PhD Student - Physical modeling via deep neural networks (Closed)

A PhD student is being sought to carry out a project in Deep Neural Networks (DNNs), based in the Dept. of Polymers and Advanced Materials: Physics, Chemistry and Technology, Faculty of Chemistry, University of the Basque Country, San Sebastian, Spain (UPV/EHU) in active collaboration with Dept. of Mechanical Engineering, Jiangsu Key Laboratory for Design and Manufacture of Micro-Nano Biomedical Instruments, Southeast University, Nanjing, China (SEU).

The project is based on a recent finding by the two groups. For physical systems exhibiting a strong spatial dependence (usually as a function of a few relevant experimental parameters) and when the number of images that display the spatial dependence is very reduced (due to excessive cost/time-consumption/legal-constrains/etc), one can directly use the spatial dependence to segment the images into physically meaningful regions, whereafter every pixel from every image will belong to one region (or class/label), thus resulting in a truly large pixel-based dataset with millions of classified pixels, even when obtained from a few images. The key idea is to use this simple piece of information (the region/class/label) as the output for two equally simple inputs: the experimental conditions and the pixel’s location within the image. In this manner, a low-complexity DNN (with a small number of hidden layers and neurons) can be easily trained into understanding the spatial dependence of the system as a function of the experimental parameters. Most importantly, after training, the simple DNN can be used out-of-the-box to generate new segmented images in a pixel-by-pixel manner for unknown experimental conditions (generalization, i.e. interpolation and extrapolation). This generative functionality is remarkable considering the simplicity of the approach (an uncomplicated DNN with spatial-dependent input). Previously, such functionality has been restricted to highly specialized Convolutional Neural Networks (CNNs) with millions of trainable parameters, such as Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs). The low complexity of the DNN enables simple/fast modifications/adaptations (e.g. for other projects or in other fields) as well as the use of less-sophisticated computational resources during both training and execution, leading to moderation in power consumption (an increasingly important aspect) while simultaneously simplifying model deployment on mobile devices with limited resources (wherever this aspect may become relevant).

The project focuses on building physically meaningful models, one for each of various physical systems, so that we can predict their behavior under unseen experimental conditions. We target an improved level of extrapolation outside the training range, which CNNs cannot provide (at least yet). Essentially, the difference here is that between data-driven learning (provided by CNNs) and data-driven discovery of underlying physical behavior (our goal).

• Contract duration: 1 year (possibility to extend up to 3 years)
• Estimated annual gross salary: Salary is commensurate with qualifications and consistent with our pay scales
• Target start date: 2023/07/01

We provide a highly stimulating research environment, and unique professional career development opportunities.

We offer and promote a diverse and inclusive environment and welcomes applicants regardless of age, disability, gender, nationality, ethnicity, religion, sexual orientation or gender identity.

The PhD project is a natural continuation of a recent collaboration between Dr. M. A. Gosalvez (UPV/EHU) and Prof. Y. Xing (SEU). Gosalvez and Xing have worked jointly for approximately 17 years, as reflected by at least 20 co-authored publications in peer reviewed journals and 15 contributions in international conferences. Xing’s group is devoted to both experimental and computational work while Gosalvez’s group is dedicated to theoretical and computational modeling. While based in Gosalvez’s group, the PhD student will have the chance to visit Xing’s group twice during 2023-2024.

Interested candidates should submit an updated CV and a brief statement of interest to the following application email below. 

Reference letters are welcome but not indispensable.

The reference of the specific opening to which the candidate is applying should be stated in the subject line, and the application must be received before the application deadline.

Although candidates are welcome to contact the project supervisors to know further details about the proposed research activity, please be aware that the application will be evaluated only if it is submitted directly to the email address listed below as application email.

• Reference: 2023/68
• Application deadline: 2023/06/19
• Application email: jobs.research@dipc.org

Applications received by the deadline will be evaluated by a Committee designed by the DIPC board on the basis of the following criteria:

• CV of the candidate (60%)
• Adequacy of the candidate’s scientific background to the project (20%)
• Reference letters (10%)
• Other: Diversity in gender, race, nationality, etc. (10%)

Evaluation results will be communicated to the candidates soon after. Positions will only be filled if qualified candidates are found.

The DIPC may revoke its decision if the candidate fails to join by the appointed time, in which case the position will be awarded to the candidate with the next highest score, provided it is above 50 (out of 100).

However, the selected candidate may keep the position if, in the opinion of the Selection Committee, the candidate duly justifies the reasons why he or she cannot join before the specified deadline, and as long as the project allows it.

This project has been supported by MCIN with funding from European Union NextGenerationEU (PRTR-C17.I1)