Peter Schaldenbrand

Research Programmer | Incoming PhD Student
Carnegie Mellon University
pschalde at cs dot cmu dot edu

Conference Papers

Content Masked Loss: Human-Like Brush Stroke Planning in a Reinforcement Learning Painting Agent.

[Accepted] 35th AAAI Conf. on Artificial Intelligence
arXiv pre-print
Peter Schaldenbrand and Jean Oh
Novel approach to increasing the human-like planning of a reinforcement learning model.

Computer-Supported Human Mentoring for Personalized and Equitable Math Learning

[Accepted] 2021 Conf. on Artificial Intelligence in Education (AIED21)
Peter Schaldenbrand, Nikki G. Lobczowski, J. Elizabeth Richey, Shivang Gupta, Elizabeth A. McLaughlin, Adetunji Adeniran, and Kenneth R. Koedinger
Novel approach to increasing the human-like planning of a reinforcement learning model.

Is the Doer Effect Robust Across Multiple Data Sets?

11th Conf. on Educational Data Mining
Kenneth Koedinger, Richard Scheines, and Peter Schaldenbrand
Establishing the causal effect between active resources in an educational environment and learning.

CVPR'21 Workshop on Bridging the Gap between Subjective and Computational Measurements of Machine Creativity

[Accepted] The 2021 Conference on Computer Vision and Pattern Recognition
Ahmed Elgammal (Rutgers) Hyeju Jang (U. British Columbia) Eunsu Kang James McCann (CMU) Jean Oh (CMU) Devi Parikh (Georgia Tech) Peter Schaldenbrand Robert Twomey (UCSD) and Jun-Yan Zhu (CMU)
We are organizing a conference to highlight the differences between the way that humans and machines evalaute creativity, and begin a conversation on how to bridge this gap.

RSS'21 Robotics x Arts Exhibition: Opportunities and Issues in Robotics Applied in the Arts

Sang Leigh (Georgia Tech), Jean Oh (CMU), Frank Dellaert (GT), Seth Hutchinson (GT), Peter Schaldenbrand, Gerry Cehn (GT), and Juan D Florez-Castillo (GT)

Workshop: Causal Discovery with Tetrad in LearnSphere’s Tigris

12th Conf. on Educational Data Mining
Richard Scheines, Peter Schaldenbrand, and Kenneth Koedinger

Machine Traditional Animation: AniPainter

Paper Website
Peter Schaldenbrand and Jean Oh
Using an affordable robot arm to "hand" paint animation frames.

Visualizing Spotify Usage Data

Static Site GitHub
Peter Schaldenbrand and Paulina Davison
Interactive visualizations of a user's Spotify data using Altair, Vega, and Streamlit.

wav2midi: An Unsupervised Music Format Translation Model

Peter Schaldenbrand and Husni Almoubayyed
Convert arbitrary waveform audio music into MIDI format. Trained without paired wav-midi examples.

Video Style Transfer

Video style transfer using Style Transfer by Relaxed Optimal Transport and Self-Similarity.

Non-Differentiable, Parametric Model Optimization via Causal Discovery

[In Progress] Optimize the parameters of a model using causal discovery algorithms. Allows for non-differentiable activation functions in the model.


Aaron Hertzmann's algorithm for stroke-based rendering in Python.

Music Genre Classification Using Deep Learning Techniques

Peter Schaldenbrand and Yizhou He

Deep Learning for Music Generation

Paper Poster
Peter Schaldenbrand, Husni Almoubayyed, and Yizhou He

Real Time Learning Analytics for Intelligent Tutoring Systems

Kenneth Holstein, Zac Yu, Octav Popescu, Jonathan Sewall, Peter Schaldenbrand, Bruce M. McLaren, and Vincent Aleven

A Deep Dive into Water Contamination

Citadel - The Data Open 2018
Peter Schaldenbrand, Husni Almoubayyed, Zecong Hu, and Xinyi Ju


In light of recent protests: Anonymize photographs by erasing humans from them.
I am a graduate student and technical staff member of Carnegie Mellon University's Human-Computer Interaction Institute.
  • At work, my goal is to address the opportunity gap that marginalized students face by creating technologies that make teaching and mentoring more effective and efficient.
  • At school, I study machine learning, computer vision, and robotics to aid my research.
  • In research, I am interested in making machine learning models that are capable of performing creative tasks such as image generation, painting, and video creation.
  • Research Programmer / Analyst

    Carnegie Mellon University
    Human-Computer Interaction Institute
    2017 — Present
    Personalized Learning²
    Full-stack, lead developer on the Personalized Learning² project; addressing the opportunity gap for marginalized students. Created an interface for teachers and mentors at public schools in the Pittsburgh region to view data from their students’ usage of multiple Math tutoring software and access resources to intervene on students’ motivational and cognitive obstacles. Gave input on the designing of the tool and managed multiple design interns throughout its development process.
    Learnsphere Tigris
    Contributed improvements in full stack development to LearnSphere’s Tigris: An online workflow tool for analyzing, visualizing, and mining educational data
    Presented and taught at workshops and CMU courses. Created videos to teach the software tools I contributed to.

    Carnegie Mellon University
    Robotics Institute

    Doctor of Philosophy
    August 2021 —

    Carnegie Mellon University
    Language Technologies Institute

    Master of Computational Data Science — Analytics
    GPA: 3.9
    2020 — Dec 2021

    The University of Pittsburgh

    Bachelor of Science: Computer Science
    GPA: 3.9
    2013 — 2017

    Central Catholic High School

    Pittsburgh, PA
    2009 — 2013


    Muscular Dystrophy Association
    Councelor and unit leader at MDA's summer camp for children with muscular diseases
    2012 — Present


    High school and college experience
    Assisted coach at Central Catholic High School
    You'll catch me outside on my bike as soon as the warm weather hits. I use it to find the best rock skipping spots in Western PA.


    Guitar, Saxophone, Bagpipes, Synthesizer in what small free time I find
    Composition: Short Film
    Background animation is a modified version of