Four Postdoctoral projects starting in new FNWI Research Priority Area
AI for Sustainable Molecules and Materials
To strengthen FNWI’s focus on smart and green, the Research Focus Area (RPA) “Artificial Intelligence for Sustainable Molecules and Materials” (AI4SMM) has been launched. AI4SMM targets breakthroughs in sustainable innovation with AI and computational methods for molecules and materials to aid in realising the materials transition needed for a sustainable society.
Read also this news item, with a short overview of the four research projects on (1) metamaterials, (2) thermal energy storage with salt hydrates, (3) safe-and-sustainable-by-design plastics, and (4) plant protein mixtures for sustainable food design, and the role that AI can play in this research.
AI4Science Lab
How can we detect, classify, and predict relevant patterns in scientific data if they are hidden within large amount of non-relevant data?
The AI4Science Lab is an initiative supported by the Faculty of Science (FNWI) at the University of Amsterdam and located in the Informatics Institute (IvI) in (LAB42}8. The AI4Science Lab is also connected to AMLAB, the Amsterdam Machine Learning Lab.
We develop and use machine learning techniques to discover patterns in data streams produced by experiments in a wide variety of scientific fields, ranging from ecology to molecular biology and from chemistry to astrophysics.
AI4Science Consortium
The AI4Science Lab is the center of a rapidly growing multidisciplinary consortium of students, researchers, and experts interested in the development and application of artificial intelligence tools for the analysis of scientific data. We keep each other informed about interesting challenges and new developments. Our activities include organising seminars and workshops, a bi-weekly colloquium series and helping each other with acquiring funding for AI4Science spin-offs. Interested in joining the AI4Science consortium?
Projects
The breakthrough discovery of the first gravitational wave signal in September 2015 (Nobel Prize 2017) has opened a new window to the Universe. Analyses of the signal waveforms during the initial inspiral, merger and final ringdown phase provide crucial information about the properties of superheavy stellar objects. An active field of research in astronomy deals with recognising rare features in data streams from space obervations in (almost) real time. This challenging task arises from the massive imaging surveys of the sky carried our at a wide range of wavelengths (optical, radio, X-ray). The purpose is no longer just to find objects that are there all or most of the time, but also to spot so-called ‘transient’ objects that appear only fleetingly. Two-dimensional liquid chromatography (2DLC) is a powerful technique to separate and detect trace molecular compounds in complex samples such as food contaminants, industrial production streams, urine, or blood, to name but a few examples. However, successful implementation requires time-consuming experiment-specific optimization of many parameters. Machine learning can aid with data-analysis and the acceleration of optimization. Gene regulatory networks (GRNs) model the biological interactions between genes and provide a better understanding of the cellular processes and regulation pathways. Correctly constructed GRNs can play a fundamental role in solving various biological and biomedical problems, such as tracking disease development. The great amount of gene expression data available on the GEO database suggests using AI-based techniques for processing and modeling such networks. Internationally operating weather radars are known to capture the mass movement of migrating birds. This offers unparalleled opportunities to quantify bird migration at large spatial and temporal scales and thereby gain insight into the effects of environmental conditions and human activities. However, the field of radar aeroecology is still in its infancy and the potential of available radar data is largely under-utilized. The stability of the antigen is crucial for the development of effective vaccines against highly contagious viruses. Mutating amino acids on spike proteins is one method of developing antigens, but it can be a time-consuming process. Deep learning-based methods have recently shown promising results in protein research, and with the abundance of protein structures available in the Protein Data Bank and AlphaFold2, we can train deep-learning models to suggest mutations and predict protein stability, which could accelerate vaccine research and design.Gravitational waves
Radio astronomy
Liquid Chromatography
Gene Regulatory Networks
Radar Aeroecology
Protein Stabilization for Vaccine Design
