Ultimately it all leads up to industrial innovations
Ultimately it all leads up to industrial innovations
Innovation Director
Being an SFI, the focus of NorwAI is research-based innovation on data-driven AI. Our main objective is to develop theories, methods and technologies for successful and responsible exploitation of data-driven artificial intelligence that ultimately leads to industrial innovations. This work package aims at coordinating all innovation activities in the research center.
The main research problems to be addressed by NorwAI are motivated and identified by real industrial needs, as evident from the following set of innovation areas:
AI Innovation in Media and Finance – with a focus on:
User modeling and behavior prediction
User data, collected and analyzed, using provable privacy-preserving and unbiased techniques, is fundamental to understand the user, and thereby provide relevant, personalized services and adapt well to a diverse set of user groups. From an industrial perspective, user modelling is
important to, e.g., improve customer satisfaction and reduce churn. Also, successful user models
and behavior predictions will enable the industry to provide improved recommendations and other personalized services and products.
Personalized and Contextualized content creation
Providing personalized and context-aware content is important for the user experience, e.g., by offering personalized advice and services. Furthermore, in many cases the provided content should be adapted to the user’s context. For text data, this amounts to providing relevant
text summarizations relevant for a user in her context.
These AI Innovation areas are linked to the Research areas of AI for Personalization and AI for Language Technologies.
AI Innovation in Energy, Maritime and Telecom with a focus on:
Predictive maintenance and operational availability
Maintenance is important for efficient operation and to avoid unplanned downtime. Predictive maintenance utilizes real-time data to optimize maintenance scheduling. Operational availability extends traditional system availability to predictive resource allocation.
IoT sensor anomaly detection
The availability of low-cost sensors for collecting data is instrumenting the physical world. Sensor data can, e.g., be used for automation and decision support. In this IA, the focus is on innovations on anomaly detection.
Hybrid digital twins
A digital twin is virtual representation of a physical system, fed with sensor information in real time. NorwAI will innovate on the use of hybrid AI techniques to handle scarce or low-quality sensor data These are linked to the Research areas of AI for Streaming & Sensor-based Data, and Hybrid AI Analytics.
Cross domain AI Innovations on Explainable AI and Data Quality – with a focus on:
Explainability and verification
Explainability and AI verification aim at remedying problems with lack of transparency and interpretability, lack of robustness, and inability to generalize to situations beyond their past experiences by developing methods for understanding how black-box models make their predictions and what are their limitations. The call for such solutions comes from the research community, the industry and high-level policy makers, who are concerned about the impact of deploying AI systems to the real world in terms of efficiency, safety, and respect for human rights.
Data quality analysis and enhancement
High quality data is the fuel of data-driven AI. NorwAI will study how data quality can be measured and assured in order to enable critical decisions to be made based on AI models.
Furthermore, data enhancement techniques will be used to increase data quality.
These are linked to the Research areas of AI in Society, Trustworthy AI and Data and Platform
for AI.