very large RDF data sets in an efficient manner requires a sophisticated distribution strategy. Several innovative solutions have recently been proposed for optimizing data distribution with predefined query workloads. This paper presents an in-depth analysis and experimental comparison of five representative and complementary distribution approaches. For achieving fair experimental results, we are using Apache Spark as a common parallel computing framework by rewriting the concerned algorithms using the Spark API. Spark provides guarantees in terms of fault tolerance, high availability and scalability which are essential in such systems. Our different implementations aim to highlight the fundamental implementation-independent characteristics of each approach in terms of data preparation, load balancing, data replication and to some extent to query answering cost and performance. The presented measures are obtained by testing each system on one synthetic and one real-world data set over query workloads with differing characteristics and different partitioning constraints.

cs.DB↗

arXiv2013

Improving tag recommendation by folding in more consistency

Tag recommendation is a major aspect of collaborative tagging systems. It aims to recommend tags to a user for tagging an item. In this paper we present a part of our work in progress which is a novel improvement of recommendations by re-ranking the output of a tag recommender. We mine association rules between candidates tags in order to determine a more consistent list of tags to recommend. Our method is an add-on one which leads to better recommendations as we show in this paper. It is easily parallelizable and morever it may be applied to a lot of tag recommenders. The experiments we did on five datasets with two kinds of tag recommender demonstrated the efficiency of our method.

cs.IR↗

arXiv2012

Dynamic recommender system : using cluster-based biases to improve the accuracy of the predictions

It is today accepted that matrix factorization models allow a high quality of rating prediction in recommender systems. However, a major drawback of matrix factorization is its static nature that results in a progressive declining of the accuracy of the predictions after each factorization. This is due to the fact that the new obtained ratings are not taken into account until a new factorization is computed, which can not be done very often because of the high cost of matrix factorization. In this paper, aiming at improving the accuracy of recommender systems, we propose a cluster-based matrix factorization technique that enables online integration of new ratings. Thus, we significantly enhance the obtained predictions between two matrix factorizations. We use finer-grained user biases by clustering similar items into groups, and allocating in these groups a bias to each user. The experiments we did on large datasets demonstrated the efficiency of our approach.

cs.LG↗