post_me_platform.png
post_me_platform.png

Saying Goodbye to MixedEmotions (in a Happy Mood)

Using emotion detection to enable visionary applications–that's what the MixedEmotions project was all about. Here's a quick review of the EU H2020 innovation action that was successfully completed in May 2017.

The Mixed Emotions consortium consisted of a diverse mix of interesting players: Universities (NUI Galway, UPM, Universität Passau, Brno University of Technology), IT companies (Expert System, Sindice Tech /SIREn SOLUTIONS, Phonexia), consultants (Paradigma), and – you’ve guessed it – DW Innovation.

post_me_consortium.png
The Mixed Emotions project consortium

MixedEmotions thus came up with an interesting mix of use cases and solutions. The consortium worked on:

  • brand reputation (mission: find out what customers say/write about a specific company on the web, i.e: what kind of emotions they express)
  • call center management (mission: use smart software to support an agent when the talk gets heated)
  • social TV (mission: recommend just the right program for a specific user)

At their core, all these services rely on emotion recognition in audio, video, text.

Use case #3 continues to be particularly important to us: As we're writing this, DW developers are working on implementing MixedEmotions tech into our official smart TV app.

Besides developing the three prototypes mentioned above, the Mixed Emotions project also established an Open Source Plattform for Big Data emotion extraction:

"The MixedEmotions platform is a big data platform for recognition of emotion and other subjective features from multiple content modalities. The platform includes modules for emotion recognition in text, audio and video, for entity recognition and linking, for social media data collection and social network analysis as well as an orchestrator for building custom data flow configurations and analysis mashups. It follows a microservices architecture built around Docker and Mesos allowing stand-alone module execution as well as scalable and flexible deployment on the cloud and easy incorporation with other big data analysis pipelines."

For those who want to learn about the platform in detail, the consortium has created a comprehensive webinar:

There’s also a GitHub repository that gives you access to all the code you need to get started in the world of emotion analysis.

After almost two years of serving the project as developers, communicators, and marketers, it certainly isn’t easy to say goodbye, but DW Innovation does so in a happy mood: We were part of a great international team, we learned a lot about big data and emotion detection, we helped create actual solutions to actual problems – and we took a part of the project’s tech to innovate DW.

Authors
team_alexander_plaum.jpg
Alexander Plaum
team_kay_macquarrie.jpg
Kay Macquarrie