Postdoctoral positions and PhD studentships
Location: Weizmann Institute of Science, Israel
Supervisors: Prof Ilya Kuprov
The Weizmann Institute of Science is, by most research rankings, the top academic institution in Israel and among the top in the world. It is located in a leafy and picturesque suburb of Tel Aviv at the feet of the Judean hills, 15 miles away from the Ben Gurion International Airport. The official language of the Institute is English; it provides furnished apartments for postdocs and PhD students either on campus or across a shopping street from the campus.
Magnetic Resonance is a significant institutional priority at Weizmann. The Institute has dozens of magnetic resonance instruments at its various departments, including 800 MHz and 1 GHz NMR spectrometers with cryoprobes, pulsed EPR spectrometers, and DNP setups. The Institute is home to over a dozen magnetic resonance spectroscopy, imaging, and spin physics research groups. This creates a uniquely vibrant atmosphere; many leading magnetic resonance researchers have either started or spent some time working at the Weizmann Institute.
The following projects are currently available at the PhD (all fees are covered and a stipend is paid) and postdoctoral (four years with a possibility of extension) levels:
Project 1: Quantum Theory of Magnetic Processes
Textbooks claim that computational complexity of quantum dynamics on classical computers is exponential, but recent research has called this into question: the inevitable presence of decoherence makes much of the Hilbert space of large quantum systems dynamically unreachable; simulations in reachable subsets have polynomial complexity scaling. This is particularly visible for spin systems that occur in magnetic resonance spectroscopy and imaging: time-domain simulations of protein NMR experiments involving hundreds of interacting spins are now routine.
Using powerful GPUs and tensor methods, quantum spin dynamics can also be modelled in the presence of classical processes: diffusion, hydrodynamics, non-linear chemical kinetics, etc. This is particularly important in magnetic resonance imaging, toxicology, and metabolomics. This project is about creating theory and software infrastructure that would be able to handle that level of multi-physics complexity. Applications range from magnetic resonance imaging of metabolic processes to geomagnetic navigation of migratory birds.
Project 2: Explainable Artificial Intelligence in Magnetic Resonance
Artificial neural networks are famously opaque – we can usually find out why a neural network gives a particular answer, but finding out exactly how it arrives at that answer is an unsolved problem.
This lack of interpretability (and therefore trust) is a much-criticised feature of deep neural networks. In fully connected nets, the signalling between inner layers is scrambled because backpropagation training does not require perceptrons to be arranged in any particular order. The result is a black box; this problem is particularly severe in scientific computing and digital signal processing.
This project is about designing neural networks and their training databases that remain transparent, or at least translucent, in their operational workflow. We use magnetic resonance spectroscopy and imaging as application areas because those are well understood at the deepest quantum mechanical level and mathematically closely matched to the procedures that artificial neural networks perform.
Project 3: Bosonic Degrees of Freedom in Magnetic Resonance
Quantum devices always come embedded into classical devices: magnets, refrigerators, control circuitry, electromagnetic traps, etc. The processes that take place at the interface between the classical and the quantum part commonly involve bosonic degrees of freedom, for example phonons in crystal lattices, rotational excitations in gas-phase molecules, and photon population numbers in microwave cavities. As the instruments become more sensitive and precise in magnetic resonance spectroscopy and imaging, accurate treatment of these degrees of freedom is becoming essential.
This project is about designing computational modelling infrastructure that takes bosonic degrees of freedom into account in spin spectroscopy and imaging, and exploring the new technologies that become possible as a result. On the software engineering side, the computational complexity of the problem requires the use of latest GPUs and tensor methods. On the applications side, simulation and optimal control infrastructure are essential in the design of new technologies based on spin.
To start the application process, please send a CV to Prof Ilya Kuprov (ilya.kuprov@weizmann.ac.il). The positions are open to applicants from anywhere in the world.
Deadline: March 2025; earlier submissions would be much appreciated for logistical reasons.