Copernicus, the European Union's Earth observation program, produces petabytes of Earth observation and climate data, offering immense potential for research, policy, and applications. However, access to these datasets requires advanced programming skills and familiarity with domain-specific formats such as NetCDF or GRIB. Moreover, general-purpose Text-to-SQL systems fail when applied naively to the meteorological domain due to a profound “Symbolic-to-Numeric” gap. To overcome these limitations, we present an end-to-end Text-to-SQL framework specifically engineered for real-world, scalable meteorological data exploration. Our system intercepts natural language to resolve spatial and semantic ambiguities before SQL generation. We design STRATOS, a Spatio-Temporal Resolution Agent for Text-to-SQL to dynamically bridge the symbolic-to-numeric gap by mapping fuzzy concepts to a localized ontology and resolving spatial entities via external knowledge bases. Further, our complexity-aware query rewriter rewrites expensive spatial predicates, reducing execution times from hours to seconds. Last, we introduce the STRATOS Evaluation Workload, comprising 7,520 complex query pairs explicitly designed by domain experts to test scalability and symbolic-to-numeric translation across challenging spatio-temporal dimensions previously unexplored by Text-to-SQL systems.

Text-to-SQL systems allow non-SQL experts to interact with relational databases using natural language. However, their tendency to generate executable SQL for ambiguous, out-of-scope, or unanswerable queries introduces a hidden risk, as outputs may be misinterpreted as correct. This risk is especially serious in biomedical contexts, where precision is critical. We therefore present Query Carefully, a pipeline that integrates LLM-based SQL generation with explicit detection and handling of unanswerable inputs. Building on the OncoMX component of ScienceBenchmark, we construct OncoMX-NAQ (No-Answer Questions), a set of 80 no-answer questions spanning 8 categories (non-SQL, out-of-schema/domain, and multiple ambiguity types). Our approach employs llama3.3:70b with schema-aware prompts, explicit No-Answer Rules (NAR), and few-shot examples drawn from both answerable and unanswerable questions. We evaluate SQL exact match, result accuracy, and unanswerable-detection accuracy. On the OncoMX dev split, few-shot prompting with answerable examples increases result accuracy, and adding unanswerable examples does not degrade performance. On OncoMX-NAQ, balanced prompting achieves the highest unanswerable-detection accuracy (0.8), with near-perfect results for structurally defined categories (non-SQL, missing columns, out-of-domain) but persistent challenges for missing-value queries (0.5) and column ambiguity (0.3). A lightweight user interface surfaces interim SQL, execution results, and abstentions, supporting transparent and reliable text-to-SQL in biomedical applications.

Text-to-SQL systems allow users to query databases using natural language. However, users often ask questions that are ambiguous, incomplete, or cannot be answered with the given database. In such cases, models may still generate SQL queries, which can lead to incorrect or misleading results. In this work, we focus on detecting unanswerable queries in text-to-SQL systems.We evaluate our approach on two datasets and databases, ScienceBenchmark (OncoMX) and SM3-Text-to-Query, and extend SM3 with a new dataset of 400 unanswerable questions across eight categories. We compare four large language models under different prompting strategies, including zero-shot and few-shot settings with answerable questions (AQ) and non-answerable questions (NAQ), as well as a Do-Not-Answer (DNA) rule.The results show that few-shot prompting improves performance for AQs, while adding NAQs and the DNA rule does not reduce overall performance. For AQs, Qwen3.5-27B performs best across both datasets, while Llama-3.3-70B-Instruct achieves the best results for detecting NAQs. Overall, Llama-3.3-70B-Instruct provides the best balance between AQs and detecting NAQs. We also show that the SQL correction loop for handling SQL syntax errors has limited impact, with most improvements happening within one or two iterations and almost no effect on detecting NAQs. In summary, the results show that detecting NAQs can be integrated into text-to-SQL systems without reducing performance. This is important for building reliable text-to-SQL systems with real-world applications - especially in biomedical settings.

Flood detection and monitoring are critical for disaster management and risk assessment. Traditional approaches in Switzerland often rely on numerical models for flood prediction. However, these hydrological models typically do not use satellite images. In this paper, we test the hypothesis of whether the image data carries useful flood-related signals or not. The final goal of the research is to improve the accuracy of existing flood prediction models through a multimodal machine learning approach that leverages both numeric and image data. This study evaluates the application of pre-trained deep learning models trained on global Sentinel-2 satellite imagery for water-land classification in the canton of Zurich, Switzerland. We apply a post-processing pipeline using JRC (Joint Research Centre of the European Commission) permanent water data to extract water differences in time-series images that correspond to flood events. The methodology involves applying a pre-trained model to distinguish between water and land, followed by a post-processing step to identify flood-relevant increases of water levels. We validate the approach by calculating the detection accuracy against official flood event labels from the canton of Zurich, revealing both data consistencies and inconsistencies between satellite-derived flood detection and administrative records. While our current implementation uses JRC data, we discuss the advantages of transitioning to permanent water data provided by the Swiss Federal Office for the Environment, which is more accurate and specifically collected for Switzerland. Our findings demonstrate that optical remote sensing captures valuable flood-relevant information that is currently not utilized by existing Swiss hydrological models, including WaSiM, PREVAH, COSMO-LEPS, and EQRN. This work establishes a foundation for enhancing operational flood forecasting systems in Switzerland by integrating satellite imagery with existing hydrological models.

When a Vision-Language Model (VLM) sees a blue banana and answers "yellow", is the problem of perception or arbitration? We explore the question in ten VLMs with various sizes and reveal an Encoding–Grounding Dissociation: models that fail to report what they see (and thus provide a wrong answer) still encode the visual evidence as strongly as models that provide the correct answer. Using Multimodal Arbitration Crossover (MAC) analysis with layer-by-layer Logit Lens probing, we track the competition between visual and prior signals across every layer of each model. We show that visual attributes can be linearly decodable from early layers (AUC > 0.86). The accuracy remains nearly identical for both successful and failed samples. However, the gap in the final-layer logit – not the strength of encoding – better predicts grounding outcomes with a correlation of . After having studied when VLMs base their answers on image clues rather than prior knowledge, we want to understand the causal relationships. We establish causality through full-sequence activation patching. The standard last-token interventions in LLM interpretability do not affect VLMs. In contrast, replacing the full token sequence at layers identified by MAC alters 60 to 84

Many different methods for prompting large language models have been developed since the emergence of OpenAI's ChatGPT in November 2022. In this work, we evaluate six different few-shot prompting methods. The first set of experiments evaluates three frameworks that focus on the quantity or type of shots in a prompt: a baseline method with a simple prompt and a small number of shots, random few-shot prompting with 10, 20, and 30 shots, and similarity-based few-shot prompting. The second set of experiments target optimizing the prompt or enhancing shots through Large Language Model (LLM)-generated explanations, using three prompting frameworks: Explain then Translate, Question Decomposition Meaning Representation, and Optimization by Prompting. We evaluate these six prompting methods on the newly created Spider4SPARQL benchmark, as it is the most complex SPARQL-based Knowledge Graph Question Answering (KGQA) benchmark to date. Across the various prompting frameworks used, the commercial model is unable to achieve a score over 51%, indicating that KGQA, especially for complex queries, with multiple hops, set operations and filters remains a challenging task for LLMs. Our experiments find that the most successful prompting framework for KGQA is a simple prompt combined with an ontology and five random shots.

We present TransClean, a method for detecting false positive predictions of entity matching algorithms under real-world conditions characterized by large-scale, noisy, and unlabeled multi-source datasets that undergo distributional shifts. TransClean is explicitly designed to operate with multiple data sources in an efficient, robust and fast manner while accounting for edge cases and requiring limited manual labeling. TransClean leverages the Transitive Consistency of a matching, a measure of the consistency of a pairwise matching model f_theta on the matching it produces G_f_theta, based both on its predictions on directly evaluated record pairs and its predictions on implied record pairs. TransClean iteratively modifies a matching through gradually removing false positive matches while removing as few true positive matches as possible. In each of these steps, the estimation of the Transitive Consistency is exclusively done through model evaluations and produces quantities that can be used as proxies of the amounts of true and false positives in the matching while not requiring any manual labeling, producing an estimate of the quality of the matching and indicating which record groups are likely to contain false positives. In our experiments, we compare combining TransClean with a naively trained pairwise matching model (DistilBERT) and with a state-of-the-art end-to-end matching method (CLER) and illustrate the flexibility of TransClean in being able to detect most of the false positives of either setup across a variety of datasets. Our experiments show that TransClean induces an average +24.42 F1 score improvement for entity matching in a multi-source setting when compared to traditional pair-wise matching algorithms.

Modern technology and scientific experiments increasingly generate larger and larger amounts of data. This data is sometimes redundant, incomplete or inaccurate and needs to be cleaned and merged with other data before becoming useful for scientific exploration. Hence, entity matching, i.e. the process of linking data about a given entity gathered from multiple data sets, is a major problem in artificial intelligence with applications in science and industry. Typical methods for entity matching either use specialized algorithms or supervised machine learning. Although the problem has been well studied on classical computers, it is unclear how quantum approaches would tackle these challenges. In this paper, we evaluate quantum machine learning algorithms for entity matching on a hand-crafted data set and compare them to similar classical algorithms. We do this by implementing a neural network with a classical embedding layer and extending it with quantum layers. Our experimental results suggest that our hybrid quantum neural network reaches similar performance as classical approaches while requiring an order of magnitude fewer parameters than its classical counterpart. Furthermore, we also show that a model trained on a quantum simulator is portable and thus transferable to a real quantum computer. From a practical perspective and as long as quantum hardware is a scarce resource, experiments, e.g. addressing performance, can profit from producing good initial configurations for quantum neural networks via a simulator, thus only leaving the fine-tuning to quantum computations.

Anomaly detection is an important problem with applications in various domains such as fraud detection, pattern recognition or medical diagnosis. Several algorithms have been introduced using classical computing approaches. However, using quantum computing for solving anomaly detection problems in time series data is a widely unexplored research field. This paper explores the application of quantum autoencoders to time series anomaly detection. We investigate two primary techniques for classifying anomalies: (1) Analyzing the reconstruction error generated by the quantum autoencoder and (2) latent representation analysis. Our simulated experimental results, conducted across various ansaetze, demonstrate that quantum autoencoders consistently outperform classical deep learning-based autoencoders across multiple datasets. Specifically, quantum autoencoders achieve superior anomaly detection performance while utilizing 60-230 times fewer parameters and requiring five times fewer training iterations. In addition, we implement our quantum encoder on real quantum hardware. Our experimental results demonstrate that quantum autoencoders achieve anomaly detection performance on par with their simulated counterparts.

Query optimization has become a research area where classical algorithms are being challenged by machine learning algorithms. At the same time, recent trends in learned query optimizers have shown that it is prudent to take advantage of decades of database research and augment classical query optimizers by shrinking the plan search space through different types of hints (e.g. by specifying the join type, scan type or the order of joins) rather than completely replacing the classical query optimizer with machine learning models. It is especially relevant for cases when classical optimizers cannot fully enumerate all logical and physical plans and, as an alternative, need to rely on less robust approaches like genetic algorithms. However, even symbiotically learned query optimizers are hampered by the need for vast amounts of training data, slow plan generation during inference and unstable results across various workload conditions. In this paper, we present GenJoin - a novel learned query optimizer that considers the query optimization problem as a generative task and is capable of learning from a random set of subplan hints to produce query plans that outperform classical optimizers. GenJoin is the first learned query optimizer that significantly and consistently outperforms PostgreSQL as well as state-of-the-art methods on two well-known real-world benchmarks across a variety of workloads using rigorous machine learning evaluations.

As Large Language Models continue to grow in size and sophistication, more benchmarks are needed to assess their capabilities. Benchmarks are crucial for measuring progress and pushing the scientific community forward. However, they are often expensive and exploitative to build. In this paper, we demonstrate how to translate the rich and complex existing resources for SQL and relational databases that have been used by the database community, for SPARQL and Knowledge Graph Question Answering (KGQA). We describe how we bootstrapped the largest and most complex KGQA benchmark to date, explore the complexities of converting a relational database into a knowledge graph and describe the methods used to learn rich schema information in the KG from its relational counterpart using a query workload-based analysis. In addition, we show how this method can be used to bootstrap any existing SQL benchmark for KGQA by translating the newly created ScienceBenchmark dataset into a KGQA dataset.

The analysis of spectra, such as Nuclear Magnetic Resonance (NMR) spectra, for the comprehensive characterization of peaks is a challenging task for both experts and machines, especially with complex molecules. This process, also known as deconvolution, involves identifying and quantifying the peaks in the spectrum. Machine learning techniques have shown promising results in automating this process. With the advent of quantum computing, there is potential to further enhance these techniques. In this work, inspired by the success of classical Convolutional Neural Networks (CNNs), we explore the use of Quanvolutional Neural Networks (QuanvNNs) for the multi-task peak finding problem, involving both peak counting and position estimation. We implement a simple and interpretable QuanvNN architecture that can be directly compared to its classical CNN counterpart, and evaluate its performance on a synthetic NMR-inspired dataset. Our results demonstrate that QuanvNNs outperform classical CNNs on challenging spectra, achieving an 11% improvement in F1 score and a 30% reduction in mean absolute error for peak position estimation. Additionally, QuanvNNs appear to exhibit better convergence stability for harder problems.

Retrieval-Augmented Generation (RAG) systems fail when documents evolve through versioning-a ubiquitous characteristic of technical documentation. Existing approaches achieve only 58-64

This paper introduces a novel anomaly detection approach tailored for time series data with exclusive reliance on normal events during training. Our key innovation lies in the application of kernel-density estimation (KDE) to scrutinize reconstruction errors, providing an empirically derived probability distribution for normal events post-reconstruction. This non-parametric density estimation technique offers a nuanced understanding of anomaly detection, differentiating it from prevalent threshold-based mechanisms in existing methodologies. In post-training, events are encoded, decoded, and evaluated against the estimated density, providing a comprehensive notion of normality. In addition, we propose a data augmentation strategy involving variational autoencoder-generated events and a smoothing step for enhanced model robustness. The significance of our autoencoder-based approach is evident in its capacity to learn normal representation without prior anomaly knowledge. Through the KDE step on reconstruction errors, our method addresses the versatility of anomalies, departing from assumptions tied to larger reconstruction errors for anomalous events. Our proposed likelihood measure then distinguishes normal from anomalous events, providing a concise yet comprehensive anomaly detection solution. The extensive experimental results support the feasibility of our proposed method, yielding significantly improved classification performance by nearly 10% on the UCR benchmark data.

Classical and learned query optimizers (LQOs) use cardinality estimations as one of the critical inputs for query planning. Thus, accurately predicting the cardinality of arbitrary queries plays a vital role in query optimization. A recent boom in novel deep learning methods stimulated not only the rise of LQOs but also contributed to the appearance of learned cardinality estimators (LCEs). However, the majority of them are based on classical neural networks, ignoring that multivariate correlations between attributes across different tables could be naturally represented via entanglements in quantum circuits. In this paper, we introduce QardEst - Quantum Cardinality Estimator - a novel quantum neural network approach to estimate the cardinality of join queries. Our experiments conducted with a similar number of trainable parameters suggest that quantum neural networks executed on a quantum simulator outperform classical neural networks in terms of mean squared error as well as the q-error.

Motivation:With the proliferation of research means and computational methodologies, published biomedical literature is growing exponentially in numbers and volume. Cancer cell lines are frequently used models in biological and medical research that are currently applied for a wide range of purposes, from studies of cellular mechanisms to drug development, which has led to a wealth of related data and publications. Sifting through large quantities of text to gather relevant information on cell lines of interest is tedious and extremely slow when performed by humans. Hence, novel computational information extraction and correlation mechanisms are required to boost meaningful knowledge extraction.Results:In this work, we present the design, implementation, and application of a novel data extraction and exploration system. This system extracts deep semantic relations between textual entities from scientific literature to enrich existing structured clinical data concerning cancer cell lines. We introduce a new public data exploration portal, which enables automatic linking of genomic copy number variants plots with ranked, related entities such as affected genes. Each relation is accompanied by literature-derived evidences, allowing for deep, yet rapid, literature search, using existing structured data as a springboard.Availability and implementation:Our system is publicly available on the web at https://cancercelllines.org.

The current boom of learned query optimizers (LQO) can be explained not only by the general continuous improvement of deep learning (DL) methods but also by the straightforward formulation of a query optimization problem (QOP) as a machine learning (ML) one. The idea is often to replace dynamic programming approaches, widespread for solving QOP, with more powerful methods such as reinforcement learning. However, such a rapid "game change" in the field of QOP could not pass without consequences - other parts of the ML pipeline, except for predictive model development, have large improvement potential. For instance, different LQOs introduce their own restrictions on training data generation from queries, use an arbitrary train/validation approach, and evaluate on a voluntary split of benchmark queries. In this paper, we attempt to standardize the ML pipeline for evaluating LQOs by introducing a new end-to-end benchmarking framework. Additionally, we guide the reader through each data science stage in the ML pipeline and provide novel insights from the machine learning perspective, considering the specifics of QOP. Finally, we perform a rigorous evaluation of existing LQOs, showing that PostgreSQL outperforms these LQOs in almost all experiments depending on the train/test splits.

Electronic health records (EHRs) are stored in various database systems with different database models on heterogeneous storage architectures, such as relational databases, document stores, or graph databases. These different database models have a big impact on query complexity and performance. While this has been a known fact in database research, its implications for the growing number of Text-to-Query systems have surprisingly not been investigated so far. In this paper, we present SM3-Text-to-Query, the first multi-model medical Text-to-Query benchmark based on synthetic patient data from Synthea, following the SNOMED-CT taxonomy – a widely used knowledge graph ontology covering medical terminology. SM3-Text-to-Query provides data representations for relational databases (PostgreSQL), document stores (MongoDB), and graph databases (Neo4j and GraphDB (RDF)), allowing the evaluation across four popular query languages, namely SQL, MQL, Cypher, and SPARQL. We systematically and manually develop 408 template questions, which we augment to construct a benchmark of 10K diverse natural language question/query pairs for these four query languages (40K pairs overall). On our dataset, we evaluate several common in-context-learning (ICL) approaches for a set of representative closed and open-source LLMs. Our evaluation sheds light on the trade-offs between database models and query languages for different ICL strategies and LLMs. Last, SM3-Text-to-Query is easily extendable to additional query languages or real, standard-based patient databases.

Text-to-SQL systems (also known as NL-to-SQL systems) have become an increasingly popular solution for bridging the gap between user capabilities and SQL-based data access. These systems translate user requests in natural language to valid SQL statements for a specific database. Recent Text-to-SQL systems have benefited from the rapid improvement of transformer-based language models. However, while Text-to-SQL systems that incorporate such models continuously reach new high scores on -- often synthetic -- benchmark datasets, a systematic exploration of their robustness towards different data models in a real-world, realistic scenario is notably missing. This paper provides the first in-depth evaluation of the data model robustness of Text-to-SQL systems in practice based on a multi-year international project focused on Text-to-SQL interfaces. Our evaluation is based on a real-world deployment of FootballDB, a system that was deployed over a 9 month period in the context of the FIFA World Cup 2022, during which about 6K natural language questions were asked and executed. All of our data is based on real user questions that were asked live to the system. We manually labeled and translated a subset of these questions for three different data models. For each data model, we explore the performance of representative Text-to-SQL systems and language models. We further quantify the impact of training data size, pre-, and post-processing steps as well as language model inference time. Our comprehensive evaluation sheds light on the design choices of real-world Text-to-SQL systems and their impact on moving from research prototypes to real deployments. Last, we provide a new benchmark dataset to the community, which is the first to enable the evaluation of different data models for the same dataset and is substantially more challenging than most previous datasets in terms of query complexity.