PhD Research Fellow in Machine Learning for Cognitive Neuroscience in Oslo

ELLIS; Oslo; Hybrid; NOK 540,000 - 660,000; Full time

Job summary

ELLIS in Oslo invites applications for a PhD Research Fellow in Machine Learning for Cognitive Neuroscience. The role focuses on building ML models to analyze time series data and understand how the brain encodes information.

The project bridges neurophysiology and algorithm design, employing transformer architectures on neural recordings from rodents and humans, in collaboration with DS B Group and IFI.

Qualifications

• PhD candidate with strong ML background and interest in neuroscience.
• Experience with time series modeling and neural data preferred.
• Ability to work across disciplines and publish results.

Responsibilities

• Develop ML models and frameworks for time series analysis of neural data.
• Implement transformer-based architectures for rodent and human datasets.
• Collaborate with neuroscience and ML groups to interpret results.

Skills

Machine Learning

Time series analysis

Neural networks

Transformer architectures

Education

PhD in a related field

Tools

Python

PyTorch

TensorFlow

Job description

PhD Research Fellow in Machine Learning for Cognitive Neuroscience in Oslo

We are seeking an ambitious candidate to develop Machine Learning models and frameworks for time series analysis, aimed at understanding how the human brain encodes information.

This cross-disciplinary project is a high-level collaboration between the Digital Signal Processing Group (DSB), at the Section for Machine Learning (IFI), and the Group for Cognitive Neurophysiology (Medical Faculty). The team has research links with Bradley Voytek at the Halıcıoğlu Data Science Institute (University of California San Diego, USA). By bridging experimental neurophysiology with advanced algorithmic design, we aim to significantly enhance the understanding of high-dimensional neural activity patterns.

The successful candidate will work with open available datasets obtained in rodents and unique datasets of neural activity. Your primary focus will be to design new learning frameworks and neural network architectures to advance our fundamental understanding of how the human brain forms perception and memories. In detail, you will use transformer architectures to analyze time series of local field potentials recorded in rodents and compare performance to time series of action potentials recorded in the same animals (the state of the art); apply and interpret the architecture to local field potential data recorded in humans who have seen a vast number of images from the CoCo-database (https://cocodataset.org); and apply and interpret the architecture to local field potential data recorded in humans who have seen movies.

The data for the project is already collected and available. Project leaders are Adin Ramirez Rivera (DSB, IFI) and Jørgen Sugar (Medical Faculty).