Belén Martín-Urcelay

Ph.D. student in Computer Engineering at Georgia Tech.

burcelay3 [AT] gatech.edu

Bio

I am an ML & AI Intern at Hippo Harvest and an incoming Postdoctoral Researcher at the University of Stuttgart, where I will work with Andreas Bulling starting in September. My research focuses on efficient and adaptive human-in-the-loop learning, with a particular interest in leveraging implicit multimodal signals, such as eye gaze, to reduce the cost of human feedback and improve the alignment of machine learning systems.

I recently completed my Ph.D. in Computer Engineering at Georgia Tech, advised by Matthieu Bloch and Christopher Rozell. I began my academic career at Universidad de Navarra, where I obtained both my B.Sc. and M.Sc. in Telecommunications Engineering. . I have also had the opportunity to collaborate internationally, including a research visit to ETH Zurich with Andreas Krause. More broadly, my research interests include active learning, human feedback, and multimodal learning. Outside of research, I am passionate about teaching, mentoring students, and promoting women in STEM.

Research

Most recent publications on Google Scholar.
‡ indicates equal contribution.

Beyond Labels: Information-Efficient Human-in-the-Loop Learning using Ranking and Selection Queries

Belén Martín-Urcelay, Yoonsang Lee, Matthieu R. Bloch, Christopher J. Rozell

Under review. Submitted February 2026.

Paper Code

From Words To Rewards: Leveraging Natural Language For Reinforcement Learning.

Belén Martín-Urcelay, Andreas Krause, Giorgia Ramponi

TMLR '26 : Transactions on Machine Learning Research. January 2026.

Paper Code

Online Machine Teaching under Learner's Uncertainty: Gradient Descent Learners of a Quadratic Loss

Belén Martín-Urcelay, Christopher J. Rozell, Matthieu R. Bloch

SIMODS '25: SIAM Journal on Mathematics of Data Science. June 2025.

Paper Code

Preference Instability in Reward Models: Detection and Mitigation via Sparse Autoencoders

Shunchang Liu, Xin Chen, Belén Martín-Urcelay, Francesco Croce

Under review. Submitted May 2026.

Paper Code

Beyond Labels: Information-Efficient Human-in-the-Loop Learning using Ranking and Selection Queries

Belén Martín-Urcelay, Yoonsang Lee, Matthieu R. Bloch, Christopher J. Rozell

Under review. Submitted February 2026.

Paper Code

From Words To Rewards: Leveraging Natural Language For Reinforcement Learning.

Belén Martín-Urcelay, Andreas Krause, Giorgia Ramponi

TMLR '26 : Transactions on Machine Learning Research. January 2026.

Paper Code

MANGO: Disentangled Image Transformation Manifolds with Grouped Operators

Brighton Ancelin‡, Yenho Chen‡, Alex Saad-Falcon‡, Peimeng Guan‡, Chiraag Kaushik‡, Nakul Singh‡, Belén Martín-Urcelay‡

SampTA '25 [Oral]: Sampling Theory and Applications. July 2025.

Paper Slides

Online Machine Teaching under Learner's Uncertainty: Gradient Descent Learners of a Quadratic Loss

Belén Martín-Urcelay, Christopher J. Rozell, Matthieu R. Bloch

SIMODS '25: SIAM Journal on Mathematics of Data Science. June 2025.

Paper Code

Enhancing Human-in-the-Loop Learning for Binary Sentiment Word Classification

Belén Martín-Urcelay, Christopher J. Rozell, Matthieu R. Bloch

CDC '24: Conference on Decision and Control. December 2024.

Paper Slides Code

Reinforcement Learning from Human Text Feedback: Learning a Reward Model from Human Text Input.

Belén Martín-Urcelay, Andreas Krause, Giorgia Ramponi

ICML '24 Workshop on Models of Human Feedback for AI Alignment. July 2024.

Paper Poster

Temperature-Dependent I/Q Imbalance Compensation in Ultra-Wideband Millimeter-Wave Multi-Gigabit Transmitters

Ainhoa Rezola, Juan F. Sevillan, David del Río, Belén Martín-Urcelay, Iñaki Gurutzeaga, Igone Vélez, Roc Berenguer

IEEE Transactions on Microwave Theory and Techniques. January 2020.

Paper

Teaching Experience

Introduction to Probability and Statistics for ECE (ECE 3077)
  • Introduction to probability, random variables, distributions, estimation, confidence intervals, linear regression and other tools for describing and managing uncertainty in electrical and computer engineering.
  • Lecture and develop course materials for 60 third year undergraduate engineering students.
  • Outstanding ECE Graduate Teaching Award.

Instructor of Record | Georgia Tech. | January 2025 - May 2025

Leadership

Mentoring
I have been fortunate to supervise and collaborate with the following students at Georgia Tech:
  • Yoonsan Lee – Focused on statistical analysis.
  • Zhihao Qin – Focused on experiment design.
  • Vien Tran – Focused on experimental infrastructure and data collection.
  • John Burchfield – Focused on parallel computating and data formatting.
Graduate Chair, Women in Electrical and Computer Engineering (WECE)
Founded and held the Graduate Chair role from 2021 to 2026 at Georgia Tech, leading initiatives to build community among female graduate students in ECE. Primary focus on mentoring and networking to foster personal and professional growth.
Workshop Organization

Vitæ

Full Resume in PDF.

  • Hippo Harvest Summer 2026
    Machine Learning Intern
    AI for greenhouse optimization
  • Georgia Tech. 2022 - 2026
    Ph.D. Student
    Advisors: Christopher Rozell and Matthieu Bloch
  • ETH Zurich Fall 2023
    Guest Graduate Researcher
    Advisor: Andreas Krause
  • Georgia Tech. 2020 - 2022
    M.Sc. Student
    Electrical and Computer Engineering
  • University of Sheffield Spring 2020
    Master Thesis
    Advisor: Iñaki Esnaola
  • Universidad de Navarra 2019 - 2020
    M.Sc. Student
    Telecommunications Engineering
  • Developair Jul 2019 - Jan 2020
    Software Developer
    Automatic simulation parameter adjustment tools
  • Fraunhofer IIS Sep-Dec 2018
    Research Assistant
    Audio Department
I made this webpage using Martin Saveski's template, available here.
Belén Martín-Urcelay
burcelay3 [AT] gatech.edu