@article{schule2026_115103,
  title = {Editorial},
  author = {Schüle, Maximilian E. and Klettke, Meike and Störl, Uta},
  year = {2026},
  pages = {183-186},
  publisher = {Otto-Friedrich-Universität},
  address = {Bamberg},
  urn = {urn:nbn:de:bvb:473-irb-115103x},
  url = {https://fis.uni-bamberg.de/handle/uniba/115103},
  language = {deu}
}

@inproceedings{wolfschmidt2026_115101,
  title = {Back to the Roots : Decision Trees within Modern Database Systems},
  author = {Wolfschmidt, Lena Renate and Schüle, Maximilian Emanuel},
  year = {2026},
  pages = {687-695},
  publisher = {Otto-Friedrich-Universität},
  address = {Bamberg},
  urn = {urn:nbn:de:bvb:473-irb-115101x},
  url = {https://fis.uni-bamberg.de/handle/uniba/115101},
  language = {eng},
  abstract = {HUK-COBURG relies on decision trees to classify e-mails into topics. As the data is stored within relational database systems, this paper investigates the performance improvement of in-database evaluation of decision trees. The paper first summarises the theory behind binary decision trees, then describes the implementation as SQL case-when statements. The evaluation shows the benefit of the in-database application of decision trees by eliminating expensive extraction-transform-load time. In the future, we intend to optimise the training as well as inference of decision trees in SQL to fully eliminate the need for data extraction.}
}

@inproceedings{gutjahr2025_111827,
  title = {DuoLingo-AutoDiff : In-Database Automatic Differentiation with MLIR},
  author = {Gutjahr, Kevin and Ruck, Clemens and Schüle, Maximilian E.},
  year = {2025},
  pages = {1-8},
  publisher = {Otto-Friedrich-Universität},
  address = {Bamberg},
  urn = {urn:nbn:de:bvb:473-irb-111827x},
  url = {https://fis.uni-bamberg.de/handle/uniba/111827},
  language = {eng},
  abstract = {Forward and reverse mode automatic differentiation evaluate the gradient of a model function efficiently by caching the results of partial derivatives. Just-in-time compilation improves the runtime of automatic differentiation by eliminating function calls and storing partial derivatives in virtual registers. This paper discusses the first open-source implementation of automatic differentiation with MLIR and LingoDB. The evaluation compares optimizations applied to forward and reverse modes. It showed that sub-expressions, that appear frequently within the calculation, will be reused after MLIR performs its optimization. Additionally, reverse mode outperforms forward mode due to less generated code.},
  keywords = {Automatic Diferentiation, In-Database Machine Learning, Query Compilation, MLIR}
}

@inproceedings{gutjahr2025_111824,
  title = {DuoLingo-AutoDiff : In-Database Automatic Differentiation with MLIR},
  author = {Gutjahr, Kevin and Ruck, Clemens and Schüle, Maximilian E.},
  year = {2025},
  booktitle = {DEEM '25: Proceedings of the Workshop on Data Management for End-to-End Machine Learning},
  pages = {1-8},
  publisher = {ACM},
  address = {New York},
  isbn = {979-8-4007-1924-0},
  doi = {10.1145/3735654.3735943},
  url = {https://fis.uni-bamberg.de/handle/uniba/111824},
  language = {eng},
  abstract = {Forward and reverse mode automatic differentiation evaluate the gradient of a model function efficiently by caching the results of partial derivatives. Just-in-time compilation improves the runtime of automatic differentiation by eliminating function calls and storing partial derivatives in virtual registers. This paper discusses the first open-source implementation of automatic differentiation with MLIR and LingoDB. The evaluation compares optimizations applied to forward and reverse modes. It showed that sub-expressions, that appear frequently within the calculation, will be reused after MLIR performs its optimization. Additionally, reverse mode outperforms forward mode due to less generated code.},
  keywords = {Automatic Diferentiation, In-Database Machine Learning, Query Compilation, MLIR}
}

@article{schule2025_107115,
  title = {The Duck’s Brain : Training and Inference of Neural Networks within Database Engines},
  author = {Schüle, Maximilian and Neumann, Thomas and Kemper, Alfons},
  year = {2025},
  pages = {209-221},
  publisher = {Otto-Friedrich-Universität},
  address = {Bamberg},
  urn = {urn:nbn:de:bvb:473-irb-1071155},
  url = {https://fis.uni-bamberg.de/handle/uniba/107115},
  language = {eng},
  abstract = {Although database systems perform well in data access and manipulation, their relational model hinders data scientists from formulating machine learning algorithms in SQL. Nevertheless, we argue that modern database systems perform well for machine learning algorithms expressed in relational algebra. To overcome the barrier of the relational model, this paper shows how to transform data into a coordinate relational representation for training neural networks in SQL: We first describe building blocks for data transformation, model training and inference in SQL-92 and their counterparts using an extended array data type. Then, we compare the implementation for model training and inference using array data types to the one using a coordinate relational representation in SQL-92 only. The evaluation in terms of runtime and memory consumption proves the suitability of modern database systems for matrix algebra, although specialised array data types perform better than matrices in coordinate relational representation.},
  keywords = {SQL-92, Neural Networks, Automatic Differentiation, In-Memory Database Systems, Machine Learning}
}

@inproceedings{wolfschmidt2025_106799,
  title = {Back to the Roots : Decision Trees within Modern Database Systems},
  author = {Wolfschmidt, Lena Renate and Schüle, Maximilian Emanuel},
  year = {2025},
  booktitle = {Datenbanksysteme für Business, Technologie und Web: Bamberg, 3.-7. März 2025 : Proceedings},
  pages = {687-695},
  publisher = {Gesellschaft für Informatik},
  address = {Bonn},
  doi = {10.18420/btw2025-37},
  url = {https://fis.uni-bamberg.de/handle/uniba/106799},
  language = {eng},
  abstract = {HUK-COBURG relies on decision trees to classify e-mails into topics. As the data is stored within relational database systems, this paper investigates the performance improvement of in-database evaluation of decision trees. The paper first summarises the theory behind binary decision trees, then describes the implementation as SQL case-when statements. The evaluation shows the benefit of the in-database application of decision trees by eliminating expensive extraction-transform-load time. In the future, we intend to optimise the training as well as inference of decision trees in SQL to fully eliminate the need for data extraction.}
}

@article{schule2024_106796,
  title = {Editorial},
  author = {Schüle, Maximilian E. and Klettke, Meike and Störl, Uta},
  year = {2024},
  journal = {Datenbank-Spektrum},
  volume = {24},
  number = {3},
  pages = {183-186},
  publisher = {Springer Science and Business Media LLC},
  issn = {1610-1995},
  doi = {10.1007/s13222-024-00493-2},
  url = {https://fis.uni-bamberg.de/handle/uniba/106796},
  language = {deu}
}

@article{schule2024_106797,
  title = {The Duck’s Brain : Training and Inference of Neural Networks within Database Engines},
  author = {Schüle, Maximilian and Neumann, Thomas and Kemper, Alfons},
  year = {2024},
  journal = {Datenbank-Spektrum},
  volume = {24},
  number = {3},
  pages = {209-221},
  publisher = {Springer},
  address = {Berlin ; Heidelberg},
  issn = {1610-1995},
  doi = {10.1007/s13222-024-00485-2},
  url = {https://fis.uni-bamberg.de/handle/uniba/106797},
  language = {eng},
  abstract = {Although database systems perform well in data access and manipulation, their relational model hinders data scientists from formulating machine learning algorithms in SQL. Nevertheless, we argue that modern database systems perform well for machine learning algorithms expressed in relational algebra. To overcome the barrier of the relational model, this paper shows how to transform data into a coordinate relational representation for training neural networks in SQL: We first describe building blocks for data transformation, model training and inference in SQL-92 and their counterparts using an extended array data type. Then, we compare the implementation for model training and inference using array data types to the one using a coordinate relational representation in SQL-92 only. The evaluation in terms of runtime and memory consumption proves the suitability of modern database systems for matrix algebra, although specialised array data types perform better than matrices in coordinate relational representation.},
  keywords = {SQL-92, Neural Networks, Automatic Differentiation, In-Memory Database Systems, Machine Learning}
}

@inproceedings{sachnov2024_106798,
  title = {Give a JIT on GPUs : NVRTC for Code-Generating Database Systems},
  author = {Sachnov, Anton and Merzljak, Leonard von and Schüle, Maximilian E.},
  year = {2024},
  booktitle = {2024 IEEE 40th International Conference on Data Engineering Workshops (ICDEW)},
  pages = {384-387},
  publisher = {IEEE},
  doi = {10.1109/icdew61823.2024.00061},
  url = {https://fis.uni-bamberg.de/handle/uniba/106798},
  language = {eng}
}

@inproceedings{schule2024_106800,
  title = {Higher-Order SQL Lambda Functions},
  author = {Schüle, Maximilian E. and Hornung, Jakob},
  year = {2024},
  booktitle = {2024 IEEE 40th International Conference on Data Engineering (ICDE)},
  pages = {5622-5628},
  publisher = {IEEE},
  doi = {10.1109/icde60146.2024.00450},
  url = {https://fis.uni-bamberg.de/handle/uniba/106800},
  language = {eng}
}

@inproceedings{ruck2023_94682,
  title = {Teaching Blue Elephants the Maths for Machine Learning},
  author = {Ruck, Clemens and Schüle, Maximilian Emanuel},
  year = {2023},
  booktitle = {DEEM '23: Proceedings of the Seventh Workshop on Data Management for End-to-End Machine Learning},
  pages = {1-4},
  publisher = {ACM},
  isbn = {979-8-4007-0204-4},
  doi = {10.1145/3595360.3595852},
  url = {https://fis.uni-bamberg.de/handle/uniba/94682},
  language = {eng}
}

@inproceedings{schule2023_59173,
  title = {Recursive SQL and GPU-Support for In-Database Machine Learning},
  author = {Schüle, Maximilian},
  year = {2023},
  pages = {931},
  publisher = {Otto-Friedrich-Universität},
  address = {Bamberg},
  urn = {urn:nbn:de:bvb:473-irb-591733},
  url = {https://fis.uni-bamberg.de/handle/uniba/59173},
  language = {eng},
  abstract = {But to retrieve the latest data from a database, time-consuming extraction is necessary as database systems have rarely been used for operations such as matrix algebra and gradient descent.In this work, we demonstrate that SQL with recursive tables makes it possible to express a complete machine learning pipeline out of data preprocessing, model training and its validation.To facilitate the specification of loss functions, we extend the code-generating database system Umbra by an operator for automatic differentiation for use within recursive tables:With the loss function expressed in SQL as a lambda function, Umbra generates machine code for each partial derivative.We further use automatic differentiation for a dedicated gradient descent operator, which generates LLVM code to train a user-specified model on GPUs.We fine-tune GPU kernels at hardware level to allow a higher throughput and propose non-blocking synchronisation of multiple units.In our evaluation, automatic differentiation accelerated the runtime by the number of cached subexpressions compared to compiling each derivative separately.Our GPU kernels with independent models allowed maximal throughput even for small batch sizes, making machine learning pipelines within SQL more competitive.},
  keywords = {SQL, Machine Learning}
}

@inproceedings{schule2023_59172,
  title = {NN2SQL : Let SQL Think for Neural Networks},
  author = {Schüle, Maximilian and Kemper, Alfons and Neumann, Thomas},
  year = {2023},
  pages = {183-194},
  publisher = {Otto-Friedrich-Universität},
  address = {Bamberg},
  urn = {urn:nbn:de:bvb:473-irb-591725},
  url = {https://fis.uni-bamberg.de/handle/uniba/59172},
  language = {eng},
  abstract = {Although database systems perform well in data access and manipulation, their relational model hinders data scientists from formulating machine learning algorithms in SQL. Nevertheless, we argue that modern database systems perform well for machine learning algorithms expressed in relational algebra. To overcome the barrier of the relational model, this paper shows how to transform data into a relational representation for training neural networks in SQL: We first describe building blocks for data transformation in SQL. Then, we compare an implementation for model training using array data types to the one using a relational representation in SQL-92 only. The evaluation proves the suitability of modern database systems for matrix algebra, although specialised array data types perform better than matrices in relational representation.},
  keywords = {SQL-92, Neural Networks, Automatic Differentiation}
}

@inproceedings{schule2023_58982,
  title = {Recursive SQL and GPU-Support for In-Database Machine Learning},
  author = {Schüle, Maximilian},
  year = {2023},
  booktitle = {Lecture Notes in Informatics (LNI) : Proceedings ; P-331},
  pages = {931},
  publisher = {Gesellschaft für Informatik e.V.},
  address = {Bonn},
  isbn = {978-3-88579-725-8},
  doi = {10.18420/BTW2023-62},
  url = {https://fis.uni-bamberg.de/handle/uniba/58982},
  language = {eng},
  abstract = {In machine learning, continuously retraining a model guarantees accurate predictions based on the latest data as training input. But to retrieve the latest data from a database, time-consuming extraction is necessary as database systems have rarely been used for operations such as matrix algebra and gradient descent.In this work, we demonstrate that SQL with recursive tables makes it possible to express a complete machine learning pipeline out of data preprocessing, model training and its validation.To facilitate the specification of loss functions, we extend the code-generating database system Umbra by an operator for automatic differentiation for use within recursive tables:With the loss function expressed in SQL as a lambda function, Umbra generates machine code for each partial derivative.We further use automatic differentiation for a dedicated gradient descent operator, which generates LLVM code to train a user-specified model on GPUs.We fine-tune GPU kernels at hardware level to allow a higher throughput and propose non-blocking synchronisation of multiple units.In our evaluation, automatic differentiation accelerated the runtime by the number of cached subexpressions compared to compiling each derivative separately.Our GPU kernels with independent models allowed maximal throughput even for small batch sizes, making machine learning pipelines within SQL more competitive.},
  keywords = {SQL, Machine Learning}
}

@inproceedings{schule2023_58981,
  title = {NN2SQL : Let SQL Think for Neural Networks},
  author = {Schüle, Maximilian and Kemper, Alfons and Neumann, Thomas},
  year = {2023},
  booktitle = {Lecture Notes in Informatics (LNI) : Proceedings ; P-331},
  pages = {183-194},
  publisher = {Gesellschaft für Informatik e.V.},
  address = {Bonn},
  isbn = {978-3-88579-725-8},
  doi = {10.18420/BTW2023-09},
  url = {https://fis.uni-bamberg.de/handle/uniba/58981},
  language = {eng},
  abstract = {Although database systems perform well in data access and manipulation, their relational model hinders data scientists from formulating machine learning algorithms in SQL. Nevertheless, we argue that modern database systems perform well for machine learning algorithms expressed in relational algebra. To overcome the barrier of the relational model, this paper shows how to transform data into a relational representation for training neural networks in SQL: We first describe building blocks for data transformation in SQL. Then, we compare an implementation for model training using array data types to the one using a relational representation in SQL-92 only. The evaluation proves the suitability of modern database systems for matrix algebra, although specialised array data types perform better than matrices in relational representation.},
  keywords = {SQL-92, Neural Networks, Automatic Differentiation}
}

@inproceedings{schule2023_58983,
  title = {Blue Elephants Inspecting Pandas: Inspection and Execution of Machine Learning Pipelines in SQL},
  author = {Schüle, Maximilian E. and Scalerandi, Luca and Kemper, Alfons and Neumann, Thomas},
  year = {2023},
  publisher = {OpenProceedings.org},
  doi = {10.48786/EDBT.2023.04},
  url = {https://fis.uni-bamberg.de/handle/uniba/58983},
  language = {eng}
}

