Winter School 2021

The first winter school of the KnowGraphs takes place on February 23rd – 25th, 2021. The goals of this school are:

  1. Disciplinary Training
    • Topic: Scientific foundations
    • Approach: Beginner talks on techniques related to KGs
  2. Complementary Training
    • Topic: Research methodology
    • Approach: Data challenge
  3. Professional Development
    • Topic: Possible professional paths in science and industry, IPR
    • Approach: Industry talks + talk on data and law

Programme

Details of the single talks will be added soon.

What are knowledge graphs?

Presenter: Prof. Dr. Sören Auer

Abstract: The availability of large-scale datasets has unleashed an enormous potential of making computers smarter and gave rise to cognitive computing. In order to realize the potential of AI a common understanding of the structure and meaning of data (e.g. to be used as training or evaluation data) must be established. We can leverage vocabularies, Linked Data, knowledge graphs or Semantic Data Lakes for that purpose. In this talk we give an overview on knowledge graphs, which all help to realize the emerging concept of hybrid AI, where large-scale, rich semantic data and knowledge tightly interacts with machine learning and analytics. We discuss some enterprise applications and present with the Open Research Knowledge Graph a use case and approach for semantically describing and organizing scientific contributions to empower researchers to master the flood of staticscientific publications.

Application talk 1: Intelligent Content Extraction from Documents using Machine Learning and Knowledge Graphs

Presenter: Dr. Martin Voigt

Abstract: Artificial Intelligence (AI) is implemented in business basically in a project-driven manner. Due to unclear and wrong expectation on customer-side, due to the lack of joint knowledge in various fields like data management, machine & deep learning (ML, DL) or DevOps to deploy the solutions, many of the projects fail, are in-efficient and never made it into production. The only reasonable approach to implement AI in businesses, especially SMEs, is to provide tailored AI-based products. These illustrates fast the benefits, are integrated in enterprise IT environments in short time frames without required AI-knowledge at customer-side and deliver business value immediately. In this talk, we give a technical overview of our ®Cognitive Business Robotics (CBR) platform and focus on the product CBR Secretary, which allows for content extraction from various kinds of documents. Therefore, it facilities technologies from the ML areas of Natural Language Processing and Computer Vision as well from Knowledge Graphs. We will emphasize on the separation of concerns in different points of view which enables to easy product integration with various customers and domains.

Introduction to Description Logics – to classical reasoning and beyond

Presenter: PD. Dr.-Ing. habil. Anni-Yasmin Turhan

Abstract: Description Logics (DLs) are an intensively studied family of logics tailored towards building ontologies and reasoning over these. Most DLs are decidable fragments of first-order logic and are the basis for most OWL profiles. The formal semantics of DLs allow to define reasoning services that, in turn facilitate the extraction of implicit knowledge from ontologies. In this course we introduce the building blocks of DL ontologies and the fundamental DL reasoning services, such as subsumption and query answering. The second half of the course gives an overview on inferences for maintaining ontologies and on reasoning under non-classical semantics that offer resilience against inconsistencies.

Knowledge Graphs in Natural Language Processing

Presenter: Prof. Dr. Roberto Navigli

Abstract: This session will be an overview of Knowledge Graphs in Natural Language Processing (NLP) and particularly in the areas of word-level and sentence-level semantics, as well as multilinguality. At the lexical level, we will analyze the structure of historical resources like WordNet and then move on to more recent endeavors like BabelNet (and how to build such resources from unstructured information). At the sentence level, we will overview current resources in the area, including VerbAtlas, which is strongly connected to WordNet and BabelNet, and the Abstract Meaning Representation formalism. We will survey the use of KGs within recent NLP tasks, such as Word Sense Disambiguation, Semantic Role Labeling and Semantic Parsing.

Application talk 2: PoolParty Semantic Suite and Knowledge Graph Management

Presenter: Martin Kaltenböck and Artem Revenko

Abstract: The amount of structured, semi-structured and unstructured data in organisations is growing heavily and data and information is often stored in silos inside of department solutions, but organisations are interested in unified 360 degree views on their data and information to be able to make precise and sustainable decisions. Enterprise Knowledge Graphs can solve this problem by integrating data and information from heterogeneous sources in different formats by mapping it to one or more central knowledge models and linking it to each other in the form of a Knowledge Graph. This talk will provide a short introduction into the problem statement and why and how Knowledge Graphs can help, furthermore it gives a short introduction into PoolParty Semantic Suite, a Semantic Middleware for powerful Knowledge Graph applications and finally it will provide an overview of the most important use cases that Enterprise Knowledge Graphs can solve.

Further reading:

Machine Learning on KGs

Presenter: Prof. Dr. Steffen Staab

Data ownership, control and access

Presenter: Asst. Prof. Dr. Sabrina Kirrane

Abstract: The goal of this talk is to introduce the audience to the topic of data ownership, control and access. We will start by discussing different types of constraints, for instance access policies, usage control, and regulatory requirements. Following on from we will examine how policy languages can be used for policy specification, enforcement, and governance. Finally we will examine initiatives, such as Digi.me and Solid, that aim to give individuals and organisations more control over their data.

Application talk 3: FAIR data applications using Knowledge Graphs

Presenter: Dr. Peter Haase

Abstract: This talk will highlight our lessons learned and best practices for building Knowledge Graph-driven FAIR applications to drive digital transformation. We will focus on agile data management and governance with Knowledge Graphs, following the FAIR Data principles. We will cover topics such as data catalogues, ontology modelling, taxonomy integration, user experience development, and how low-code platforms enable the rapid delivery of value.

Legal Aspects of KGs

Presenter: Prof. Dr. Jeanne Pia Mifsud Bonnici

Practical Information

The winter school will take place online using the BigBlueButton platform. Registered participants will receive a link to the room via mail before the winter school starts.

People external to the KnowGraphs project can participate in the winter school. Please contact Michael Röder before February 20th via mail for registration.

Schedule

All times are in CET.

Day 1

9:00 – 9:30
Introduction
9:30 – 11:30
What are knowledge graphs?
Prof. Dr. Sören Auer
11:30 – 11:45
Coffee break ☕
11:45 – 12:30
Application talk 1: Intelligent Content Extraction from Documents using Machine Learning and Knowledge Graphs
Dr. Martin Voigt
12:30 – 13:30
Lunch 🍔
13:30 – 15:30
Introduction to Description Logics – to classical reasoning and beyond
PD. Dr.-Ing. habil. Anni-Yasmin Turhan
15:30 – 15:45
Coffee break ☕
15:45 – 18:00
Data Challenge 💻
  • Group formation (3 people per group)
  • Setup for benchmarking
  • Questions to helpers

Day 2

9:00 – 9:30
Introduction
9:30 – 11:30
Knowledge Graphs in Natural Language Processing
Prof. Dr. Roberto Navigli
11:30 – 11:45
Coffee break ☕
11:45 – 12:30
Application talk 2: PoolParty Semantic Suite and Knowledge Graph Management
Martin Kaltenböck and Artem Revenko
12:30 – 13:30
Lunch 🍝
13:30 – 15:30
Machine Learning on KGs
Prof. Dr. Steffen Staab
15:30 – 15:45
Coffee break ☕
15:45 – 18:00
Data Challenge 💻

Day 3

9:00 – 9:30
Introduction
9:30 – 11:30
Data ownership, control and access
Asst. Prof. Dr. Sabrina Kirrane
11:30 – 11:45
Coffee break ☕
11:45 – 12:30
Application talk 3: FAIR data applications using Knowledge Graphs
Dr. Peter Haase
12:30 – 13:15
Lunch 🍜
13:15 – 15:00
15:30 – 16:00
Coffee break ☕
16:00 – 17:30
Data Challenge Presentations
17:30 – 17:59
Wrap up