Multiple PhD Positions in Symmetry and Symmetry Breaking in Deep Learning

Laplace Apply; Amsterdam; Ibrido; EUR 36.000 - 45.000; Tempo pieno

Vantaggi offerti da questo lavoro

Full-time appointment

Paid salary

Descrizione del lavoro

University of Amsterdam offers multiple PhD positions in symmetry and symmetry breaking in deep learning, based in Amsterdam, focusing on mathematical and statistical physics with neural network experiments. Applicants should hold a strong master's in mathematics, physics, CS, or AI, and have strong Python skills.

The positions are full-time with a four-year contract and Dutch doctoral salary scale. Funding is EUR 3,204–4,051 gross per month, period per month, with application by 8 December

Competenze

• Strong mathematical background and research programming experience.
• Experience with neural networks and symmetry in ML.
• Proven ability to work independently on theoretical and computational projects.

Mansioni

• Conduct research on symmetry and symmetry breaking in deep learning.
• Test, analyze and validate theoretical models with experiments.
• Collaborate with supervisors and other PhD researchers on publications.

Conoscenze

Python

Neural networks

Statistical physics

Symmetry in ML

Formazione

Master's degree in mathematics/physics/CS/AI

Descrizione del lavoro

Multiple PhD Positions in Symmetry and Symmetry Breaking in Deep Learning

University of Amsterdam

Netherlands

Apply by 08 Dec 2026

About This Opportunity

University of Amsterdam offers Multiple PhD Positions in Symmetry and Symmetry Breaking in Deep Learning in Amsterdam, focused on symmetries, statistical physics and theoretical deep learning.

The University of Amsterdam advertises a coordinated doctoral recruitment call for research on symmetries and their breaking in deep-learning systems. This is one source vacancy advertising multiple funded doctoral places; it is kept as a single record. The projects combine mathematical or statistical-physics ideas with neural-network analysis, generative modelling and experiments. Topics can include invariance, symmetry-breaking mechanisms and implications for models such as diffusion or autoregressive networks; doctoral candidates may use contemporary machine-learning frameworks to evaluate research hypotheses. A strong master's degree in mathematics, physics, computer science, AI or a closely related quantitative discipline with research programming and theoretical preparation is expected. A full-time doctoral employment contract with potential limited teaching duties. Compensation or funding: EUR 3,204-4,051 gross monthly on Dutch PhD employment scale. Stated appointment or programme length: Four years. Application timing: 8 December 2026 at 23:59 Europe/Amsterdam on source application page. Who This Opportunity Is Best Suited For: Applicants who enjoy both mathematical explanation of learning phenomena and practical model-based testing may be well matched; prior numerical experimentation and strong Python skills are useful. Potential Challenges / Who May Find It Less Suitable: This is conceptually demanding theory-led machine learning. Multiple positions are advertised together and should not be split into fabricated source records; final project allocation may vary. Overall Opportunity Assessment: The call creates an interdisciplinary bridge between symmetry concepts in physics and mechanisms of modern deep neural networks.

Symmetries, statistical physics and theoretical deep learning

Level PhD

Time Commitment Full-time

Duration Four years

Field Theoretical Machine Learning

Specialization Computer Science

Organization Type University

Funding & Compensation

Amount EUR 3,204-4,051 / Per Month

Funding Type Salary

Funding Status Paid

Period Per Month

Requirements

Education: A strong master's degree in mathematics, physics, computer science, AI or a closely related quantitative discipline with research programming and theoretical preparation is expected.

Experience

Technical Skills: Statistical physics, Symmetry in machine learning, Python, Neural network modelling