DSC DACH 2023
| title | author | length |
|---|---|---|
| Open Data vs. Privacy - How Synthetic Data Can Democratize AI Innovation | Wolfgang Weidinger | 54m |
Defines synthetic data as using generative AI with privacy mechanisms to learn patterns from sensitive records and create artificial records that are… | ||
| Scaling & Industrialization of AI | Lukas Kölbl | 19m |
Survey finds 84% of executives say scaling AI is necessary while 76% can build pilots but fail to scale, with banking and insurance surprisingly behind despite… | ||
| Creating Production-Ready Shiny Apps | Mario Annau | 01h 32m |
Turns a vanilla Old Faithful Shiny app into a proper R package with documented exports and zero R CMD check errors, uses usethis and roxygen2 to generate… | ||
| Go with the flow–Track your machine learning lifecycle | Philipp Danninger | 01h 04m |
Uses MLflow to version experiments on the California housing dataset, running a baseline Linear Regression and an XGBoost pipeline while logging… | ||
| Graph Neural Networks in Drug Discovery | Tin Tunjic | 01h 23m |
Compares a fingerprint multitask classifier (ECFP, 1,024 bits) to graph approaches on Tox21, 7,800 compounds, 12 tasks, using DeepChem + RDKit; fingerprint… | ||
| The Modern Data Stack | Bogdan Pirvu | 01h 35m |
Claim: the modern data stack replaces Hadoop/Spark ops with cloud services so analysts become analytics engineers. Demo builds a working pipeline: MongoDB +… | ||
| Learnings integrating a machine learning model to existing software | Clemens Zauchner | 24m |
Built an anomaly-alerting module for Trinity smart alerts that learns daily and weekday patterns from production metrics and replaces brittle thresholds. A… | ||
| Connecting the Dots: Graph Analytics for Economic Sustainability | Diogo Braga | 27m |
Claims that property graph databases let teams traverse multi-source relationships in real time instead of waiting days or weeks, with examples: a 30,000… | ||
| Beyond Deep Learning: Next Generation AI | Danko Nikolic | 51m |
Claims a Guided Transfer Learning method that cuts training data needs by 3–5 orders of magnitude, based on a neuro-inspired architecture that inserts gates… | ||
| Toxic management | Josip Saban | 34m |
Research based on 15–16 interviews and qualitative analysis across about 190 companies surveyed claims toxic managers use slow agendas, favoritism… | ||
| Lyrics Generator + Results: AI Usecases for climate change | Felix Stockhammer | 28m |
In the Lumos Lyrics Generator project, we learned how to build a text generator from scratch. Using existing lyrics as training data, our model can write a… | ||
| LLM-Based Development | Bogdan Pirvu | 29m |
A prototype booking app wires voice transcription to GPT-4 and GPT-3.5 as the decision engine, calling a calendar API to schedule appointments; one demo… | ||
| The building blocks of a successful data strategy | Mario Meir-Huber | 34m |
Many data programs fail because culture and governance lag behind tools; success requires raising three domains — technology governance culture — not just… | ||
| AI Regulation - How to implement AI legally compliant? | Alexandra Cirnau | 30m |
Overview of concrete legal steps for AI projects: treats anonymization and synthetic data as a processing step that requires a legal basis, warns that training… | ||
| Empowering Education and Therapy with Data Science | Nicole Weinert | 20m |
Describes a pipeline that first builds a machine learning emotion classifier that distinguishes six expressions (anger, disgust, fear, happiness, sadness… | ||
| How Deep Learning is Revolutionizing Sports Analytics | Philipp Singer | 42m |
Predicts a rusher’s final yardage at handoff from NFL shoulder‑sensor tracking by encoding 21 players into tensors, applying convolutions and pooling, and… | ||
| Digitalization and Space Tech | Martin Mossler | 55m |
Space sector revenue hits roughly $500 billion by 2030 and most of that value comes from downstream data — open Copernicus imagery lets models spot millimeter… | ||
| Driving data use cases in the insurance business | Andreas Magenschab | 31m |
UNIQA reports running ~40 companies across 18 countries with 16 million customers and a ten‑person group data analytics team that focuses on rapid MVPs… | ||
| How generative AI is Changing the way people interact with computers | Iveta Lohovska | 32m |
Claims that large generative models like ChatGPT/GPT‑4, Whisper and Llama now generate working code, translate across languages and collapse years of study… | ||
| Panel: Data-Driven Sustainability | 30m | |
A coalition of cases shows concrete wins and clear limits: an 80 acres sensor network feeds AI models that cut water, energy and transport waste on European… | ||
| Let's Terminate Climate Change with AI | Stefan Papp | 22m |
Argues that AI and data are applied to three concrete problems: scheduling heavy industrial processes by region-specific emission forecasts (models use wind… | ||
| Geospatial Machine Learning at Scale | Karsten Schroer | 26m |
Amazon SageMaker adds geospatial connectors for Sentinel‑2, Landsat and OpenStreetMap, a notebook SDK to launch managed processing jobs (demo job returns 25… | ||
| Digital Solutions for ESG | Stefan Blachfellner | 27m |
A practical evaluation framework assesses digital products by measuring model energy and hardware use, testing whether a stronger data model actually reduces… | ||
| (Un)Ethical Machines? Why AI Bias Is a Problem and What to Do About It | Alexandra Ebert | 34m |
Examples where models deny women access to credit and label dark‑skinned people as gorillas anchor the claim that bias comes from training data and historical… | ||
| PeaceEye - How state of the art technology helps to save lives | Andreas Papp | 26m |
PeaceEye fuses satellite Earth observation, microwave soil‑moisture algorithms and Geo‑Social Media sentiment analysis to detect hotspots and guide… | ||
| Beyond a Machine Learning Model | Dana Jomar | 27m |
Uses Johns Hopkins data and the CoronaNet dataset to train models that predict country-level COVID case growth rates and then interprets effects with ALE plots… | ||
| Using data to improve physical therapy for people with neurological diseases | Gernot Greimler | 22m |
Collects synchronized sensor streams — 50 Hz motor and joint-angle data, pressure insoles, finger-position encoders, SpO2 and blood pressure — and attaches… | ||
| A simple model to estimate an entities’ contribution to global warming | Stefan Janecek | 27m |
Builds a simple climate model that converts an entity’s emission time series and projected changes into an expected global temperature rise called warming… | ||
| nothing matches | ||