<html><head><meta http-equiv="content-type" content="text/html; charset=utf-8"></head><body dir="auto"><p style="margin: 0cm 0cm 0.0001pt;"><span style="background-color: rgba(255, 255, 255, 0);"><b>PhD position in Machine Learning in the Clopath lab under the joint supervision of Dr Claudia Clopath (Imperial College London) and Dr Razvan Pascanu (DeepMind London)</b><o:p></o:p></span></p><p class="MsoNormal" style="margin: 0cm 0cm 0.0001pt;"><o:p style="background-color: rgba(255, 255, 255, 0);"> </o:p></p><p style="margin: 0cm 0cm 0.0001pt;"><span style="background-color: rgba(255, 255, 255, 0);"><i>-----------------Description--------------</i><o:p></o:p></span></p><p style="margin: 0cm 0cm 0.0001pt;"><span style="background-color: rgba(255, 255, 255, 0);">This is a 3-year, fully-funded PhD position at Imperial College London. The position will be co-supervised by Dr. Claudia Clopath (head of the lab) and Dr. Razvan Pascanu (Research Scientist at DeepMind). The student will be based in the Clopath lab. The focus of the PhD will be on machine learning while drawing inspiration from computational neuroscience. In particular, the main themes will revolve around Deep Learning and Deep Reinforcement Learning. The research will be published at conferences such as NeurIPS, ICML or ICLR. </span></p><p style="margin: 0cm 0cm 0.0001pt;"><span style="background-color: rgba(255, 255, 255, 0);"><br></span></p><p style="margin: 0cm 0cm 0.0001pt;"><span style="background-color: rgba(255, 255, 255, 0);">An *incomplete* list of topics of interest could include:<o:p></o:p></span></p><ul type="disc" style="margin-bottom: 0cm; margin-top: 0cm;"><li style="margin-top: 0cm; margin-bottom: 0.0001pt; vertical-align: baseline;"><span style="background-color: rgba(255, 255, 255, 0);">Theory of Deep Learning/Deep Reinforcement Learning<o:p></o:p></span></li><ul type="circle" style="margin-bottom: 0cm; margin-top: 0cm;"><li style="margin-top: 0cm; margin-bottom: 0.0001pt; vertical-align: baseline;"><span style="background-color: rgba(255, 255, 255, 0);">Understanding learning dynamics<o:p></o:p></span></li><li style="margin-top: 0cm; margin-bottom: 0.0001pt; vertical-align: baseline;"><span style="background-color: rgba(255, 255, 255, 0);">Representational power of neural networks<o:p></o:p></span></li></ul><li style="margin-top: 0cm; margin-bottom: 0.0001pt; vertical-align: baseline;"><span style="background-color: rgba(255, 255, 255, 0);">Exploiting our understanding of dopamine for training RL systems<o:p></o:p></span></li><li style="margin-top: 0cm; margin-bottom: 0.0001pt; vertical-align: baseline;"><span style="background-color: rgba(255, 255, 255, 0);">Topics around introducing structure in neural networks (e.g. Graph-Nets)<o:p></o:p></span></li><li style="margin-top: 0cm; margin-bottom: 0.0001pt; vertical-align: baseline;"><span style="background-color: rgba(255, 255, 255, 0);">Optimization for neural networks (theory/application)<o:p></o:p></span></li><li style="margin-top: 0cm; margin-bottom: 0.0001pt; vertical-align: baseline;"><span style="background-color: rgba(255, 255, 255, 0);">Understanding memory in recurrent models/attention based models (improve the memory capacity of RNNs)<o:p></o:p></span></li><li style="margin-top: 0cm; margin-bottom: 0.0001pt; vertical-align: baseline;"><span style="background-color: rgba(255, 255, 255, 0);">Meta-learning/Continual learning or exploration in RL systems<o:p></o:p></span></li></ul><p style="margin: 0cm 0cm 0.0001pt;"><span style="background-color: rgba(255, 255, 255, 0);"><br></span></p><p style="margin: 0cm 0cm 0.0001pt;"><span style="background-color: rgba(255, 255, 255, 0);">Although we welcome the student’s suggestion for potential projects.<o:p></o:p></span></p><p class="MsoNormal" style="margin: 0cm 0cm 12pt;"><o:p style="background-color: rgba(255, 255, 255, 0);"> </o:p></p><p style="margin: 0cm 0cm 0.0001pt;"><span style="background-color: rgba(255, 255, 255, 0);"><i>---------------Requirements--------------------</i><o:p></o:p></span></p><p style="margin: 0cm 0cm 0.0001pt;"><span style="background-color: rgba(255, 255, 255, 0);">The perfect candidate has a strong background in maths, physics, engineering or related fields. Machine Learning background is a plus but not necessary. The candidate should have strong programming skills (knowledge of tensorflow/pytorch would be a plus but not required). Interest in neural networks research and experience is a plus.<o:p></o:p></span></p><p class="MsoNormal" style="margin: 0cm 0cm 0.0001pt;"><o:p style="background-color: rgba(255, 255, 255, 0);"> </o:p></p><p style="margin: 0cm 0cm 0.0001pt;"><span style="background-color: rgba(255, 255, 255, 0);">Due to the funding requirement of the position, the candidate should be an *EU or British citizen*. The funding/position will not be affected by the Brexit outcome. <o:p></o:p></span></p><p class="MsoNormal" style="margin: 0cm 0cm 0.0001pt;"><o:p style="background-color: rgba(255, 255, 255, 0);"> </o:p></p><p style="margin: 0cm 0cm 0.0001pt;"><span style="background-color: rgba(255, 255, 255, 0);"><i>------------ The lab ------------</i><o:p></o:p></span></p><p style="margin: 0cm 0cm 0.0001pt;"><span style="background-color: rgba(255, 255, 255, 0);">The Computational Neuroscience Laboratory is very young and dynamic, and publishes in prestigious journals and conferences. It is part of the Department of Bioengineering, which conducts state-of-the-art multidisciplinary research. The lab is at Imperial College London, the 3rd ranked university in Europe, is in the top 10 worldwide, and is located in the city centre of London. More information can be found at: <a href="http://www.bg.ic.ac.uk/research/c.clopath/">http://www.bg.ic.ac.uk/research/c.clopath/</a><o:p></o:p></span></p><p class="MsoNormal" style="margin: 0cm 0cm 0.0001pt;"><o:p style="background-color: rgba(255, 255, 255, 0);"> </o:p></p><p style="margin: 0cm 0cm 0.0001pt;"><span style="background-color: rgba(255, 255, 255, 0);">DeepMind is a company focusing on AI and ML research, with world-wide recognized expertise. Its research centers around Deep Learning and Reinforcement Learning, with successes such as AlphaGo, first AI winning against a human professional at the game of Go. More information here: <a href="https://deepmind.com/about/">https://deepmind.com/about/</a>. For more info on Razvan Pascanu’s work see <a href="https://sites.google.com/view/razp">https://sites.google.com/view/razp</a>.<o:p></o:p></span></p><p class="MsoNormal" style="margin: 0cm 0cm 0.0001pt;"><o:p style="background-color: rgba(255, 255, 255, 0);"> </o:p></p><p style="margin: 0cm 0cm 0.0001pt;"><span style="background-color: rgba(255, 255, 255, 0);"><i>----------------- How to apply:-----------------</i><o:p></o:p></span></p><p style="margin: 0cm 0cm 0.0001pt;"><span style="background-color: rgba(255, 255, 255, 0);">Candidates should send a single pdf file, consisting of a 1 to 2-page motivation letter followed by a CV (including publication list) to <a href="mailto:judith@mypajudith.com" dir="ltr" x-apple-data-detectors="true" x-apple-data-detectors-type="link" x-apple-data-detectors-result="8">judith@mypajudith.com</a>, with the subject containing 'PhD_ML_2019'. In addition, candidates should organize two letters of reference to be sent to <a href="mailto:judith@mypajudith.com" dir="ltr" x-apple-data-detectors="true" x-apple-data-detectors-type="link" x-apple-data-detectors-result="9">judith@mypajudith.com</a>, with the subject containing 'PhD_ML_2019'. The position is open until filled but the earlier applications will be considered first (first round of consideration will take applications sent <a href="x-apple-data-detectors://10" dir="ltr" x-apple-data-detectors="true" x-apple-data-detectors-type="calendar-event" x-apple-data-detectors-result="10" style="-webkit-text-decoration-color: rgba(0, 0, 0, 0.258824);">before April 10th</a>).<o:p></o:p></span></p><p class="MsoNormal" style="margin: 0cm 0cm 0.0001pt;"><span lang="EN-US" style="background-color: rgba(255, 255, 255, 0);"> </span></p></body></html>