Big Data Engineering
Data Science Abstractions and Systems, Performance-Accuracy Tradeoffs in Data Science, Data Cleaning Pipelines and Optimization
The mission of the Big Data Engineering group, led by Prof. Dr. Matthias Böhm, is to simplify data science by providing high-level, data-science-centric abstractions and building systems and tools to execute these tasks in an efficient and scalable manner. The general research interests include the exploration of performance-accuracy tradeoffs, tooling (script generators, label generation, advisors, etc.), seamless data augmentation, cleaning, feature engineering, model debugging and deployment, cost-effective cloud deployments, advanced optimization techniques, adaptive data storage and indexing, and the exploitation of modern hardware.
Current research focuses on:
• Data Cleaning Pipelines: Automatic enumeration of data cleaning pipelines for target ML application, hyperparameter optimization of cleaning primitives.
• Model Debugging: Finding the top-k data slices where a trained model underperforms, linear-algebra-based enumeration and pruning algorithms.
• Fine-grained Lineage Tracing and Reuse: Fine-grained, multi-level lineage tracing for versioning and reuse, lineage deduplication, full and partial reuse of intermediates.
• Federated Linear Algebra and Parameter Servers: ML model training on federated raw data without central data consolidation, plan generation under awareness of privacy constraints, federated linear algebra programs and parameter servers.
• Workload-aware Data Reorganization: Compression under awareness of data and workload (linear algebra program) characteristics, asynchronous data reorganization in standing executors (e.g., at standing federated workers).
• Code Generation for Heterogeneous HW: Extended operator fusion and code generation for GPUs and heterogeneous devices, including sparsity exploitation across operations.
POLAR: Adaptive and Non-invasive Join Order Selection via Plans of Least Resistance
Cross-Language Differential Testing of JSON Parsers
Special issue on “Machine learning and databases”
“POLAR” lowers the adoption barrier for adaptive query processing in database systems
A preprint by BIFOLD researchers titled "POLAR: Adaptive and Non-invasive Join Order Selection via Plans of Least Resistance" is set to be presented at the VLDB conference in 2024. The database engineering paper introduces a technique for reordering joins that is adaptive, with a focus on non-invasive integration and low overhead.
DIMA-DAMS Annual Retreat
The three-day retreat was aimed at discussing the research results and ongoing work of doctoral students in the BIFOLD Data Systems Lab: currently, comprised of the Database Systems and Information Management (DIMA) and Big Data Engineering (DAMS) groups.
Photo recap: BIFOLD New Year's reception
At its New Year's reception BIFOLD welcomed a series of distinguished guests and friends from Berlin's AI community.
Research Group Lead
Doctoral Researcher
Postdoctoral Researcher
Secretary MLSEC & DAMS
Doctoral Researcher
Doctoral Researcher