CV

Education

• MEng in Electrical and Electronic Engineering, Bristol University, 1st Class Honours 78%, 2014-2018
• Ph.D in Computer Vision, Imperial College London, 2018-2023

Work experience

• Huawei Noah’s Ark Lab UK Research Scientist: 06/23 - present
  • Foundational research on knowledge distillation for vision-language models.
  • Working with multi-modality models and LLMs.
• Samsung Research UK Machine Learning Research Scientist Intern: 06/22 - 12/22
  • Introduced a novel method for semi-supervised video object segmentation on mobile devices, using a unification of representation distillation and contrastive learning.
  • Achieved competitive performance despite running up to ×5 faster, and with ×32 fewer parameters.
  • This work has been accepted for presentation at CVPR 2023.
• Waymont Consulting Software Engineer Intern: 06/17 - 09/17
  • Developed a responsive web GUI with a C++ backend to automate tests for signal processing equipment.
  • Replaced previously manual testing system and increased efficiency by hundredfold.
• Toshiba Research Europe Research Engineer Intern: 06/16 - 09/16
  • Designed out-of-tree blocks in C++ using the GNURadio signal processing environment to characterise and linearise a communication channel between two USRP devices.
  • Conducted research on utilising the RC-5 Infrared (IR) remote control standard for short-range, low-cost, and low-data rate IoT applications
• Toshiba Research Europe Research Engineer Intern: 06/15 - 09/15
  • Developed a MATLAB GUI to operate a motor-driven positioner in conjunction with a Vector Network Analyzer for measuring antenna patterns within an anechoic chamber.
  • Modified existing VB.NET code to synchronize two X-Y positioners for analyzing propagation channels.

Skills

• Programming: C++, Python, CUDA, C#, VHDL, MATLAB
• Frameworks: PyTorch, Tensorflow, CuDNN, OpenCV, Vivado HLS, Qt, Django
• Web Dev: HTML, CSS, JavaScript, Python, SQL

Publications

Cascaded channel pruning using hierarchical self-distillation

Miles, R., & Mikolajczyk, K. (2020). Cascaded channel pruning using hierarchical self-distillation. BMVC.

Compression of descriptor models for mobile applications

Miles, R., & Mikolajczyk, K. (2021). Compression of descriptor models for mobile applications. ICASSP.

Reconstructing Pruned Filters using Cheap Spatial Transformations

Miles, R., & Mikolajczyk, K. (2021). Reconstructing Pruned Filters using Cheap Spatial Transformations. ICCVW.

Information Theoretic Representation Distillation

Miles, R., Rodríguez, A. L., & Mikolajczyk, K. (2021). Information Theoretic Representation Distillation. BMVC

MobileVOS: Real-Time Video Object Segmentation Contrastive Learning meets Knowledge Distillation

Miles, R., & Kerim Yucel, M., & Manganelli, B., Saa-Garriga, A. (2021). MobileVOS: Real-Time Video Object Segmentation Contrastive Learning meets Knowledge Distillation. CVPR.

Understanding the Role of the Projector in Knowledge Distillation

Miles, R., & Mikolajczyk, K. (2023). Understanding the Role of the Projector in Knowledge Distillation. AAAI.

Teaching

Graduate teaching assistant for the following final year courses:

• Deep Learning
• Machine Learning
• Computer Vision and Pattern Recognition

Awards and Achievements

• Highly Commendable Award in the ARM 24 Hour Hackathon : Developed a wearable electronic glove to identify the alphabetic letters in sign language gestures. The glove could then convert the series of gestures into speech using the built in speakers.