​

At the Department of Computer Science (IDI), we have a vacancy for a postdoctoral fellowship position in Artificial Intelligence for developing and building trust in deep learning models applied to the operation of the power grid

This project is a part of NorwAI and the work will be performed in close cooperation with Statnett, the Norwegian transmission system operator.

The power sector is in rapid development. The green energy transition introduces large amount of power from renewable and intermittent energy sources such as wind and solar. This leads to large, rapid fluctuations in the system, and a need for predictive machine learning models for safe and efficient operation. In particular, Statnett increasingly relies on accurate forecasts on many different timescales.

Deep learning, and in particular transformer models, have been successfully applied to a broad range of problems. These models are particularly successful in large language models but are also increasingly applied to time series forecasting. However, their application remains limited to industrial problems, partly due to limited availability of training data, and partly due to lack of transparency and explainability.

The objective of this position is to further the uptake of deep learning in industrial processes, particularly in the operation of critical infrastructure such as the power grid. The focus will be on methods for building model trust by adding explainability and quantifying uncertainty, as well as studying ways to exploit domain knowledge in combination with observational data, for instance by data augmentation through simulation or by incorporating expert knowledge into the models (e.g. through the use of hybrid models or deep probabilistic models).

The use of transformers for time series forecasting is a natural first application, but other deep learning architectures and other applications such as anomaly detection are also within the scope of this position. In particular, we are interested in explaining anomalies to reduce the number of false positives.

Application deadline: 2024-08-15

Read more and apply

​

At the Department of Computer Science (IDI), we have a vacancy for a PhD position in Fact Checking/Information Verification, the position is a part of NorwAI.

Fact checking or information verification is the basis for many pressing real-world problems such as fake news detection. False information that spreads very fast as news is a growing problem with serious real-world consequences. With the advancements of large language models (LLMs), this issue has become even more important. LLMs have proved themselves very useful in certain tasks and their usage in many areas is increasing. However, these models are not very reliable in means of the truthfulness of the information they generate. For whatever purpose it is, the importance of fact checking or information verification is increasing. External knowledge sources, such as knowledge graphs (KGs) can be used for fact checking purposes, also in combination with LLMs. However, this process has its challenges such as KGs being the bottleneck and not being novel enough in certain domains. In this position, the PhD candidate will work towards fact checking/information verification where one of the important use cases is trustworthy LLMs. The candidate might explore the use and integration of knowledge graphs and other external knowledge resources with LLMs as well as using the fake news domain as a use-case.

Application deadline: 2024-08-15

Read more and apply