Understanding column semantics in tabular datasets is a fundamental yet costly task that often relies on human interpretation of ambiguous attributes. We propose GaV (Guess and Verification), a multi-agent architecture that automates this process through an iterative hypothesis generation-and-verification paradigm. A Hypothesis Generator formulates candidate semantic statements for each column with respect to a user-specified semantic aspect (from a configurable aspect inventory), while a Verifier empirically validates them against the data. This mechanism grounds large language model (LLM) reasoning in empirical evidence, yielding verifiable and interpretable semantic profiles. We formalize the Aspect-Based Column Understanding (ABCU) task and introduce a benchmark comprising 46 datasets. Experimental results show that GaV attains up to 88.2% accuracy, effectively balancing accuracy and cost. By coupling LLM-agents with data-grounded validation, GaV establishes a scalable and transparent framework for semantic data understanding.

The idea of using an ontology to enrich user queries with domain knowledge has attracted considerable attention from the database and KR communities during the last fifteen years or so. The ontology and the user query can be conveniently seen as two components of one composite query, called ontology-mediated query (omq), while an omq language (OL, QL) collects all such omqs where the ontology is expressed using the ontology language OL and the user query comes from the query language QL. The evaluation problem for rule-based omq languages of the form (OL, CQ), where OL is a rule-based ontology language, i.e., it collects ontologies modelled using tuple-generating dependencies (a.k.a. existential rules), and CQ is the language of conjunctive queries, has been extensively studied in the literature. In particular, the notion of first-order rewritability of such languages, i.e., the property of being able to rewrite every omq from the language in question to an equivalent first-order query, has been studied in depth. This research effort led an algorithmic characterization of when a rule-based omq language (OL, CQ) is first-order rewritable. More precisely, there is a uniform algorithm Rewrite such that, for every rule-based ontology language OL, the omq language (OL, CQ) is first-order rewritable iff for every omq O from (OL, CQ), the algorithm Rewrite on input O terminates and constructs a first-order rewriting of O. The question that we are interested in is whether the above algorithmic characterization can be extended to rule-based omq languages of the form (OL, CQ(sic)), where CQ (sic) is the language of conjunctive queries with the useful feature of negation. The goal of this work is to initiate effort towards the settlement of the above highly non-trivial question. To this end, we provide a new algorithm, which is a non-trivial extension of the algorithm Rewrite for positive omqs, and show the following: under the Skolem semantics, a well-established approach for defining the answer to a rule-based omq when the user query can use negation, the proposed algorithm is a first-order rewriter for (OL, CQ (sic)), where OL is the language of linear or acyclic tuple-generating dependencies, two central rule-based ontology languages that ensure first-order rewritability for positive omqs. We strongly believe that the new algorithm can serve as a good starting point towards the full settlement of our main question.

This paper explores Machine Unlearning (MU), an emerging field that is gaining increased attention due to concerns about neural models unintentionally remembering personal or sensitive information. We present SeUL, a novel method that enables selective and fine-grained unlearning for language models. Unlike previous work that employs a fully reversed training objective in unlearning, SeUL minimizes the negative impact on the capability of language models, particularly in terms of generation. Furthermore, we introduce two innovative evaluation metrics, sensitive extraction likelihood (S-EL) and sensitive memorization accuracy (S-MA), specifically designed to assess the effectiveness of forgetting sensitive information. In support of the unlearning framework, we propose efficient automatic online and offline sensitive span annotation methods. The online selection method, based on language probability scores, ensures computational efficiency, while the offline annotation involves a two-stage LLM-based process for robust verification. In summary, this paper contributes a novel selective unlearning method (SeUL), introduces specialized evaluation metrics (S-EL and S-MA) for assessing sensitive information forgetting, and proposes automatic online and offline sensitive span annotation methods to support the overall unlearning framework and evaluation.

Query optimization has played a central role in database research for decades. However, more often than not, the proposed optimization techniques lead to a performance improvement in some, but not in all, situations. Therefore, we urgently need a methodology for designing a decision procedure that decides for a given query whether the optimization technique should be applied or not. In this work, we propose such a methodology with a focus on Yannakakis-style query evaluation as our optimization technique of interest. More specifically, we formulate this decision problem as an algorithm selection problem and we present a Machine Learning based approach for its solution. Empirical results with several benchmarks on a variety of database systems show that our approach indeed leads to a statistically significant performance improvement.

Hypertree decompositions provide a way to evaluate Conjunctive Queries (CQs) in polynomial time, where the exponent of this polynomial is determined by the width of the decomposition. In theory, the goal of efficient CQ evaluation therefore has to be a minimisation of the width. However, in practical settings, it turns out that there are also other properties of a decomposition that influence the performance of query evaluation. It is therefore of interest to restrict the computation of decompositions by constraints and to guide this computation by preferences. To this end, we propose a novel framework based on candidate tree decompositions, which allows us to introduce soft hypertree width (shw). This width measure is a relaxation of hypertree width (hw); it is never greater than hw and, in some cases, shw may actually be lower than hw. Most importantly, shw preserves the tractability of deciding if a given CQ is below some fixed bound, while offering more algorithmic flexibility. In particular, it provides a natural way to incorporate preferences and constraints into the computation of decompositions. A prototype implementation and preliminary experiments confirm that this novel framework can indeed have a practical impact on query evaluation.

This invited talk first delves into the division between the two primary branches of AI research: symbolic AI, which predominantly focuses on knowledge representation and logical reasoning, and sub-symbolic AI, primarily centered on machine learning employing neural networks. We explore both the notable accomplishments and the challenges encountered in each of these approaches. We provide instances where traditional deep learning encounters limitations, and we elucidate significant obstacles in achieving automated symbolic reasoning. We then discuss the recent groundbreaking advancements in generative AI, driven by language models such as ChatGPT. We showcase instances where these models excel and, conversely, where they exhibit shortcomings and produce erroneous information. We identify and illustrate five key reasons for potential failures in language models, which include: (i) information loss due to data compression, (ii) training bias, (iii) the incorporation of incorrect external data, (iv) the misordering of results, and (v) the failure to detect and resolve logical inconsistencies contained in a sequence of LLM-generated prompt-answers. Lastly, we touch upon the Chat2Data project, which endeavors to leverage language models for the automated verification and enhancement of relational databases, all while mitigating the pitfalls (i)-(v) mentioned earlier.

Existential rules form an expressive ${{\textsf{Datalog}}}$ -based language to specify ontological knowledge. The presence of existential quantification in rule-heads, however, makes the main reasoning tasks undecidable. To overcome this limitation, in the last two decades, a number of classes of existential rules guaranteeing the decidability of query answering have been proposed. Unfortunately, only some of these classes fully encompass ${{\textsf{Datalog}}}$ and, often, this comes at the price of higher computational complexity. Moreover, expressive classes are typically unable to exploit tools developed for classes exhibiting lower expressiveness. To mitigate these shortcomings, this paper introduces a novel general syntactic condition that allows us to define, systematically and in a uniform way, from any decidable class $\mathcal{C}$ of existential rules, a new class called ${{\textsf{Dyadic-}\mathcal{C}}}$ enjoying the following properties: ( i ) it is decidable; ( ii ) it generalizes ${{\textsf{Datalog}}}$ ; ( iii ) it generalizes $\mathcal{C}$ ; ( iv ) it can effectively exploit any reasoner for query answering over $\mathcal{C}$ ; and ( v ) its computational complexity does not exceed the highest between the one of $\mathcal{C}$ and the one of ${{\textsf{Datalog}}}$ .

One of the main challenges in the area of Neuro-Symbolic AI is to perform logical reasoning in the presence of both neural and symbolic data. This requires combining heterogeneous data sources such as knowledge graphs, neural model predictions, structured databases, crowd-sourced data, and many more. To allow for such reasoning, we generalise the standard rule-based language Datalog with existential rules (commonly referred to as tuple-generating dependencies) to the fuzzy setting, by allowing for arbitrary t-norms in the place of classical conjunctions in rule bodies. The resulting formalism allows us to perform reasoning about data associated with degrees of uncertainty while preserving computational complexity results and the applicability of reasoning techniques established for the standard Datalog setting. In particular, we provide fuzzy extensions of Datalog chases which produce fuzzy universal models and we exploit them to show that in important fragments of the language, reasoning has the same complexity as in the classical setting.

One of the main challenges in the area of Neuro-Symbolic AI is to perform logical reasoning in the presence of both neural and symbolic data. This requires combining heterogeneous data sources such as knowledge graphs, neural model predictions, structured databases, crowd-sourced data, and many more. To allow for such reasoning, we generalise the standard rule-based language Datalog with existential rules (commonly referred to as tuple-generating dependencies) to the fuzzy setting, by allowing for arbitrary t-norms in the place of classical conjunctions in rule bodies. The resulting formalism allows us to perform reasoning about data associated with degrees of uncertainty while preserving computational complexity results and the applicability of reasoning techniques established for the standard Datalog setting. In particular, we provide fuzzy extensions of Datalog chases which produce fuzzy universal models and we exploit them to show that in important fragments of the language, reasoning has the same complexity as in the classical setting.

Various classic reasoning problems with natural hypergraph representations are known to be tractable if a hypertree decomposition (HD) of low width exists. The resulting algorithms are attractive for practical use in fields like databases and constraint satisfaction. However, algorithmic use of HDs relies on the difficult task of first computing a decomposition of the hypergraph underlying a given problem instance, which is then used to guide the algorithm for this particular instance. The performance of purely sequential methods for computing HDs is inherently limited, yet the problem is, theoretically, amenable to parallelisation. In this article, we propose the first algorithm for computing hypertree decompositions that is well suited for parallelisation. The newly proposed algorithm log- k -decomp requires only a logarithmic number of recursion levels and additionally allows for highly parallelised pruning of the search space by restriction to so-called balanced separators. We provide a detailed experimental evaluation over the HyperBench benchmark and demonstrate that log- k -decomp outperforms the current state of the art significantly.

Join queries involving many relations pose a severe challenge to today's query optimisation techniques. To some extent, this is due to the fact that these techniques do not pay sufficient attention to structural properties of the query. In stark contrast, the Database Theory community has intensively studied structural properties of queries (such as acyclicity and various notions of width) and proposed efficient query evaluation techniques through variants of Yannakakis' algorithm. However, although most queries in practice actually are acyclic or have low width, structure-guided query evaluation techniques based on Yannakakis' algorithm have not found their way into mainstream database technology yet. The goal of this work is to address this gap between theory and practice and to demonstrate that the consideration of query structure can improve query evaluation performance on modern DBMSs significantly in cases that have been traditionally challenging. In particular, we study the performance of structure-guided query evaluation in three architecturally distinct DBMSs by rewriting SQL queries into a sequence of SQL statements that express an execution of Yannakakis' algorithm. Moreover, we identify a class of queries that is particularly well suited for our approach and allows query answering in a variety of common scenarios without materializing any join. Through empirical evaluation we show that structure-guided query evaluation can make the evaluation of many difficult join queries feasible whereas their evaluation requires a prohibitive amount of time and memory on current DBMSs.

Over the past decade, Knowledge Graphs have received enormous interest both from industry and from academia. Research in this area has been driven, above all, by the Database (DB) community and the Semantic Web (SW) community. However, there still remains a certain divide between approaches coming from these two communities. For instance, while languages such as SQL or Datalog are widely used in the DB area, a different set of languages such as SPARQL and OWL is used in the SW area. Interoperability between such technologies is still a challenge. The goal of this work is to present a uniform and consistent framework meeting important requirements from both, the SW and DB field.

Fractional (hyper-)graph theory is concerned with the specific problems that arise when fractional analogues of otherwise integer-valued (hyper-)graph invariants are considered. The focus of this paper is on fractional edge covers of hypergraphs. Our main technical result generalizes and unifies previous conditions under which the size of the support of fractional edge covers is bounded independently of the size of the hypergraph itself. We show how this combinatorial result can be used to extend previous tractability results for checking if the fractional hypertree width of a given hypergraph is ≤k for some constant k. Moreover, we show a dual version of our main result for fractional hitting sets.

Modern data processing applications often combine information from a variety of complex sources. Oftentimes, some of these sources, like Machine-Learning systems or crowd-sourced data, are not strictly binary but associated with some degree of confidence in the observation. Ideally, reasoning over such data should take this additional information into account as much as possible. To this end, we propose extensions of Datalog and Datalog+/- to the semantics of Lukasiewicz logic Ł, one of the most common fuzzy logics. We show that such an extension preserves important properties from the classical case and how these properties can lead to efficient reasoning procedures for these new languages.

We consider the scenario of ontology-based data access where a conjunctive query is evaluated against a database enriched with intensional knowledge via an ontology. It is generally accepted that true scalability of query answering in this setting can only be achieved by using standard relational database management systems (RDBMSs). An approach to query answering that enables the use of RDBMSs is the so-called polynomial combined approach. We investigate this approach for the main guarded- and sticky-based classes of existential rules, and we highlight the assumptions on the underlying schema which are sufficient for the polynomial combined first-order rewritability of those classes. To the best of our knowledge, this is the first work which explicitly studies the polynomial combined approach for existential rules.

Join queries involving many relations pose a severe challenge to today’s query optimisation techniques. To some extent, this is due to the fact that these techniques do not pay sufficient attention to structural properties of the query. In stark contrast, the Database Theory community has intensively studied structural properties of queries (such as acyclicity and various notions of width) and proposed efficient query evaluation techniques through variants of Yannakakis’ algorithm for many years. However, although most queries in practice actually are acyclic or have low width, structure-guided query evaluation techniques based on Yannakakis’ algorithm have not found their way into mainstream database technology yet. The goal of this work is to address this gap between theory and practice. We want to analyse the potential of considering the query structure for speeding up modern DBMSs in cases that have been traditionally challenging. To this end, we propose a rewriting of SQL queries into a sequence of SQL statements that force the DBMS to follow a Yannakakis-style query execution. Through first empirical results we show that structure-guided query evaluation can indeed make the evaluation of many difficult join queries significantly faster.

Competitor data constitutes information significantly valuable for many business applications. Meltwater provides users with access to a large Company Information System (CIS), Owler, which contains competitor pairs and other useful information about companies. Meltwater has been seeking a practical solution to discover more competitor pairs in Owler. The first attempt, a fully-manual workflow (called MW_Manual) for finding more competitor pairs in Owler consisted of two manual steps: a filtering step that excludes obvious non-competitor company pairs, and a further inspection process that inspects each left company pair after the filtering step. MW_Manual was cost prohibitive because the results of the filtering step contained too many non-competitor pairs. Inspecting such non-competitor pairs caused an overhead to the overall workload. To reduce the manual workload, especially the required human effort in the manual inspection process, Meltwater has transformed MW_Manual into a semi-automatic workflow (called MW_CPFilter) by replacing the manual filtering with an automatic yet more precise process that adopts a system called CPFilter. This paper presents CPFilter, a system used in the filtering process of MW_CPFilter. CPFilter automatically pre-computes likely competitor pairs from existing competitor pairs in Owler. CPFilter combines (i) the generation of new competitor candidate pairs by inference from existing competitors and other company-specific knowledge, with (ii) the validation of each candidate competitor pair of two companies by checking whether or not empirical evidence that indicates the competitor relationships of these two companies can be found. CPFilter has three key advantages compared with the manual filtering process and previous works: (i) it resulted in a high workload reduction rate of 0.81, (ii) it is domain-independent so that it can be applied to different sectors in Owler, and (iii) its results are explainable so that humans can easily understand its results.