We interrupt this blog for a very important announcement: Faction XYZ has joined the MARCONI consortium! We are very excited to be working together with this applied artificial intelligence engineering service provider and tackle new challenges with their language processing services. Introductions are in order: meet Faction XYZ!
Welcome to the team! Could you please introduce your organisation to our readers?
Being founded in 2016 by veterans in the Computer Science and Artificial Intelligence industry, Faction XYZ is a team of 25 experts in deep learning with a focus on text and signal processing.
Faction XYZ as your A.I. partner
Enjoying the trust of dozens of global companies, including global players in the banking, telecom and automotive industry, we apply state of the art deep learning algorithms in order to automate business processes or support new revenue streams.
Applying data science for business value
At Faction XYZ, we apply A.I. algorithms to provide business value. As such, we take on a pragmatic, agile and application-oriented approach to solve problems that really matter to our client's business.
Data science meets IT Engineering
Our unique blend of data science and IT engineering makes us a solid partner to companies that require assistance in streaming data platforms, producing deep learning algorithms at scale, and production-oriented solutions.
You can find out more on www.faction.xyz
Who will work on the MARCONI project from your organisation?
From our side, it will be mostly Aleksandra Vercauteren, our senior NLU Engineer, who will be responsible for the MARCONI project. Aleksandra started working for Faction XYZ in November 2017. She focuses on the development and continuous improvement of NLP algorithms.
Of course, she can also count on the full support of our Machine Learning team at Faction XYZ, which consists of experts in different domains that are relevant to machine learning and Natural Language Processing.
What specific part of the project will you be working on?
We will take care of several language-related aspects of the project, such as the chatbots. Our Named Entity Recognition and Topic Detection algorithms will make the life of radio makers easier as it will allow them to filter through messages by selecting a certain topic. In a later stage, we will also implement sentiment analysis of the messages that listeners send in or post, so that radio makers are able to easily detect what makes their audience happy!
Which new technologies or applications are you bringing into the project?
We are providing the chatbot platform Chatlayer (www.chatlayer.ai), which allows users to easily "build" their own chatbots. Radio stations can provide a list of sentences and their corresponding intent, based on what they know listeners will often say or ask. For instance, think of sentences like "I would like to hear song X" or "Can you play song X" or "I want to surprise my grandmother who really likes song X, so can you please play it in this afternoon show?", all three of which have a "request song" intent. The language model underlying the chatbot platform is trained with the data provided by the radio station. The chatbot that makes use of the model will be able to correctly "interpret" sentences similar to the ones in the training data set. The sentences above will be recognized as song requests, and the title of the song will automatically be extracted and, for instance, automatically added to a playlist. Finally, the radio station can create a dialogue flow on the chatbot platform that defines what needs to be done with each intent: reply to it, ask further questions, etc. Our chatbot platform will enable radio stations to give their audience the attention they deserve, with personalised and real time responses to their messages.
“It is so exciting to be part of a project that will reinvent radio, a medium that has been around for about a century, filling the gap between a one-directional and an interactive way of making radio.”
Apart from the chatbot platform, we will also provide some other language processing services that will make the life of radio makers easier. Nowadays, they are dealing with hundreds of unfiltered messages and incoming posts all the time, and it is hard to figure out which of these messages might be interesting to include in the show. We will provide Named Entity Recognition and Topic extraction services that will contribute to a more manageable incoming message stream, making interactive radio so much easier!
Why did you join MARCONI?
Faction XYZ is always looking for challenging projects to put new technologies and innovations to work. The MARCONI project has an ambitious but relevant goal to improve the overall interaction between radio stations and the audience. This will not only change the audience experience, but will most of all be quite revolutionary in the way radio stations will operate and stay relevant in the future.
Shaping these kinds of innovations that could be real game changers, is what we love to do, and what we do best.
“Shaping these kinds of innovations that could be real game changers, is what we love to do, and what we do best.”
What are you most looking forward to within the project?
It is so exciting to be part of a project that will reinvent radio, a medium that has been around for about a century, filling the gap between a one-directional and an interactive way of making radio. Radio making has become more and more interactive in the past decades, and it will be interesting for us to see how new technologies can be effectively used in a changing world. We are very curious to find out how both radio makers and listeners will react to this new radio experience!
Thank you and good luck!
