Vortragende
Peter Steinbach

Moderation
Maik Fiedler

Thema
Data and AI

As AI models become increasingly sophisticated, the importance of high-quality, well-prepared data cannot be overstated. However, in practice, data is often messy, inconsistent, and poorly formatted, leading to frustrated data scientists, stalled projects, and sub-optimal model performance.

Drawing from real-world experiences, we’ll explore the common pitfalls and best practices for ensuring data is „AI-ready“. We’ll discuss the importance of standardized data formats, and how to navigate the complexities of data quality, from handling missing values and outliers to dealing with inconsistent labeling and annotation.

We’ll also examine the critical role of data handling systems, including data warehouses, lakes, and pipelines, in facilitating efficient data preparation and curation. Additionally, we’ll touch on the often-overlooked topic of holdout dataset curation, and why it’s essential for robust model evaluation and deployment.

This talk is designed for data scientists, IT admins or engineers, and researchers who want to improve their data preparation skills, streamline their workflows, and ultimately, want to help build more accurate and reliable AI models.

The talk will be held in English.

Anmeldung zur SaxFDM-Digital Kitchen