LPLennart Purucker

Lennart Purucker

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Lennart Purucker has a total of 26 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    MARVIS: Modality Adaptive Reasoning over VISualizations

    by Benjamin Feuer, Lennart Purucker, Oussama Elachqar, Chinmay Hegde

  • Open Access

    Enhancing Intra-Continental Biogeographical Ancestry Prediction Through a Machine Learning Marker Selection Method

    by Theresa Maurer, Lennart Purucker, Frank Hutter, Peter Pfaffelhuber, Carola Sophia Heinzel

    While classifiers such as TabPFN (Hollmann et al., 2025) and SNIPPER (Phillips et al., 2007a) achieve strong intercontinental performance (Heinzel et al., 2025), their accuracy in classifying individuals within Europe remains low. One major factor contributing to this limitation is the set of genetic markers used for classification. Marker panels such as the VISAGE Enhanced Tool (Xavier et al., 2022) are commonly employed in forensic genetics because they contain ancestry-informative markers (AIMs) that distinguish very well between major continental populations. However, these panels are often not optimized for fine-scale differentiation within continents, where genetic variation is more subtle and population structure is rather continuous. We apply machine learning to select informative markers for intra-European classification, using data from Consortium et al. (2015). Compared with the VISAGE Enhanced Tool and allele frequency–based approaches (Phillips et al., 2007b; Kosoy et al., 2009; Nassir et al., 2009; Kidd et al., 2014; Phillips et al., 2014a), our marker sets achieve substantially higher accuracy within Europe: For four European populations, accuracy improves from 68.2% (VISAGE, 104 markers) to 73.7% (100 new markers) and 82.3% (200 new markers). For five populations, accuracy rises from 56.1% (VISAGE) to 64.5% (100 new markers). Our results show that tailored marker selection markedly improves intra-continental classification. While optimized here for Europe, the method can be applied to any region with sufficient training data.

  • Open Access

    Agentic NL2SQL to Reduce Computational Costs

    by Dominik Jehle, Lennart Purucker, Frank Hutter

    Translating natural language queries into SQL queries (NL2SQL or Text-to-SQL) has recently been empowered by large language models (LLMs). Using LLMs to perform NL2SQL methods on a large collection of SQL databases necessitates processing large quantities of meta-information about the databases, which in turn results in lengthy prompts with many tokens and high processing costs. To address this challenge, we introduce Datalake Agent, an agentic system designed to enable an LLM to solve NL2SQL tasks more efficiently. Instead of utilizing direct solvers for NL2SQL that call the LLM once with all meta-information in the prompt, the Datalake Agent employs an interactive loop to reduce the utilized meta-information. Within the loop, the LLM is used in a reasoning framework that selectively requests only the necessary information to solve a table question answering task. We evaluate the Datalake Agent on a collection of 23 databases with 100 table question answering tasks. The Datalake Agent reduces the tokens used by the LLM by up to 87\% and thus allows for substantial cost reductions while maintaining competitive performance.

  • Open Access

    Advancing biogeographical ancestry predictions through machine learning

    by Carola Sophia Heinzel, Lennart Purucker, Frank Hutter, Peter Pfaffelhuber


  • Open Access

    Advancing Biogeographical Ancestry Predictions Through Machine Learning

    by Carola Sophia Heinzel, Lennart Purucker, Frank Hutter, Peter Pfaffelhuber

    Tools like Snipper or the Admixture Model count as state-of-the-art methods in forensic science for biogeographical ancestry. However, they have not been systematically compared to classifiers widely used in other disciplines. Noting that genetic data have a tabular form, this study addresses this gap by benchmarking forensic classifiers against TabPFN, a cutting-edge, general-purpose machine learning classifier for tabular data. The comparison evaluates performance using metrics such as accuracy—the proportion of correct classifications—and ROC AUC. We examine classification tasks for individuals at both the intracontinental and continental levels, based on a published dataset for training and testing. Our results reveal significant performance differences between methods, with TabPFN consistently achieving the highest accuracy and ROC AUC across the datasets. Based on these findings, we recommend adopting TabPFN for population classification in forensic science.

  • Top Cited Papers
    Open Access

    Accurate predictions on small data with a tabular foundation model

    by Noah Hollmann, Samuel Müller, Lennart Purucker, Arjun Krishnakumar, Max Körfer, Shi Bin Hoo, Robin Tibor Schirrmeister, Frank Hutter

    Tabular data, spreadsheets organized in rows and columns, are ubiquitous across scientific fields, from biomedicine to particle physics to economics and climate science1,2. The fundamental prediction task of filling in missing values of a label column based on the rest of the columns is essential for various applications as diverse as biomedical risk models, drug discovery and materials science. Although deep learning has revolutionized learning from raw data and led to numerous high-profile success stories3–5, gradient-boosted decision trees6–9 have dominated tabular data for the past 20 years. Here we present the Tabular Prior-data Fitted Network (TabPFN), a tabular foundation model that outperforms all previous methods on datasets with up to 10,000 samples by a wide margin, using substantially less training time. In 2.8 s, TabPFN outperforms an ensemble of the strongest baselines tuned for 4 h in a classification setting. As a generative transformer-based foundation model, this model also allows fine-tuning, data generation, density estimation and learning reusable embeddings. TabPFN is a learning algorithm that is itself learned across millions of synthetic datasets, demonstrating the power of this approach for algorithm development. By improving modelling abilities across diverse fields, TabPFN has the potential to accelerate scientific discovery and enhance important decision-making in various domains.


  • Open Access

    AMLTK: A Modular AutoML Toolkit in Python

    by Edward Bergman, Matthias Feurer, Aron Bahram, Amir Rezaei Balef, Lennart Purucker, Sarah Segel, Marius Lindauer, Frank Hutter, Katharina Eggensperger

    Machine Learning is a core building block in novel data-driven applications. Practitioners face many ambiguous design decisions while developing practical machine learning (ML) solutions. Automated machine learning (AutoML) facilitates the development of machine learning applications by providing efficient methods for optimizing hyperparameters, searching for neural architectures, or constructing whole ML pipelines (Hutter et al., 2019). Thereby, design decisions such as the choice of modelling, pre-processing, and training algorithm are crucial to obtaining well-performing solutions. By automatically obtaining ML solutions, AutoML aims to lower the barrier to leveraging machine learning and reduce the time needed to develop or adapt ML solutions for new domains or data.


  • Open Access

    Revealing the Hidden Impact of Top-N Metrics on Optimization in Recommender Systems

    by Lukas Wegmeth, Tobias Vente, Lennart Purucker

    The hyperparameters of recommender systems for top-n predictions are typically optimized to enhance the predictive performance of algorithms. Thereby, the optimization algorithm, e.g., grid search or random search, searches for the best hyperparameter configuration according to an optimization-target metric, like nDCG or Precision. In contrast, the optimized algorithm, internally optimizes a different loss function during training, like squared error or cross-entropy. To tackle this discrepancy, recent work focused on generating loss functions better suited for recommender systems. Yet, when evaluating an algorithm using a top-n metric during optimization, another discrepancy between the optimization-target metric and the training loss has so far been ignored. During optimization, the top-n items are selected for computing a top-n metric; ignoring that the top-n items are selected from the recommendations of a model trained with an entirely different loss function. Item recommendations suitable for optimization-target metrics could be outside the top-n recommended items; hiddenly impacting the optimization performance. Therefore, we were motivated to analyze whether the top-n items are optimal for optimization-target top-n metrics. In pursuit of an answer, we exhaustively evaluate the predictive performance of 250 selection strategies besides selecting the top-n. We extensively evaluate each selection strategy over twelve implicit feedback and eight explicit feedback data sets with eleven recommender systems algorithms. Our results show that there exist selection strategies other than top-n that increase predictive performance for various algorithms and recommendation domains. However, the performance of the top ~43% of selection strategies is not significantly different. We discuss the impact of our findings on optimization and re-ranking in recommender systems and feasible solutions.