The web as we know it has reached its limits due to a lack of semantics

June 28, 2018 by Stefan Summesberger

Andreas is one of the founding fathers of SEMANTiCS Conference and the CEO and co-founder of the Semantic Web Company. He holds a degree in computer science and is a distinguished Semantic Web expert. Also, he is the ‘product architect’ of PoolParty Semantic Suite and responsible for all sales activities of the company. In this interview we talk about challenges in the data lifecycle, semantic AI, SEMANTiCS' evolution as well as the conference's specific offerings for CIOs and much more.

The power of data to drive business success is undoubted. However, management expectations and project execution often disperge. Which part of the data lifecycle is most challenging for companies?

What most frequently is missing, is the overall data and information strategy with a focus on data quality and governance. Still, not many enterprises treat data as a first class citizen from the very beginning of the data life cycle, as a valuable asset on its own. All over the places a new species of ‚magicians‘, aka ‘data scientists’ are asked to step in to overcome this miserable situation. They are asked to do kind of ‘semantic reengineering’ at the very end of the data life cycle, bringing data back to life when fished out of the data lake. This works to a certain degree for isolated use cases, nevertheless it doesn‘t support integrated data governance across the growing numbers and types of data assets and governance needs. As quoted by SEMANTiCS keynote speaker Ivo Willems: „When you automate a mess, you get an automated mess“.

How can companies start to implement Big Data or Artificial IntelIigence technologies?

A broader acceptance for AI technologies can only be developed when results are good enough to either automate an existing process step or when whole new processes can be established due to the higher level of automatisation, which is only made possible through AI. This challenge is strongly related to HR and the skills of people involved, who usually come from different departments starting to work on AI projects. AI is not at all a technological issue only! Bootstrapping AI projects in companies means on one side to develop various skills, on the other side a strategy must be developed that tackles the fundamental question “make, buy, or outsource?”. Obviously little steps with many iterations at the beginning of an AI initiative will overcome this challenge. At SWC, we offer the ‘Semantic Web Starter Kit’ for this purpose, which in many cases has been transferred successfully into productive projects later.

Artificial Intelligence, Machine Learning, Data Science - these are all trends that are currently at the focus of public attention. What comes next?

In the immediate future, I cannot see anything on top of that but rather a fusion of all of that. This also means that the focus will be set on the whole data life cycle. Organisations will develop greater awareness that the often-quoted data-driven business can only succeed if data and information management can be established as a core process rather than an annoying appendage to the ‘actual business processes’.

How does the Semantic Web and Artificial Intelligence relate to each other?

Currently Machine Learning and Artificial Intelligence are frequently used synonymously. True is that AI has developed into two main branches, which are ‘Symbolic AI’ and ‘Statistical AI’. Semantic Web is based on the approach mentioned first, while most ML techniques such as Deep Learning are based on the second version. Over the past months we have seen promising developments into a new kind of AI that we call ‘Semantic AI’. Recently a team of researchers at Free University of Amsterdam has published a paper that will guide us the way towards a fully developed ‘Semantic AI’: The Knowledge Graph as the Default Data Model for Machine Learning. The main idea is not to use single and isolated input data, typically a CSV file, to feed the ML algorithms, instead using an integrated and linked data set based on a more expressive semantic data model.

You are the co-founder of the Semantic Web Company and acknowledged as pioneer in Semantic Web technologies. Why did you decide to focus on these technologies in such an early stage and what drives you today?

It’s one main goal that has driven me from day 1: only data and information that reflects well enough the complexity you can find within most knowledge domains will help us to understand and make better informed decisions. In the days of fake news we can see very clearly why information should be better linked to some curated knowledge bases serving as the ‘ground truth’. As a customer it’s about time to call for more transparency in supply chains to better understand what’s inside of all those shiny product packages. The web as we know it has reached its limits due to a lack of semantics. It’s not only about finding quickly something that is related to a search term, it’s about getting informed and being able to learn from available resources efficiently.

You are the co-founder of the SEMANTiCS. This year the conference takes place for the 14th time in Vienna. Why should CIOs attend the conference?

CIOs should know that semantic technologies are an important cornerstone of an enterprise AI architecture. SEMANTiCS is exactly the right place to meet practitioners and researchers who are top experts in their fields. CIOs will get a comprehensive overview over various fields of applications: Process Automation, Enhanced Customer Experience, Knowledge Management, Information Management, Knowledge Engineering, and Agile Data Integration.

14 years of SEMANTiCS. What are your thoughts on the evolution of the conference?

It’s great to see how the conference and the whole topic has evolved over the years. In 2005, it was quite a manageable number of people involved in semantic web activities. In the meantime, thousands of people all across the globe deal with various aspects of semantics. Somewhere around the year 2010 more and more topics related to text mining got covered while we got connected with the NLP community. Two years ago we have already started to strive to fuse our approaches and technologies with the machine learning and data science community, this year even more of ‘Semantic AI’ will be on the agenda.

Which speaker are you looking forward to hear at SEMANTiCS 2018?

Of course any of the keynote speakers will contribute with sharing some of their latest insights, and I am looking forward to all of them. Nevertheless, it’s sometimes a talk in a break, or a initially inconspicuous poster somewhere in a corner that makes a change. So be awake when you’re around at SEMANTiCS and help to shape the future!

Discuss the potential of semantic AI with with Andreas at SEMANTiCS 2018. Register NOW!

About SEMANTiCS The annual SEMANTiCS conference is the meeting place for professionals who make semantic computing work, and understand its benefits and know its limitations. Every year, SEMANTiCS attracts information managers, IT-architects, software engineers, and researchers, from organisations ranging from NPOs, universities, public administrations to the largest companies in the world.

This interview has originally been published in german on 6-22-2018 by BigData-Insider.