<!DOCTYPE html><html><head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
  </head>
  <body>
    <div class="field field--name-body field--type-text-with-summary field--label-hidden field--item">
      <div class="tex2jax_process">
        <p><span>We invite applications for four Postdoctoral Research
            Associates (PDRA) to join the EPSRC Hub on the Mathematical
            and Computational Foundations of Artificial Intelligence.
            One PDRA will be recruited for each of the following four
            research themes: Learning with Structured & Geometric
            Models, Low Effective-dimensional Learning Models, Implicit
            Regularization, and Reinforcement Learning through
            Stochastic Control. </span><span>A brief description of
            each these is as follows (additional details are in the
            further particulars): </span></p>
        <p><span><strong>Learning with Structured and Geometric Models.</strong> 
            We will apply tools from manifold learning and Riemannian
            optimisation to leverage the underlying manifold structure
            for better training and novel network designs. </span></p>
        <p><span><strong>Low Effective-dimensional Learning Models.</strong>
             We will extend foundational theory of how large ML systems
            can be regularised to have dramatically fewer trainable
            parameters without sacrificing accuracy by analysing the use
            of low-dimensional building blocks </span></p>
        <p><span><strong>Implicit Regularization.</strong> We aim to
            develop mathematical understanding of implicit
            regularisation properties in deep neural networks to guide
            the development of algorithmic paradigms aimed at combining
            statistical optimality with computational efficiency. </span></p>
        <p><span><strong>Reinforcement Learning through Stochastic
              Control.</strong>  We will develop methods from stochastic
            control, which will provide a mathematically grounded
            approach that has a well-posed continuous-time limit (as
            opposed to traditional RL methods that are inherently
            discrete and do not scale favourably for high frequency
            observations without judicious hyper-parameter tuning). </span></p>
        <p><span>The PDRAs will work with faculty across the
            multi-university Hub, but will be employed by and directly
            supervised by faculty within the Mathematical Institute at
            the University of Oxford.  Faculty within the Mathematical
            Institute associated with the above work packages include
            Profs. Cartis, Cohen, Hauser, Lambiotte, Reisinger,
            Sirignano, and Tanner. </span></p>
        <p><span>These are two-year, fixed-term position, funded by a
            research grant from the EPSRC. The starting date of this
            position is flexible with an earliest start date of 01 March
            2025.  </span></p>
        <p><span>The successful candidates will be expected to conduct
            research which falls within the remit of this large-scale
            project and will have the opportunity to do so
            collaboratively with other members of the hub, both at
            Oxford and/or with hub partners which include universities
            as well as companies and governmental organisations. </span><span>They
            will contribute to the activities of the wider machine
            learning and data science research group and write up the
            results of their work, with co-authors, for publication in
            refereed journals and proceedings. There will be
            opportunities to contribute a small amount of teaching to
            the department, of at most three hours a week during the
            academic terms.</span></p>
        <p><span>You will have, or be close to completing, a PhD in
            mathematics or a related discipline, and possess sufficient
            specialist knowledge in the discipline to work within
            established research programmes. Excellent communication
            skills are essential, including the ability to write for
            publication, present research proposals and results, and
            represent the research group at meetings.</span></p>
        <p> </p>
        <p><span>We proudly hold a departmental Athena SWAN Silver Award
            and an institutional Race Equality Charter Bronze Award,
            which guide our progress towards advancing racial and gender
            equality. As part of our strategic aim to improve staff
            equality and diversity, we would particularly welcome
            applications from women and BME candidates, who are
            currently under-represented in positions of this type within
            the department.</span></p>
        <p> </p>
        <p><span>Please direct informal enquiries to the Recruitment
            Coordinator (email: </span><a href="mailto:recruitment@maths.ox.ac.uk" title="mailto:recruitment@maths.ox.ac.uk" class="moz-txt-link-freetext">recruitment@maths.ox.ac.uk</a><span>),
            quoting vacancy reference <strong>176180</strong>.</span></p>
        <p><span>Applicants will be selected for interview purely based
            on their ability to satisfy the selection criteria as
            outlined in full in the job description. You will be
            required to upload a statement setting out how you meet the
            selection criteria, a curriculum vitae including full list
            of publications, a statement of research interests, and the
            contact details of two referees as part of your online
            application. <strong>(NOTE: Applicants are responsible for
              contacting their referees and making sure that their
              letters are received by the closing date).</strong></span></p>
        <p><span>Applications for this vacancy are to be made online. To
            apply for this vacancy and for further information,
            including a job description and selection criteria, please
            click on the link below:</span></p>
        <p><span lang="EN-US"></span></p>
        <p><a href="https://my.corehr.com/pls/uoxrecruit/erq_jobspec_details_form.jobspec?p_id=176180" title="Unmangled Microsoft Safelink"><span lang="EN-US">https://my.corehr.com/pls/uoxrecruit/erq_jobspec_details_form.jobspec?p_id=176180</span></a><span lang="EN-US"></span></p>
        <p><span lang="EN-US">Applications received before <strong>12.00
              noon</strong> UK time on <strong>Monday, 02 December
              2024 </strong>will receive full consideration.
            Applications after this date will be considered at the
            discretion of the committee.</span></p>
      </div>
    </div>
    <div class="field field--name-field-vacancy-file field--type-file field--label-above">
      <div class="field--label">Job description</div>
      <div class="field--item"><span class="file file--mime-application-pdf file--application-pdf icon-before"><span class="file-icon"><span class="icon glyphicon glyphicon-file text-primary" aria-hidden="true"></span></span><span class="file-link"><a href="https://www.maths.ox.ac.uk/system/files/vacancies/176180_PDRA%20JD%20and%20Selection%20Criteria_0.pdf" title="Unmangled Microsoft Safelink">176180_PDRA JD and
              Selection Criteria_0.pdf</a></span><span class="file-size">562.59
            KB</span></span></div>
    </div>
    <p></p>
    <p>With best wishes Coralia Cartis</p>
    <p>Professor of Numerical Optimization,</p>
    <p>Mathematical Institute, University of Oxford </p>
    <br>
    <br>
    ------------------------------<br>
    Coralia Cartis<br>
    Oxford<br>
    ------------------------------<br>
    <br>
  </body>
</html>