Generative learning and non-equilibrium statistical mechanics

A fully funded postdoctoral position is available in the group of Prof. Tristan Bereau (tristanbereau.com) at the Institute for Theoretical Physics, Heidelberg University. We invite applications for a position at the intersection of generative deep learning and non-equilibrium statistical mechanics. The initial appointment will be for two years, with a possible extension of one year. The starting date is negotiable, preferably Fall 2023. The position is funded according to the German public service scale (TV-L E13) commensurate with experience. It includes funds for travel, visits, and publications. Both CPU and GPU computing resources are available.

About the Role

The successful candidate will focus on employing and advancing deep diffusion probabilistic machine learning techniques to understand and predict the behavior of molecular systems. Connections to non-equilibrium statistical mechanics will play a central role.

Relevant group publications around generative learning, non-equilibrium statistical mechanics:

import {Plot} from "@mkfreeman/plot-tooltip"
import {map} from "@martien/ramda"

papers_r = FileAttachment("../data/all_papers.json").json()
n_columns = Object.entries(papers_r["doi"]).length
index_array = Array.from({length: n_columns}, (_, i) => i)
papers = map(i => map(x => x[i], papers_r), index_array)

function displayTable(papers_) {
    return Inputs.table(papers_, {
        columns: [
        "doi",
        "title",
        "journal.name",
        "journal.volume",
        "year",
      ],
      header: {
        "journal.name": "journal",
        "journal.volume": "volume",
      },
      format: {
        doi: doi => htl.html`<a href=https://doi.org/${doi} target=_blank>${doi}</a>`,
        title: title => htl.html`${title}`,
        "journal.volume": volume => htl.html`<b>${volume}</b>`,
        year: year => htl.html`${year}`,
      },
      width: {
        "title": 10
      },
      layout: "auto",
    })
}

displayTable(
    papers.filter(function(p) {
        return (
            p.authors.map(a => a.full_name).includes("Marc Stieffenhofer")
            || p.authors.map(a => a.full_name).includes("Marius Bause")
            || p.authors.map(a => a.full_name).includes("Kirill Shmilovich")
        )
    })
)

  import {Plot as Plot} from "@mkfreeman/plot-tooltip"

  import {map as map} from "@martien/ramda"

papers_r = Object {doi: Object, title: Object, publication_date: Object, authors: Object, citation_count: Object, bib_entry: Object, abstract: Object, tldr: Object, year: Object, embedding: Object, journal.name: Object, journal.url: Object, journal.issue: Object, journal.abbreviation: Object, journal.pages: Object, journal.volume: Object, open_access.is_open_access: Object, open_access.open_access_status: Object, open_access.landing_page_url: Object, open_access.pdf_url: Object}

n_columns = 79

index_array = Array(79) [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, …]

papers = Array(79) [Object, Object, Object, Object, Object, Object, Object, Object, Object, Object, Object, Object, Object, Object, Object, Object, Object, Object, Object, Object, …]

displayTable = ƒ(papers_)

	doi	title	journal	volume	year
	10.3389/fchem.2022.982757	Benchmarking coarse-grained models of organic semiconductors via deep backmapping	Frontiers in Chemistry	10	2022
	10.1039/D2SC00116K	Data-driven discovery of cardiolipin-selective small molecules by computational active learning	Chemical Science	13	2022
	10.1063/5.0042972	Reweighting non-equilibrium steady-state dynamics along collective variables	The Journal of Chemical Physics	154	2021
	10.1063/5.0039102	Adversarial reverse mapping of condensed-phase molecular structures: Chemical transferability	APL Materials	9	2021
	10.1088/2632-2153/abb6d4	Adversarial reverse mapping of equilibrated condensed-phase molecular structures	Machine Learning: Science and Technology	1	2020
	10.1103/PhysRevE.100.060103	Microscopic reweighting for nonequilibrium steady-state dynamics	Physical Review E	100	2019

About You

The ideal candidate will possess:

1. A Ph.D. in Physics, Computer Science, Applied Mathematics, or a related field.
2. Ideally both, but at least either of:
   1. Strong theoretical background in statistical mechanics, particularly in non-equilibrium systems.
   2. Proficiency in machine learning, including experience with deep learning frameworks such as TensorFlow or PyTorch.
3. Excellent programming skills in Python or a similar high-level language.
4. Evidence of high-quality scientific publications.
5. Excellent communication skills in English, both written and oral.

Why Join Us

As the oldest university in Germany, Heidelberg University is a world-leading institution and provides an internationally recognised and collaborative environment. Joining Heidelberg University’s Institute for Theoretical Physics will give you the opportunity to:

1. Work on exciting topics at the intersection of statistical mechanics and machine learning.
2. Access state-of-the-art facilities and resources.
3. Engage with a vibrant intellectual community in Heidelberg, including Simplaix, the Interdisciplinary Center for Scientific Computing (IWR), the cluster of excellence STRUCTURES, ELLIS, the Heidelberg Institute for Theoretical Studies, and Scientific Machine Learning.
4. Live in the beautiful and historic city of Heidelberg.

We welcome applications from all individuals, regardless of their gender, ethnicity, disability, or any other protected characteristic. We encourage diversity and foster a culture of inclusion where everyone feels welcome.

Application Process

Interested candidates should submit the following via email as a single PDF:

1. Curriculum Vitae;
2. List of publications;
3. Contact information of two supervisors/collaborators to request recommendation letters;
4. Brief summary of previous research projects (one or two pages).

Applications submitted by July 31, 2023 will receive full consideration.

For more information or to apply, please contact Prof. Tristan Bereau at bereau@thphys.uni-heidelberg.de