Big-Data.AI Summit 2022
2022-09-14..2022-09-15
Berlin, Germany
| title | author | length |
|---|---|---|
| Welcome – Susanne Dehmel | Susanne Dehmel | 08m |
Welcome and conference logistics: on-site Berlin return with a packed program and partners listed including Accenture, Dataiku, Google Cloud; sessions focus on… | ||
| Opening Keynote | Ina Schieferdecker | 14m |
Germany reports six AI centers of excellence with long term funding, over 100 new AI professorships of which about one third come from abroad, and an upcoming… | ||
| The Time Is Right For Everyday AI! | Dr. Sebastian Werner | 23m |
Argues that AI is market ready for everyday use with concrete metrics and operational lessons: cites McKinsey's ~1.2% GDP boost, reports 75–85% fail to reach… | ||
| Custom Training of AI for Language Services: Translation is Just the Beginning | Alexander Gigga | 13m |
Custom models run as per-customer instances called Halos human-augmented language OS that ingest translation memories via automated pipelines containing… | ||
| Panel: Scaling AI successfully | Florian Doerr, Thomas Edward Hahn, Hana Niederbrucker, Feiyu Xu | 56m |
Scaling succeeds when companies stop treating AI as isolated pilots and adopt a data-first architecture with a central data hub, global metadata and governance… | ||
| Establishing AI Governance in Practice: How to Benchmark against the Gold Standard | Ryan Donnelly | 13m |
Enzai Technologies Limited launches an AI governance risk management platform that integrates with MLflow and SAS via a flexible API, bundles framework-aligned… | ||
| AI as Medical Device - Interaction between the AI Act and the Medical Device Regulation | Heinz-Uwe Dettling | 18m |
Explains that software is a medical device only when tied to a specific disease and patient use, and that an AI using learning techniques is high-risk AI… | ||
| Panel: Regulating Artificial Intelligence – The AI Act and its Implications | Andreas Hartl, Tobias Heimann, Nadja Kroha, Svea Windwehr | 46m |
Panel summarizes concrete frictions in the AI Act: one medical firm reports 60 AI-enabled products already CE‑marked under the Medical Devices Regulation… | ||
| Creating Markets, Creating Trustworthiness: Standards for AI | Ralf Banisch, Joachim Lonien | 21m |
Mindpeak describes an AI that locates and classifies single tumor cells on gigapixel whole-slide images, trained on data totaling about 5–10x COCO in pixels to… | ||
| Training Data and its Impact on Cultural Understanding & Ethical Values in Giant AI Models | Jonas Andrulis | 19m |
Aleph Alpha built its own data center (ranked Place 72 in Europe) and trains models across five languages, claiming much stronger German/French/Spanish results… | ||
| Scaling AI: From Experimentation to Transformation | Florian Schaudel | 19m |
Thousands surveyed show the bottleneck is people and adoption not models or missing data: organizations spend millions over a decade but still have “more… | ||
| Successfully Scaling AI and the Role of Trust | Sarah Engel | 15m |
A concrete claim: scaling AI fails when users can’t rely on outputs — a neuroprosthetic user chooses a simple, deterministic arm over a high‑tech one because… | ||
| Development of a Transformation Roadmap with the AppliedAI Maturity Assessment | Philipp Hartmann, Rainer Hoffmann | 18m |
A web‑based AI Maturity Assessment of roughly 109 items across nine dimensions is used to map where organizations stand, with a partner edition run first with… | ||
| Empowering Analysts With No-Code Machine Learning | Riley MacDonald | 16m |
Claims analysts can build supervised models without code using Qlik AutoML, with automatic feature engineering, algorithm selection and hyperparameter search… | ||
| How to Connect Data & AI with Dog Intelligence | Florian Dohmann, Nina Stumme | 19m |
A new DB Netz division builds production-ready data products and claims concrete wins: training sniffer dogs + scent data and centralizing those observations… | ||
| How Hashtag You Builds New Social Media Brands Based on Data | Ashish Kalra | 15m |
Hashtag You builds B2C social brands by plumbing 65–70 data sources into a single stack and claims the small data team moves from setup to dashboards in 21–25… | ||
| How to Drive Impact and Scale with Your Big Data/AI Applications? | Klaas Mertens | 24m |
Claim: data attracts compute so move algorithms to data not the other way round. Concrete constraints: fiber economics make 40 km a practical radius, Germany’s… | ||
| Data Mesh, Data Fabric & Co - the End of Monolithic Architecture Concepts? | Ulrich Hohmann | 20m |
Argues that data virtualization is a practical foundation for both data fabric and data mesh, delivering a semantic layer that leaves data in place and… | ||
| Goodbye by Bitkom, Closing of Day 1 | Merle Uhl | 01m |
Thanks and sponsor roll call names specific partners: Accenture, Google, premium partners including D R E C O, de noodle, Dean5Trend, Click and Stack, Fuel… | ||
| Lessons from Analysing all of Europe's Data | Guillaume Aimetti, Georg Wittenburg | 18m |
A system autonomously crawls EU and national portals, processes 140,000 CSVs in eight days on 256 Azure instances and averages 4 minutes per dataset, using a… | ||
| Initializing a Sustainable Data Mesh Transformation | Carsten Fritsch | 21m |
Deutsche Bahn runs a 10‑week pilot that produces a tunnel image data product collecting inspectors’ photos (picked 3–5 of ~50 inspectors) and a central sharing… | ||
| Data Fabric or Data Mesh - Similarities, Differences & Optimal Possible Combination of Both | Harald Gröger | 20m |
Argues that data fabric = technology (catalogs, knowledge graphs, metadata activation, ML-driven automation) while data mesh = architecture (domain-owned data… | ||
| Behavioral Data Creation for AI - Delivering Value and Insights | Alexander Dean | 18m |
Argues that owning high‑resolution first‑party behavioral events creates real competitive advantage, not just more data; Snowplow collects atomic event… | ||
| A Brief History of Data Traffic at DB - about the Big Bang and Black Holes | Stephan Kaulbach | 17m |
Deutsche Bahn centralizes group data since 2020, spins up a Data Intelligence Center and ships 15 operational use cases tied to punctuality, capacity and… | ||
| Forecasting for Everyone - Foundation Models and the Future of Time Series Prediction | Fabian Müller | 19m |
A foundation-model pipeline trains transformer-based networks on 250 million time series points from many industries, then performs pure inference on unseen… | ||
| AI or Not AI? That's the Answer! | Christian Stephan | 22m |
Project CovidDataNet.NRW gathers standardized hospital and wearable data to decode sepsis with pattern analysis and trains models that predict 30‑day mortality… | ||
| Turn that Data into a Differentiator - How Analytics Can Drive Product Value | Martin Aussenhof | 16m |
Claims concrete ROI: embedding Sisense Infuse raises subscription revenue up to 25% and product value by 43% in cited customers, a logistics client lifts win… | ||
| Synthetic Data for Privacy, AI Fairness and Democratizing Data Access | Alexandra Ebert | 20m |
Synthetic data generated by MOSTLY AI and demonstrated with NVIDIA research claims near-identical statistical utility to real records while remaining out of… | ||
| Using AI to Prevent Residential Burglary in Berlin | Lukas Lötters | 23m |
Krimpro is a predictive system deployed in Berlin since 2017 that forecasts residential burglary risk per 400×400m quadrant for the next 72 hours, trained on… | ||
| How EnBW Utilizes Data & AI Trainings to Upskill Employees & Enable Cross-Divisional Work | Rainer Hoffmann, Caroline Hirsch | 13m |
EnBW runs a year-long Data Professional Development Program (DP)2 for 20 data experts that spends 30% time commitment on four coordinated pillars: cohort group… | ||
| The Treasure Trove of Credit Data: Managing Risk | Edeltraud Leibrock, Christian Thun | 22m |
Market-owned European DataWarehouse runs an Azure cloud + DevOps platform that captures loan-level ABS data since 2013, holding 100 million loans with 50–60… | ||
| Can AI Support Our City Trees in Changing Climate? | Marcus Voß | 22m |
Project named Quantified Trees uses soil moisture sensors at 30 60 and 90 cm and open datasets on 800,000 city trees plus radar rainfall and watering logs to… | ||
| From Data Strategy to a Scalable Data Mesh – How to Successfully Navigate the Journey | Martin Whyte | 23m |
Case study shows a large Swedish retailer (IKEA) moving from shared warehouses to a scalable data mesh by defining domains, creating data products with SLAs… | ||
| Moving Forward with Explainable AI: Adoption Rate & Essential Dimensions for a Reliable Use | Lukas Böhres, Manuel Marowski | 19m |
PwC reports 4.5 operationalized AI use cases per company and finds 46% fear productivity loss, 44% fear lost customer confidence and 25% fear legal risk when… | ||
| "Automation of Automation" for Competitive Production Systems | Werner Kraus | 24m |
Fraunhofer IPA shows concrete systems: automated layout generation, self‑learning bin picking and AMR material flow integrated with simulation, point‑cloud… | ||
| Gaia-x FAIC: Democratizing AI and Technology for Sustainable Finance and ESG | Jochen Papenbrock | 19m |
Gaia‑X FAIC claims a marketplace for validated models and a set of reproducible methodologies and is already shipping demonstrators: an Earth 2 digital twin… | ||
| Scaling Data Mesh @ Allianz Direct - Insights, achievements & learnings | Martin Günther, Lukas Feuerstein | 18m |
Allianz Direct moves from central BI to domain teams by shipping configurable connectors that business users trigger via automatic deployment workflows, so… | ||
| Extending Mobility - How AI Contributes towards the Future of Railway Transportation | Thomas Thiele | 18m |
Central unit builds and deploys concrete systems: a centralized construction-site platform (prototype finished April) feeding capacity planning; a… | ||
| Marispace-X: Connect the Dots | Lars Francke | 22m |
Marispace‑X builds a Gaia‑X compatible data space (€15M funding, 11 projects chosen) that moves maritime sensor data from edge to fog to cloud with open‑source… | ||
| Complementing Political/Social Research with AI for a Deeper Understanding of Society | Martin Schniedermeier, Georg Wittenburg | 19m |
A hybrid rule-based system plus neural classifiers automates survey analysis by implementing about 420 rules to emulate human reasoning, producing pre-cleaned… | ||
| Automated Risk Evaluation and Mitigation for AI Systems | Florens Greßner | 18m |
neurocat condenses five years of field experience into a software module that automates the systematic risk assessment demanded by the EU AI Act, integrating… | ||
| AI Systems Engineering - Challenges of a new professional discipline | Julius Pfrommer | 22m |
Competence Center CC-KING develops predictive quality models for 300 m wood‑panel lines that cut energy and material costs by millions of euros; prototypes… | ||
| MLOps End-to-End Speedrun: Build an APIs around a Spam Classifier Model | Dr Stefano Lottini | 19m |
A working pipeline trains an LSTM recurrent neural network on extracted features, stores models with explicit versioning, and serves them behind Python FastAPI… | ||
| Beyond MLOps - How Federated MLOps Unlocks the Next Frontier of AI Innovation | Robin Röhm | 20m |
Claim: federated MLOps lets organizations keep data in place and run analytics or training across partners to unlock commercial value without sharing raw… | ||
| Free Texts to Knowledge Graphs - Holy Grail of Semantic Computing on an Industrial Scale | Bernd Geiger | 18m |
Claims a working pipeline that turns PDFs into JSON with layout, extracts n‑grams, builds word‑level graphs and maps them into active ontologies implemented in… | ||
| Exploratory Data Science on Federated Raw Data | Claus Neubauer | 19m |
A federated exploratory pipeline is built that ships algorithms to data sites instead of centralizing raw streams, using Apache SystemDS for parallelized model… | ||
| AI in Quality Control – Real Use Cases and Best Practices | Tobias Lehrer-Ruhland | 17m |
Optical inspection systems use infrared camera images plus meta tags and labeled training data to train a visual‑inspection AI; a cookie line PoC achieves… | ||
| Intelligent Decision Support: Enabling CSM & FSM with Health Monitoring | Sebastian Baumann | 20m |
DATAbility combines physics-based engineering rules with machine learning to predict and explain manufacturing outcomes using hybrid models. A deployed MVP… | ||
| Vertical Federated Learning across the Value Chain | Michael Kühne-Schlinkert | 22m |
Katulu applies vertical federated learning across industrial value chains by slicing models into a U-shape split model and having suppliers train local layers… | ||
| Key Points and Lessons Learned in AI Projects in Mechanical Engineering | Vanessa Müller | 16m |
Company builds production AI for mechanical engineering using the Python ecosystem and GitLab/Azure/GCP, runs a one‑year POC→cloud rollout, and now operates… | ||
| Data Driven Production Process Optimization Machine Learning Toolbox & Beverage Industry | Ralf Klinkenberg | 26m |
A modular DaPro toolbox built on RapidMiner with industry-specific data schemas and semi-automated data mapping enables brewers and process engineers to run… | ||
| AI as a Service: Quick and Scalable Machine Learning POCs with No-Code/Low-Code Tools | Julius Otto | 15m |
Proof of concept uses Amazon Recognition with synthetic data augmentation to detect weld spatters on motorcycle frames, scaling a tiny dataset (~100 images) up… | ||
| Welcome – Merle Uhl | Merle Uhl | 02m |
Second day agenda lays out concrete sessions: a deep dive into industrial use cases on the main stage, ground-floor tracks on cloud enabling innovation and… | ||
| AI Gardening: Harvesting and Pruning the AI Tree | Christoph Peylo | 15m |
Founded Bosch Center for AI in 2017, advocates hybrid modeling that combines physical process knowledge with ML and uses anomaly detection for wafer production… | ||
| Mapping the Future of Data & AI | Frank Theisen | 15m |
Argues that solving data gravity needs a data fabric that gives virtual, low‑duplication access with automated business and technical metadata and master‑data… | ||
| Extending European Strength with Embedded AI | Nico Kelling | 22m |
Argues that processing moves from cloud to embedded AI as a continuum, claiming millisecond control needs imply servers <100 km away and many use cases must… | ||
| Computer Vision Toolbox - Connected Computer Vision & AI Quality Inspection at Scale! | Anja Fessler, Hana Niederbrucker | 23m |
ZF builds an in‑house Computer Vision Toolbox: an edge agent that triggers cameras and sends images + metadata to the cloud over DMP, stores and trains in AAP… | ||
| AI Assisted Load Sharing for Improved Energy Efficiency | Anja Vedder | 17m |
AI builds a physics‑aware digital twin of a five‑compressor air supply for two clarifiers, fits thermodynamic parameters to historical and real‑time… | ||
| Data Scientist: The Sexiest Job of the Century 10 Years Later - Still Sexy? | Alexander Friedenberger | 22m |
Claims that talent—more than data or process—is the biggest scaling hurdle, and that a second generation of grads plus new roles (data engineer, ML engineer… | ||
| Collaborative Self-Learning for Security, Fun and Profit | Prof. Dr. Ahmad-Reza Sadeghi | 22m |
Describes building a production federated learning platform for mobile cyber risk intelligence that runs across nine companies and trains on 24 million users… | ||
| Enabling Medical AI through Federated Learning | Taner Topal | 18m |
Medical AI needs diverse training data but hospitals rarely share raw records so federated evaluation first sends a global model to sites for local metrics and… | ||
| Introduction of MedModels | Philipp Koch | 20m |
Claims that a computer scientist will win a Nobel in medicine within 15 years and that machine learning can unlock longitudinal EHR and claims data are backed… | ||
| Cloud Security Megatrends - No Data-Driven Innovation without Security | Phil Venables | 23m |
Argues eight measurable megatrends that push hyperscale cloud security upward: a $10B investment and Titan security chip on millions of servers, commoditized… | ||
| Panel: Cloud as an Enabler for Data-Driven Innovation in Healthcare | Christina Busmalis, Julia Hagen, Harald Wagener, Maria Barros Weiss | 47m |
Cloud projects here are concrete: a vendor-built Health Suite platform layers healthcare services on public clouds; a Healthcare Data Engine joins clinical… | ||
| Digital Ethics in Business Practice: Operationalizing a Principle-Based Ethics Framework | Simon Lucas | 19m |
Merck establishes a Digital Ethics Advisory Panel (DEAP) in 2021 and a board‑approved Code of Digital Ethics (CoDE) of five core principles, then embeds them… | ||
| Sustainability of AI | Dr. Maria Börner | 20m |
AI training carbon depends on location, datacenter PUE, model complexity and training time: CIFAR-10 48s vs ImageNet 11,000s to reach ~50% accuracy shows… | ||
| The Power of Data Culture as a Precondition for a Smart Utilities Organization | Christiana Von floto, Tatiana vibosh | 22m |
Argument: data culture is a precondition, not a checkbox. The company builds a linked data asset model and pilots appointed data owners for the… | ||
| How to Make Sure Your AI Initiative Fails | Dr Dominik Haitz | 18m |
Argues that enterprise AI succeeds when strategy follows local domain experts (the "foxes"), not blanket buzzwords, and when teams use a product mindset with… | ||
| Computing AI - Obstacles and Implications | Simon Knowles | 19m |
Simon Knowles | Co-Founder, CTO and EVP Engineering, Graphcore
More about Big-Data.AI Summit: https://www.big-data.ai/ | ||
| Goodbye | Merle Uhl | 01m |
Closing remarks that list sponsors (Accenture, Google, PWC, Snowplow and others), thank partners and media collaborators, and ask attendees to submit a… | ||
| nothing matches | ||