<div dir="ltr"><div dir="ltr"><p class="MsoNormal" align="center" style="text-align:center"><b><span lang="EN" style="font-size:12pt;line-height:17.12px;font-family:Roboto,serif">A fully-funded 4-year PhD scholarship</span></b><span lang="EN" style="font-size:14pt;line-height:19.9733px;font-family:Roboto,serif;color:rgb(51,51,51);background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial"> </span><b><span lang="EN" style="font-size:12pt;line-height:17.12px;font-family:Roboto,serif">in UKRI Centre for Doctoral Training in Artificial Intelligence, Machine Learning & Advanced Computing</span></b></p><p class="MsoNormal" align="center" style="margin-bottom:0.0001pt;text-align:center;line-height:normal"><b><span lang="EN" style="font-size:12pt;font-family:Roboto,serif">Project:</span></b><span lang="EN" style="font-size:12pt;font-family:Roboto,serif"> Few-shot Learning for Environment Adaptive Multi-modal Vision System</span></p><p class="MsoNormal" style="margin-bottom:0.0001pt;line-height:normal"><span lang="EN-US" style="font-family:"Times New Roman",serif"> </span></p><p class="MsoNormal" style="margin-bottom:0.0001pt;text-align:justify;line-height:normal"><span lang="EN-US" style="font-size:10pt;font-family:"Times New Roman",serif">As a dominating technique in AI, deep learning has been successfully used to facilitate a multitude of visual tasks, such as recognizing faces, tracking emotions, or monitoring physical activities. However, each of these tasks requires training a neural network on a </span><b><span lang="EN-US" style="font-size:10pt;font-family:"Times New Roman",serif">very large image dataset</span></b><span lang="EN-US" style="font-size:10pt;font-family:"Times New Roman",serif"> specifically collected and annotated for that task. Though the trained networks are experts for the target task, they only understand the 'world' experienced during training and can 'say' nothing about other content, nor can they be adaptive to other tasks without retraining. Moreover, most visual algorithms are learning from ‘single modal’, but pay no attention to other vision modalities, such as depth and thermal sensors. </span></p><p class="MsoNormal" style="margin-bottom:0.0001pt;text-align:justify;line-height:normal"><span lang="EN-US" style="font-size:10pt;font-family:"Times New Roman",serif"> </span></p><p class="MsoNormal" style="margin-bottom:0.0001pt;text-align:justify;line-height:normal"><span lang="EN-US" style="font-size:10pt;font-family:"Times New Roman",serif">The core objective of the project is to develop the next generation of machine learning algorithms that can mimic human vision and intelligence – continually learning to adapt to the new environment from </span><b><span lang="EN-US" style="font-size:10pt;font-family:"Times New Roman",serif">a few visual shots</span></b><span lang="EN-US" style="font-size:10pt;font-family:"Times New Roman",serif"> without requiring the traditional ‘strong supervision’ of a new dataset of each new task. Compared to the conventional supervised setting, learning from few shots poses several challenges due to, e. g. insufficient training data in a new environment; the bias between old and new tasks; the continually emerging tasks; the catastrophic forgetting problem when learning new tasks. In this project, the prospective student will solve the above problems in three stages: 1) based on our recent findings of self-paced (curriculum) learning for image classification [1] and saliency detection [2], we address the lack of annotated training data problem for real-life semantic segmentation task; 2) incorporating self-paced learning into the framework of Meta-learning, aiming to be environment adaptive by learning emerging new tasks data from old tasks; 3) delivering, distilling, and reasoning semantic and geometric information over a multitude of visual data, e.g. RGB videos, depth videos and thermal videos.</span></p><p class="MsoNormal" style="margin-bottom:0.0001pt;text-align:justify;line-height:normal"><span lang="EN-US" style="font-family:"Times New Roman",serif"> </span></p><p class="MsoNormal" style="margin-bottom:0.0001pt;text-align:justify;line-height:normal"><span lang="EN-US" style="font-size:9pt;font-family:"Times New Roman",serif">[1] L. Xiang, G. Ding and </span><b><span lang="EN-US" style="font-size:9pt;font-family:"Times New Roman",serif">J. Han</span></b><span lang="EN-US" style="font-size:9pt;font-family:"Times New Roman",serif">, Learning from Multiple Experts: Self-paced Knowledge Distillation for Long-tailed Classification, in proceeding of European Conference on Computer Vision (ECCV 2020), spotlight paper (top 5%).</span></p><p class="MsoNormal" style="margin-bottom:0.0001pt;text-align:justify;line-height:normal"><span lang="EN-US" style="font-size:9pt;font-family:"Times New Roman",serif">[2] D. Zhang, H. Tian and </span><b><span lang="EN-US" style="font-size:9pt;font-family:"Times New Roman",serif">J. Han</span></b><span lang="EN-US" style="font-size:9pt;font-family:"Times New Roman",serif">, Few-Cost Salient Object Detection with Adversarial-Paced Learning, in proceeding of 34th Conference on Neural Information Processing Systems (NeurIPS 2020).</span></p><p class="MsoNormal" style="margin-bottom:0.0001pt;text-align:justify;line-height:normal"><span lang="EN-US" style="font-size:9pt;font-family:"Times New Roman",serif"> </span></p><p style="margin:0cm 0cm 8.25pt"><b><span lang="EN" style="font-size:10.5pt">Eligibility and Desired Student background</span></b></p><p style="margin:0cm 0cm 8.25pt"><span lang="EN-US" style="font-size:10pt">We are seeking an enthusiastic individual to join the Computer Science Department at Aberystwyth University, UK, with the following attributes:</span></p><p style="margin:0cm 0cm 0.0001pt 36pt;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial"><span lang="EN-US" style="font-size:10pt;font-family:Symbol">·<span style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:7pt;line-height:normal;font-family:"Times New Roman"">         </span></span><span lang="EN-US" style="font-size:10pt;color:black">A minimum 2:1 undergraduate (BEng, MEng) and/or postgraduate masters’ qualification (MSc) in a science and technology field: Computer Science, Engineering, Mathematics, with specialisation in Computer Vision, Machine Learning and AI</span><span lang="EN-US" style="font-size:10pt"></span></p><p style="margin:0cm 0cm 0.0001pt 36pt;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial"><span lang="EN-US" style="font-size:10pt;font-family:Symbol">·<span style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:7pt;line-height:normal;font-family:"Times New Roman"">         </span></span><span lang="EN-US" style="font-size:10pt;color:black">Appropriate IELTS score (overall score of 6.0 with no component below 5.5) or TOEFL.</span><span lang="EN-US" style="font-size:10pt"></span></p><p style="margin:0cm 0cm 0.0001pt 36pt;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial"><span lang="EN-US" style="font-size:10pt;font-family:Symbol">·<span style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:7pt;line-height:normal;font-family:"Times New Roman"">         </span></span><span lang="EN-US" style="font-size:10pt;color:black">Familiarity with machine learning and probabilistic models</span><span lang="EN-US" style="font-size:10pt"></span></p><p style="margin:0cm 0cm 0.0001pt 36pt;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial"><span lang="EN-US" style="font-size:10pt;font-family:Symbol">·<span style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:7pt;line-height:normal;font-family:"Times New Roman"">         </span></span><span lang="EN-US" style="font-size:10pt;color:black">Relevant software knowledge and experience, for example Python and tensor frameworks (PyTorch or TensorFlow), C++, etc</span><span lang="EN-US" style="font-size:10pt"></span></p><p style="margin:0cm 0cm 0.0001pt 36pt;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial"><span lang="EN-US" style="font-size:10pt;font-family:Symbol">·<span style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:7pt;line-height:normal;font-family:"Times New Roman"">         </span></span><span lang="EN-US" style="font-size:10pt;color:black">A driven, professional and independent work attitude</span><span lang="EN-US" style="font-size:10pt"></span></p><p style="margin:0cm 0cm 0.0001pt 36pt;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial"><span lang="EN-US" style="font-size:10pt;font-family:Symbol">·<span style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:7pt;line-height:normal;font-family:"Times New Roman"">         </span></span><span lang="EN-US" style="font-size:10pt;color:black">Excellent written and verbal communication skills</span><span lang="EN-US" style="font-size:10pt"></span></p><p class="MsoNormal" style="margin-bottom:8.25pt;line-height:normal"><b><span lang="EN" style="font-size:10.5pt;font-family:"Times New Roman",serif">Funding Notes</span></b></p><p style="margin:0cm 0cm 0.0001pt 36pt;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial"><span lang="EN-US" style="font-size:10pt;font-family:Symbol">·<span style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:7pt;line-height:normal;font-family:"Times New Roman"">         </span></span><span lang="EN-US" style="font-size:10pt;color:black">The 4-year PhD scholarship, will sit within the UKRI Centre for Doctoral Training in Artificial Intelligence, Machine Learning & Advanced Computing</span><span lang="EN-US" style="font-size:10pt"></span></p><p style="margin:0cm 0cm 0.0001pt 36pt;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial"><span lang="EN-US" style="font-size:10pt;font-family:Symbol">·<span style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:7pt;line-height:normal;font-family:"Times New Roman"">         </span></span><span lang="EN-US" style="font-size:10pt;color:black">Funding will cover the full cost of tuition fees and an annual stipend of £15,285 </span><b><span lang="EN-US" style="font-size:10pt;color:black">for 4 years</span></b><span lang="EN-US" style="font-size:10pt"></span></p><p style="margin:0cm 0cm 0.0001pt 36pt;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial"><span lang="EN-US" style="font-size:10pt;font-family:Symbol">·<span style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:7pt;line-height:normal;font-family:"Times New Roman"">         </span></span><span lang="EN-US" style="font-size:10pt;color:black">The post is open to both Home/EU and overseas students</span><span lang="EN-US" style="font-size:10pt"></span></p><p style="margin:0cm 0cm 0.0001pt 36pt;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial"><span lang="EN-US" style="font-size:10pt;font-family:Symbol">·<span style="font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:7pt;line-height:normal;font-family:"Times New Roman"">         </span></span><span lang="EN-US" style="font-size:10pt;color:black">Additional funding is available for training, research and conference expenses</span><span lang="EN-US" style="font-size:10pt"></span></p><p style="margin:0cm 0cm 0.0001pt 36pt;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial"><span lang="EN-US" style="font-size:10pt"> </span></p><p class="MsoNormal" style="margin-bottom:8.25pt;line-height:normal"><b><span lang="EN" style="font-size:10.5pt;font-family:"Times New Roman",serif">How to Apply:</span></b></p><p class="MsoNormal" style="margin-bottom:8.25pt;line-height:normal"><span lang="EN-US" style="font-size:10pt;font-family:"Times New Roman",serif">Applications through Aberystwyth’s electronic application process <u><a href="https://www.aber.ac.uk/en/postgrad/apply/#apply-online" target="_blank"><span style="color:windowtext">Aberystwyth University - Postgraduate Study : How to apply</span></a></u> must include the following attachments in pdf form:</span><br style="box-sizing:border-box"><br style="box-sizing:border-box">1. CV<br style="box-sizing:border-box">2. Degree certificates and transcripts (if you are still an undergraduate, provide a transcript of results known to date)<br style="box-sizing:border-box">3. A statement no longer than 1000 words that explains why you want to join our Centre, and your preferred topic/supervisor.<br style="box-sizing:border-box">4. Academic references - all scholarship applications require two supporting references to be submitted. Please ensure that your chosen referees are aware of the funding deadline (to be determined), as their references form a vital part of the evaluation process. Please include these with your scholarship application.<br style="box-sizing:border-box"><br style="box-sizing:border-box">In addition, email the pdf(s) of your application to Prof. Jungong Han [<a href="mailto:jungong.han@aber.ac.uk" target="_blank">jungong.han@aber.ac.uk</a>].<br style="box-sizing:border-box"><br style="box-sizing:border-box">The deadline for applications is 12 February 2021; and the start date is Oct. 2021.</p><p class="MsoNormal" style="margin-bottom:8.25pt;line-height:normal"><b><span lang="EN" style="font-size:10.5pt;font-family:"Times New Roman",serif">Supervisors:</span></b></p><p class="MsoNormal" style="margin-bottom:0.0001pt;text-align:justify;line-height:normal"><span lang="EN-US" style="font-size:10pt;font-family:"Times New Roman",serif">Prof. Jungong Han holds a Chair in Computer Science at Aberystwyth University, UK, and is an Honorary Professor at the University of Warwick, UK. Prof. Han has an active research portfolio in the area of Video Analytics, Computer Vision, and applied Machine Learning. Prof. Han has 150+ publications in leading venues, including 50+ IEEE/ACM Transactions and 50+ A* conferences (12 CVPR/ICCV/ECCV; 2 NeurIPS; 1 ICML; 15 AAAI), cited over 6100 times. He is an Associate Editor of Elsevier Neurocomputing, an Editorial Board Member of Springer Multimedia Tools and Applications, and an Associate Editor of IET Computer Vision.</span></p><p class="MsoNormal" style="margin-bottom:0.0001pt;text-align:justify;line-height:normal"><span lang="EN-US" style="font-size:10pt;font-family:"Times New Roman",serif"> </span></p><p class="MsoNormal" style="margin-bottom:0.0001pt;text-align:justify;line-height:normal"><span lang="EN-US" style="font-size:10pt;font-family:"Times New Roman",serif">Prof. Qiang Shen is an established Chair, and Pro-Vice-Chancellor: Faculty of Business and Physical Sciences at Aberystwyth University. He is a Fellow and Council member of the Learned Society of Wales (the national academy of Wales). Prof. Shen’s research interests include computational intelligence, reasoning and learning under uncertainty, pattern recognition, data mining, and real-world applications of such techniques for intelligent decision support (e.g., crime detection, space exploration, consumer profiling, systems monitoring, and medical diagnosis). Prof. Shen has published 2 research monographs and over 400 peer-refereed papers, a number of which have received prestigious international prizes, including one Outstanding IEEE Transactions Paper Award. Prof. Shen  is an  associate editor/editorial board member of several leading international journals, including <a href="http://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=91" target="_blank"><span style="color:windowtext;text-decoration-line:none">IEEE Transactions on Fuzzy Systems</span></a>, IEEE Transactions on Cybernetics, <a href="http://www.elsevier.com/wps/find/journaldescription.cws_home/505545/description#description" target="_blank"><span style="color:windowtext;text-decoration-line:none">Fuzzy Sets and Systems</span></a>, etc. </span></p></div></div>