Robin Schmucker
I am a PhD candidate in the Machine Learning Department at Carnegie Mellon University (CMU), where
I am fortunate to be advised by Professor Tom Mitchell.
Previously, I spent three wonderful years at the Karlsruhe Institute of Technology,
completing my bachelor's degree in Computer Science. I was a research assistant and part of the TECO research group led by Professor Michael Beigl.
I am grateful for the mentorship of Professor Manuela Veloso who hosted me as a research intern as part of the CLICS scholarship program.
In the industry, I worked as a research intern at AWS where I designed new algorithms and contributed to the AutoGluon project.
|
|
Research
My research interests are broadly in the areas of machine learning and optimization. Particularly, I am interested in the theoretical
and practical problems that arise when deploying data-driven paradigms in real world settings such as education and healthcare.
Currently, my research focuses on problems related to intelligent tutoring systems including student knowledge modeling, performance predictions,
personalized curriculum design and emerging types of learning technologies.
Research opportunities: I am happy to collaborate, discuss research and answer questions about CMU's academic programs. If you are interested, please feel free to send me an email.
|
Selected Publications
Visit Google Scholar for a complete and up-to-date list of publications.
|
|
Gaining Insights into Course Difficulty Variations Using Item Response Theory
Frederik Baucks*, Robin Schmucker*, Laurenz Wiskott
To appear: Int. Conf. on Learning Analytics & Knowledge, 2024
Common curriculum analytics techniques assume time-invariant course behavior and
are unable to capture temporal variations in the data. We introduce an item response theory
(IRT)-based methodology to address the open problem of monitoring course difficulty
changes over time to ensure the equal treatment of student cohorts and GPA consistency.
|
|
Ruffle&Riley: Towards the Automated Induction of Conversational Tutoring Systems
Robin Schmucker, Meng Xia, Amos Azaria, Tom Mitchell
NeurIPS'23 Workshop: Generative AI for Education (GAIED) , 2023
Conversational tutoring systems (CTSs) promote cognitive engagement and benefit learning outcomes.
Still, the time and cost required to author CTS content is a major obstacle to widespread adoption.
We leverage LLMs as foundation for a novel type of CTSs that can automatically induce and orchestrate coherent free-form adaptive dialogues.
|
|
Learning to Give Useful Hints: Assistance Action Evaluation and Policy Improvements
Robin Schmucker, Nimish Pachapurkar, Shanmuga Bala, Miral Shah, Tom Mitchell
European Conf. on Technology Enhanced Learning , 2023
We employ a multi-armed bandit framework inside an online tutoring system to learn how to provide students with effective hints.
Using data from over 190,000 students, we quantify effects on various learning outcome measures.
A live use evaluation of the optimized assistance policies showed significant improvements in student learning outcomes.
|
|
Transferable Student Performance Modeling for Intelligent Tutoring Systems
Robin Schmucker,
Tom M. Mitchell
Int. Conf. on Computers in Education, 2022 (best paper nominee)
We propose transfer learning techniques that can mitigate the student
performance modeling cold-start problem for new courses by leveraging
log data from existing courses. Our course-agnostic
models enable accurate predictions for future courses when they are first deployed.
|
|
Assessing the Performance of Online Students - New Data, New Approaches, Improved Accuracy
Robin Schmucker,
Jingbo Wang,
Shijia Hu,
Tom M. Mitchell
Journal of Educational Data Mining, 2022
Video, GitHub
We study how to utilize various types of student
log data for performance modeling using four recent large-scale
datasets. We propose various extensions over earlier methods and
define a new state of the art for logistic regression-based performance modeling.
|
|
Combination Treatment Optimization Using a Pan-Cancer Pathway Model
Robin Schmucker,
Gabriele Farina,
James Faeder,
Fabian Fröhlich,
Ali Sinan Saglam,
Tuomas Sandholm
PLOS Computational Biology, 2021
We use a pan-cancer pathway model to identify novel combination
therapies by defining multiple treatment optimization problems and solving
them by combining CMA-ES with an efficient Hamiltonian Monte-Carlo sampling scheme.
We also consider sequential treatment plans.
|
|
Bandit Linear Optimization for Sequential Decision Making and Extensive-Form Games
Gabriele Farina,
Robin Schmucker,
Tuomas Sandholm
AAAI Conference on Artificial Intelligence, 2021
We propose the first algorithm for the bandit linear optimization problem for
tree-form sequential decision making that offers both (i) linear-time iterations
(in the size of the decision tree) and (ii) O(√T) cumulative regret in
expectation compared to any fixed strategy, at all times T.
|
|