Nikhil Churamani

230 City Rd

London EC1V 2QY, UK

I work as a Senior AI Scientist at Medtronic Digital Surgery working on Medical Image analyses and Digital Surgery solutions.

Previously, I was a Post-Doctoral Research Associate at the Computer Laboratory, University of Cambridge, working with Prof Hatice Gunes at the Affective Intelligence and Robotics (AFAR) Lab. My research focussed on Continual Learning for Social and Affective Robots, aimed towards enabling lifelong adaptability in robots operating in real and open-world settings. In particular, my work focused on embedding robots with a personalised and continually adaptive understanding of human behaviour, allowing them to learn socially and contextually appropriate behaviours in human-centred environments.

My post-doctoral work, funded by Google under the GIG Grant, focused on Federated Continual Learning of Socially Appropriate Robot Behaviours where I was exploring how social robots can adapt to differences in social norms, learning from each others’ experience, in a federated manner.

I completed my PhD at Trinity Hall, University of Cambridge. My doctoral work titled, Continual Learning for Affective Robotics was funded by the EPSRC International Doctoral Scholarship and the Premium Research Studentship of the Computer Lab, University of Cambridge. My other research interests include Deep Learning, Computer Vision, Neuro-inspired AI, and Human-Robot Interaction.

I completed my MSc in Intelligent Adaptive Systems from Universität Hamburg where I was supervised by Prof Dr Stefan Wermter. My MSc dissertation titled, Modelling Affective Cores for Behaviour Modulation in Social Robots, advised by Dr Pablo Barros, focused on embedding intrisic affective motivations in social robots as an Affective Core that enables different personalities in their behaviour modelling in Human-Robot collaborative tasks.

news

Mar 11, 2024	Two papers at the ACM/IEEE International Conference on Human-Robot Interaction (HRI) 2024: Lifelong Learning and Personalization in Long-Term Human-Robot Interaction (LEAP-HRI): Open-World Learning and Causal-HRI: Causal Learning for Human-Robot Interaction.
Mar 10, 2024	Paper accepted at the LEAP-HRI Workshop at the ACM/IEEE International Conference on Human-Robot Interaction (HRI) 2024: Federated Learning of Socially Appropriate Agent Behaviours in Simulated Home Environments.
Feb 27, 2024	Happy to announce the Dungeons Neurons and Dialogues Workshop to be held at IEEE International Conference on Affective Computing and Intelligent Interaction (ACII) 2024.
Dec 2, 2023	Happy to announce the Causal-HRI Workshop to be held at ACM/IEEE International Conference on Human-Robot Interaction (HRI) 2024.
Dec 2, 2023	Happy to announce the Leap-HRI Workshop to be held at ACM/IEEE International Conference on Human-Robot Interaction (HRI) 2024.

selected publications

Continual Learning for Affective Robotics: Why, What and How?

Nikhil Churamani, Sinan Kalkan, and Hatice Gunes

In 2020 29th IEEE International Conference on Robot and Human Interactive Communication (RO-MAN), 2020

Bib PDF Slides Video

@inproceedings{Churamani2020CL4AR,
  author = {Churamani, Nikhil and Kalkan, Sinan and Gunes, Hatice},
  booktitle = {2020 29th IEEE International Conference on Robot and Human Interactive Communication (RO-MAN)},
  title = {{Continual Learning for Affective Robotics: Why, What and How?}},
  year = {2020},
  pages = {425-431},
  doi = {10.1109/RO-MAN47096.2020.9223564},
}

Learning Socially Appropriate Robo-Waiter Behaviours through Real-Time User Feedback

Emily McQuillin, Nikhil Churamani, and Hatice Gunes

In Proceedings of the 2022 ACM/IEEE International Conference on Human-Robot Interaction, 2022

Abs Bib PDF Supp Video

Current Humanoid Service Robot (HSR) behaviours mainly rely on static models that cannot adapt dynamically to meet individual customer attitudes and preferences. In this work, we focus on empowering HSRs with adaptive feedback mechanisms driven by either implicit reward, by estimating facial affect, or explicit reward, by incorporating verbal responses of the human ’customer’. To achieve this, we first create a custom dataset, annotated using crowd-sourced labels, to learn appropriate approach (positioning and movement) behaviours for a Robo-waiter. This dataset is used to pre-train a Reinforcement Learning (RL) agent to learn behaviours deemed socially appropriate for the robo-waiter. This model is later extended to include separate implicit and explicit reward mechanisms to allow for interactive learning and adaptation from user social feedback. We present a within-subjects Human-Robot Interaction (HRI) study with 21 participants implementing interactions between the robo-waiter and human customers implementing the above-mentioned model variations. Our results show that both explicit and implicit adaptation mechanisms enabled the adaptive robo-waiter to be rated as more enjoyable and sociable, and its positioning relative to the participants as more appropriate compared to using the pre-trained model or a randomised control implementation.
@inproceedings{McQuillin2022RoboWaiter, author = {McQuillin, Emily and Churamani, Nikhil and Gunes, Hatice}, title = {Learning Socially Appropriate Robo-Waiter Behaviours through Real-Time User Feedback}, year = {2022}, publisher = {IEEE Press}, booktitle = {Proceedings of the 2022 ACM/IEEE International Conference on Human-Robot Interaction}, pages = {541–550}, numpages = {10}, keywords = {facial affect, implicit feedback, explicit feedback, reinforcement learning, humanoid robo-waiter}, location = {Sapporo, Hokkaido, Japan}, series = {HRI '22}, doi = {10.5555/3523760.3523831}, }

Domain-Incremental Continual Learning for Mitigating Bias in Facial Expression and Action Unit Recognition

Nikhil Churamani, Ozgur Kara, and Hatice Gunes

IEEE Transactions on Affective Computing, 2022

arXiv Bib PDF

@article{Churamani2022DICL4Bias,
  author = {Churamani, Nikhil and Kara, Ozgur and Gunes, Hatice},
  journal = {{IEEE Transactions on Affective Computing}},
  title = {{Domain-Incremental Continual Learning for Mitigating Bias in Facial Expression and Action Unit Recognition}},
  year = {2022},
  pages = {1-15},
  doi = {10.1109/TAFFC.2022.3181033},
}

Latent Generative Replay for Resource-Efficient Continual Learning of Facial Expressions

Samuil Stoychev, Nikhil Churamani, and Hatice Gunes

In 2023 IEEE 17th International Conference on Automatic Face and Gesture Recognition (FG), 2023

Bib PDF Supp Video

@inproceedings{Stoychev2023LGR,
  author = {Stoychev, Samuil and Churamani, Nikhil and Gunes, Hatice},
  booktitle = {2023 IEEE 17th International Conference on Automatic Face and Gesture Recognition (FG)},
  title = {Latent Generative Replay for Resource-Efficient Continual Learning of Facial Expressions},
  year = {2023},
  pages = {1-8},
  doi = {10.1109/FG57933.2023.10042642},
}

Towards Causal Replay for Knowledge Rehearsal in Continual Learning

Nikhil Churamani, Jiaee Cheong, Sinan Kalkan, and Hatice Gunes

In Proceedings of The First AAAI Bridge Program on Continual Causality, 2023

Abs Bib PDF Slides

Given the challenges associated with the real-world deployment of Machine Learning (ML) models, especially towards efficiently integrating novel information on-the-go, both Continual Learning (CL) and Causality have been proposed and investigated individually as potent solutions. Despite their complimentary nature, the bridge between them is still largely unexplored. In this work, we focus on causality to improve the learning and knowledge preservation capabilities of CL models. In particular, positing Causal Replay for knowledge rehearsal, we discuss how CL-based models can benefit from causal interventions towards improving their ability to replay past knowledge in order to mitigate forgetting.
@inproceedings{pmlr-v208-churamani23a, title = {Towards Causal Replay for Knowledge Rehearsal in Continual Learning}, author = {Churamani, Nikhil and Cheong, Jiaee and Kalkan, Sinan and Gunes, Hatice}, booktitle = {Proceedings of The First AAAI Bridge Program on Continual Causality}, pages = {63--70}, year = {2023}, editor = {Mundt, Martin and Cooper, Keiland W. and Dhami, Devendra Singh and Ribeiro, Adéle and Smith, James Seale and Bellot, Alexis and Hayes, Tyler}, volume = {208}, series = {Proceedings of Machine Learning Research}, publisher = {PMLR}, }

Continual Facial Expression Recognition: A Benchmark

Nikhil Churamani, Tolga Dimlioglu, German I. Parisi, and Hatice Gunes

2023

arXiv Bib

@article{Churamani2023ConFER,
  title = {Continual Facial Expression Recognition: A Benchmark},
  author = {Churamani, Nikhil and Dimlioglu, Tolga and Parisi, German I. and Gunes, Hatice},
  year = {2023},
}