ABOUT ME
I am a Postdoctoral Research Associate at the University of Cambridge, UK. I received my Ph.D. degree in Computer Science from KTH Royal Institute of Technology, Sweden, in 2023. During my doctoral studies, I was a visiting scholar at the Georgia Institute of Technology, US (Nov 2021 – Apr 2022). I hold M.Sc. (2018) and B.Sc. (2015) degrees in Computer Engineering from Middle East Technical University (METU), Turkey.
While pursuing my M.Sc. degree, I began conducting research at the Kovan Robotics Lab with Prof. Sinan Kalkan. During my doctoral education, I continued robotics research at the Division of Robotics, Perception and Learning under the supervision of Prof. Iolanda Leite. As part of my doctoral degree, I had the opportunity to collaborate with Prof. Sonia Chernova's research group, the Robot Autonomy and Interactive Learning Lab. In my current postdoctoral research, I work with Prof. Hatice Güneş in the Affective Intelligence and Robotics Lab.
RESEARCH
My research aims to deploy autonomous robots in human environments to facilitate people's lives. To achieve this goal, there are several research challenges, such as understanding users (e.g., comprehending their natural language instructions and analyzing their nonverbal cues) and observing the interaction environment (e.g., the context for social appropriateness, the visual scene, and the semantic properties of the objects in the environment). My research has focused on such varying components to achieve my long-term vision.
During my Ph.D., I worked on natural language grounding for robots. More specifically, I addressed how to understand user instructions and resolve ambiguities with follow-up clarifications. My Ph.D. thesis contributed to efficient human-robot conversation and suggested a full pipeline for object search in a household environment, first finding the described object location and then identifying the specified object in the user instruction. The main takeaway of my Ph.D. research is that ambiguities are inevitable in HRI, and instead of asking users to repeat their request over and over again under uncertainties (as in many existing language-based systems such as Amazon Alexa or Siri), leveraging follow-up clarifications is efficient for HRI.
My ongoing work not only further explores my doctoral research topics but also focuses on using different modalities and learning approaches while maintaining my main focus on autonomous robots that help people. Currently, I focus on continual learning for robots in human environments leveraging explainability to generate socially appropriate robot behaviour.
CONTACT ME
William Gates Building
Cambridge CB3 0FD, UK
fid21@cam.ac.uk
irmakdogan92@gmail.com