Jun 01 2015
AX Semantics introduces a new era for technology with semantic software
In the media industry, the future has already arrived. Machines are already producing text drafts for editors and providing a weight off their shoulders, text variety and quality. The software taps into databases and, with the help of an algorithm, structures all relevant information into a news feed. The Industry 4.0 concept also functions along with the Media 4.0 process - man and machine working “hand in hand”. Both complement one another and profit from each other’s strengths. What may seem like pipe dreams of the future actually plays an increasingly important role for many editors already. A number of media companies have been using equivalent solutions for some time, as well as the AX Semantics software. With the release of the software to a limited number of developers, Stuttgart-based Media 4.0 has introduced a new era of machine-assisted news production.
The Henry Ford of text production
“If we were to ask people what they want in the media industry, they’d probably say more editors,” states CEO Saim Rolf Alkan, alluding to the words of Henry Ford. The reality of which is quite the opposite: job losses, downsizing and merging editorial offices, traditional newspapers and media outlets vanishing into thin air paint the picture at present. Publishers and media producers are desperately seeking out new ideas, which on the one hand should maintain or improve the quality of their publications, but on the other hand also make editorial processes more efficient and lighten the load on editors so that they have more time for research and to find better stories. This does not seem to make sense at first glance, but in the age of big data and automated processes in the so-called Media 4.0 process, it can be realized with intelligent, self-learning tools. “The aim of our solution is to become the Henry Ford of text writers and make text production effective and scalable,” says Alkan, describing his personal aspiration.
The Media 4.0 process
The starting point for Media 4.0 is a host of data sources. They provide structured information such as weather data, sports results, stock data and trading figures. AX Semantics then sorts this data and filters it. Here the software identifies structures in the data by using self-learning data mining. An article is automatically formulated from the relevant assembled data. Editorial offices can therefore provide a layout and design, so that at the end the scope and structure (title, sub-titles, number of paragraphs or article length) is suitable for its intended media format. Using these methods, AX Semantics is currently able to produce texts in 11 different languages. The finishing touches are added by journalists, which means the text will be improved with additional information, pictures and graphs, if necessary. In addition, the text is proofread and prepared for publication. This can be carried out simultaneously or according to a pre-defined schedule, written in several languages, tailored to each media channel and prepared for a variety of media output formats.
What does this machine offer editors and media?
Media 4.0 aims to relieve the load on editors’ shoulders. The machine will take over routine tasks such as writing weather or traffic reports. The result is target group-oriented, comprehensible, readable text. “These days, automatically generated texts are hardly distinguishable from texts written by people. When comparing sports and financial articles written by people, they’re sometimes even perceived as a bit more objective and easier to understand,” says Mario Haim, a scientific researcher at the Department of Communication Studies and Media Research at LMU Munich. This frees up editors to work on more qualitative journalistic output. It also saves research time. The software inspects and filters large databases relating to the most important events. The texts are then prepared and edited by an editor. Even niche topics can be better attended to: articles for smaller target groups that were previously not profitable enough to be written, can now be produced because the costs are lower. Creating texts for any reader by individualization and localization of millions of items opens the door to new sources of revenue. The option of generating texts in many languages also expands reach and makes it possible to attract more advertising revenues.
Further development with the support of editors
The German provider of sports news, sid (Sport-Informations-Dienst), as with a number of other large media companies and publishers, is already using AX Semantics; sid produces announcements for sports events. This has put the process of Media 4.0 into top gear. AX Semantics is now offering its ATML3 platform in a closed beta version. This means a selected group of programmers have the chance to test and customize the software. “We want to work hand in hand with editors to cover aspects like data sources and formats. The same also applies to our semantic software, which is also developing along these lines,” explains Frank Feulner, who is responsible for ongoing development of the AX Semantics software. The project team is still accepting applications.
Great interest in the United States
As is often the case with technology, the US is again playing a pioneering role when it comes to Media 4.0 and robot journalism. The birth of robot-assisted journalism can be traced back to last year, when the Los Angeles Times used a programmed bot to report on an earthquake practically in real time. AX Semantics was, at that time, also able to inspire many publishers and investors. Since AX Semantics’ appearance as one of only four German companies to take part in the START program at the Collision Conference, one of the biggest conferences on digital business in Las Vegas, the interest in the solution with proficiency in 11 languages has become extraordinary.
The following infographic illustrates the Media 4.0 process: