Timothy Hospedales

Professor of Artificial Intelligence

Turing Fellow, Alan Turing Institute (ATI)

Principal Scientist & Head of ML group, Samsung AI Research

Institute of Perception, Action and Behaviour (IPAB)
School of Informatics
The University of Edinburgh
t.hospedales at ed.ac.uk

I am a Professor within IPAB in the School of Informatics at the University of Edinburgh, where I head the Machine Intelligence Research group; Principal Scientist at Samsung AI Research Centre, Cambridge; Turing Fellow of the Alan Turing Institute.

Previously I was Reader ('16-'20) at Edinburgh and Senior Lecturer/Lecturer ('12-16) at QMUL within the Risk and Information Management (RIM) group and Centre for Intelligent Sensing, where I founded the Applied Machine Learning Lab. I recieved my PhD in Neuroinformatics from Edinburgh in 2008, working with Sethu Vijayakumar in the Statistical Machine Learning and Motor Control group, and my BA in Computer Science from the University of Cambridge in 2002.

My research focuses on data-efficient and robust machine learning using techniques such as meta-learning and lifelong transfer-learning, in both probabilistic and deep learning contexts. I have looked at a variety application areas including computer vision, vision and language, reinforcement learning for robot control, finance and beyond.

News


11/2016: Paper in AAAI'17: Guaranteeing rationality in neural networks!

01/2017: Deep Multi-task learning in ICLR'17!

10/2017: Program co-chair of BMVC2018!

11/2017: EPSRC ORCA Hub in Robotics and AI awarded!

11/2017: Paper at AAAI'18: Meta-learning for domain generalization!

02/2018: Nine papers accepted at CVPR'18!

07/2018: Our Meta Learning Active Learning paper awarded Best Paper Prize at ICML AutoML'18!

08/2018: Congratulations Kunkun Pang! Dynamic Ensemble Active Learning wins Best Student Paper Award @ ICPR2018!

03/2019: Two papers in CVPR'19!

03/2019: Two papers in ICML'19!

06/2019: Congratulations Carl Allen! Analogies Explained paper wins Honorable Mention @ ICML'19!

08/2019: Three papers at ICCV'19!

08/2019: TuckER model for fast and effective knowledge graph completion accepted at EMNLP2019!

09/2019: Two papers in NeurIPS'19!

11/2019: Two papers in AAAI'20!

12/2019: Joining T-PAMI as AE!

01/2020: Paper accepted to ICLR'20, congrats Miguel!

02/2020: Four papers at CVPR'20!

04/2020: Check out our new Survey on meta-learning!

07/2020: Four papers in ECCV'20!

08/2020: Promoted to Full Professor.

08/2016: Co-chairing BMVA symposium on Transfer Learning in Computer Vision

09/2016: Guest Editor - IET CV - Special Issue on Deep Learning for Computer Vision

10/2016: Keynote on transfer learning with paramaterised tasks and domains at ECCV 2016 TASK-CV workshop!

10/2016: Giving the zero-shot learning part of ACM Multimedia '16 tutorial Emerging topics in learning from noisy and missing data.

10/2017: Program co-chair of BMVC 2018

11/2017: Hiring one 3.5yr Postdoc on Machine Learning for Robotics & AI. Closing date: 14th Dec 2017. Advert. Apply.

03/2018: Keynote at Vision & Language Conference.

03/2018: Talk on sanity guarantees for deep learning at Deep Learning in Finance Summit London 2018.

09/2018: Hiring a postdoc research associate on deep learning, 2.5yr post, international applicants welcome. Closing date 4th October 2018. Apply.

05/2019: Teaching the Advanced Topics in Deep Learning Summer School, May 2019 in Verona.

05/2020: Keynote @ DIRA workshop in CVPR2020

08/2020: Tutorial on Domain Adaptation @ ECCV2020

Sample code for active learning & discovery (PAKDD'11,TKDE'11,ECCV'12)

Unsegmented sports news dataset available (ICDM'11). Includes reference extracted STIP features and ground-truth multi-label annotation.

Sample code for MCTM (ICCV'09, IJCV'11) and WSJTM (ACCV'10, PAMI'11)

Attribute Dataset for Heterogeneous Face Recognition (ACCV'14)

Code and Data for BMVC04 Fine-grained SBIR paper.

02/2015: Code for the famous ELF descriptor in re-id. First ever public release!

06/2015: Ground-truth Anomaly Annotation for ICCV'09 and IJCV'11 papers.

07/2015: Code for Sketch-a-Net BMVC'15 paper.

09/2015: Code for our ICLR'15 Multi-Task/Multi-domain paper

09/2015: Multi-camera video surveillance dataset from our TCSVT'15 paper

03/2016: 800K image+social metadata dataset from our ACM DH'16 paper

05/2017: LIMA dataset for re-id and tracking at CVPR'17

01/2018: Demo code CVPR'18 learning to compare few-shot learning

07/2018: Neural Decision tree Example Code.

05/2019: Hypernetwork KB Completion Code.

05/2019: Tucker Decomposition Link Prediction Code.

Visit my Google Scholar Profile.

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

Meta Learning

Lifelong and Transfer Learning

Zero-Shot Learning

Learning for Robotics and Control

Finance

Active Learning

Behaviour and Action Understanding

Person re-identification

Visual Attribute Learning

Deep Learning

Machine Learning

Zero-shot learning, Vision and Language

Sketches

Computer Vision

Neuroscience

Journal

Conference

Chapter / Workshop / Other

Postdocs

  • Reliable Machine Learning
Chenyang Zhao
  • Lifelong learning for Robotics and Control
  • Meta-Learning, Domain Generalisation

Postdocs, Affiliated

Kunkun Pang
  • Active Learning & Perception

PhD Students

  • Meta-Learning, Data Distillation, Cross-Domain
  • Meta-Learning, Self-Supervised Learning
Panagiotis Eustratiadis
  • Reinforcement Learning, Adversarial Robustness
Miguel Jacques (CDT)
  • Physics and Extrapolation in Deep Networks
Carl Allen (CDT)
  • Theoretical Analysis of Embedding Models
  • Knowledgebases, Link Prediction
Tanmoy Mukherjee
  • Vision & Language, Semantic Embeddings
Chenyang Zhao
  • Lifelong learning for Robotics and Control
Marija Jegorova
  • Generative Adversarial Networks
Xueting Zhang
  • Few Shot, Continual Learning
Boyan Gao
  • Unsupervised Deep Learning, Meta-Learning

PhD Students, Co-supervisor

  • Cross-domain, Sketches, Face Recognition. With Yi-Zhe Song.
  • Multi-View Learning. With Hakan Bilen.

PhD Students, Visiting

  • Structured Video Analysis, Generative Adversarial Nets
  • Person Identification
  • Generative, Graphical Models in Vision
  • Domain Generalisation. Deep RL.

Postdocs, Alum

  • Lifelong Learning, Deep Learning, Meta-Learning
Ryan Layne
  • Person Re-identification, Transfer Learning


PhD Students, Graduated

  • Lifelong Learning, Deep Learning
    → Postdoc at University of Edinburgh → Faculty at University of Surrey
Kunkun Pang
  • Active Perception, Active Learning
    → Postdoc at University of Edinburgh
  • Transfer Learning, Zero-shot Learning, Attributes, Learning to Rank
    → Faculty at Fudan University, PRC
Ryan Layne
  • Person Re-identification, Attributes
Yun Zhou
  • Transfer Learning in Bayesian Networks
    → Goldsmiths → Faculty at NUDT, PRC
  • Weakly Supervised Learning, Attributes
    → Imperial College London
Yi Li
  • Cross-domain matching, Sketches, SBIR
  • Visual Surveillance, Action Recognition, Transfer Learning
    → National University of Singapore

Classes

2018/19: Semester A
  • Introduction to Vision and Robotics, Image and Vision Computing
2017/18: Semester A
  • Introduction to Vision and Robotics, Image and Vision Computing
2016/17: Semester B
  • Inf1 - Object Oriented Programming
2015/16: Semester A
  • Data Mining, Machine Learning, Procedural Programming
2014/15: Semester A
  • Data Mining, Procedural Programming
2014/15: Semester B
  • Java (BUPT)
2013/14: Semester A
  • Data Mining
2013/14: Semester B
  • Semantic Web
2012/13: Semester A
  • Information Systems
2012/13: Semester B
  • Semantic Web