devAIce® Web API 4.1.0 Update

Web API Update Header presenting a Smartphone and Laptop with download symbol

This latest release of devAIceⓇ Web API introduces updated dimensional and categorical emotion models in the Emotion (Large) module. In benchmarks, the new versions of these models are shown to be significantly more robust against background noises and different recording conditions than the previous models, all while keeping the computational complexity of the models unchanged.

devAIce® SDK 3.7.0 Update

devAIce Update Header presenting devices like smartwatch, headphones and VR glasses

Today, we are happy to announce the public release of devAIceⓇ SDK 3.7.0. This update comes with several noteworthy model updates for emotion and age recognition, the deprecation of the Sentiment module, as well as numerous other minor tweaks, improvements and fixes.

devAIce Web API 4.0.0 Update

2022 11 08 devAIceUpdateWEBAPI blogheader

We are proud to announce version 4.0.0 as a major update to devAIce Web API that is available to customers today. Most notably, this release introduces a modernized and simplified set of new API endpoints, all-new client libraries with support for more programming languages, OpenAPI compatibility, as well as an enhanced command-line interface tool. It also includes recent model updates and performance improvements from the latest devAIce SDK release, i.e. support for the Dominance emotion dimension and accuracy improvements of up to 15 percentage points. 

devAIce® SDK 3.6.1 Update

Emotion dimensions cube: valence, arousal, dominance

The devAIce® team is proud to announce the availability of devAIce SDK 3.6.1 which comes with a number of major enhancements, exciting new functionality and smaller fixes since the last publicly announced version, 3.4.0. This blog post summarizes the most important changes that have been introduced in devAIce® SDK since then.

Closing the Valence Gap in Emotion Recognition

Julia with headphones

2021 has been an exciting year for our researches working on the recognition of emotions from speech. Benefiting from the recent advances in transformer-based architectures, we have for the first time built models that predict valence with a similar high precision as arousal.

Contrastive Analysis: Baerbock, Laschet & Scholz

Baerbock Laschet und Scholz jeweils mit Foto und Oszillogramm

The german politicians Annalena Baerbock, Armin Laschet and Olaf Scholz had been analyzed by audEERING’s Audio AI while their chancellor candidation. audEERING’s analyzes do not refer to the content of the speeches, but to acoustic, linguistic and emotional characteristics. These are identified by using scientific procedures and audEERING’s award-winning AI technology devAIce ™.