Shayan Doroudi

• Carnegie Mellon University

Humans have the ability to reason about geometric patterns in images and scenes from a young age. However, developing large multimodal models (LMMs) capable of similar reasoning remains a challenge, highlighting the need for robust evaluation methods to assess these capabilities. We introduce TurtleBench, a benchmark designed to evaluate LMMs' capa...

Computational models of human learning can play a significant role in enhancing our knowledge about nuances in theoretical and qualitative learning theories and frameworks. There are many existing frameworks in educational settings that have shown to be verified using empirical studies, but at times we find these theories make conflicting claims or...

While object recognition is one of the prevalent affordances of humans' perceptual systems, even human infants can prioritize a place system over the object recognition system, that is used when navigating. This ability, combined with active learning strategies can make humans fast learners of Turtle Geometry, a notion introduced about four decades...

Purpose: Community college counselors and students use articulation agreement websites to (a) learn how community college courses will transfer and fulfill university requirements and (b) develop an academic plan to prepare to transfer. Compared to universities that do not have them, universities that do have articulation agreements provide more tr...

Artificial intelligence (AI) has become ubiquitous in recent decades, but using AI in education is not a recent idea. Since the emergence of AI, there has been an interest in the relationship between AI and education, not only looking at AI as an applied tool to advance education but also investigating its value as an analogy to human intelligence....

In the 30th anniversary of the International Artificial Intelligence in Education Society, there is a need to look back to the past to envision the community’s future. This paper presents a new framework (AI \(\times \) Ed) to categorize different interactions between artificial intelligence (AI) and education. We use our framework to compare paper...

The bias–variance tradeoff is a theoretical concept that suggests machine learning algorithms are susceptible to two kinds of error, with some algorithms tending to suffer from one more than the other. In this letter, we claim that the bias–variance tradeoff is a general concept that can be applied to human cognition as well, and we discuss implica...

Computer-assisted instructional programs such as intelligent tutoring systems are often used to support blended learning practices in K-12 education, as they aim to meet individual student needs with personalized instruction. While these systems have been shown to be effective under certain conditions, they can be difficult to integrate into pedago...

The development of educational AI (AIEd) systems has often been motivated by their potential to promote educational equity and reduce achievement gaps across different groups of learners -- for example, by scaling up the benefits of one-on-one human tutoring to a broader audience, or by filling gaps in existing educational services. Given these nob...

In higher education, predictive analytics can provide action-able insights to diverse stakeholders such as administrators, instructors, and students. Separate feature sets are typically used for different prediction tasks, e.g., student activity logs for predicting in-course performance and registrar data for predicting long-term college success. H...

Mastery learning algorithms are used in many adaptive learning technologies to assess when a student has learned a particular concept or skill. To assess mastery, some technologies utilize data-driven models while others use simple heuristics. Prior work has suggested that heuristics may often perform comparably to model-based algorithms. But is th...

In higher education, predictive analytics can provide actionable insights to diverse stakeholders such as administrators, instructors, and students. Separate feature sets are typically used for different layers of prediction, but little is known about the overall utility of different data sources across prediction tasks. Using data from nearly 2,00...

In 1988, a book was published by the name of "Technology in Education: Looking Toward 2020." Its purpose was to have thought leaders in educational research envision what the role of technology should be in 2020's educational landscape. By reflecting on how the visions of the book's authors align with the current state of educational technology, I...

Fallace (2019) claims that the notion of learning styles has roots in ethnocentrism. While some of the work on learning styles research does indeed seem ethnocentric, neither was this work the earliest work on learning styles nor was it seemingly influential on much of the later research on learning styles.

Since the 1960s, researchers have been trying to optimize the sequencing of instructional activities using the tools of reinforcement learning (RL) and sequential decision making under uncertainty. Many researchers have realized that reinforcement learning provides a natural framework for optimal instructional sequencing given a particular model of...

In many online environments, such as massive open online courses and crowdsourcing platforms, many people solve similar complex tasks. As a byproduct of solving these tasks, a pool of artifacts are created that may be able to help others perform better on similar tasks. In this paper, we explore whether work that is naturally done by crowdworkers c...

The potential for data-driven algorithmic systems to amplify existing social inequities, or create new ones, is receiving increasing popular and academic attention. A surge of recent work, across multiple researcher and practitioner communities, has focused on the development of design strategies and algorithmic methods to monitor and mitigate bias...

Adaptive educational technologies have the capacity to meet the needs of individual students in theory, but in some cases, the degree of personalization might be less than desired, which could lead to inequitable outcomes for students. In this paper, we use simulations to demonstrate that while knowledge tracing algorithms are substantially more eq...

We consider the problem of off-policy policy selection in reinforcement learning: using historical data generated from running one policy to compare two or more policies. We show that approaches based on importance sampling can be unfair---they can select the worse of two policies more often than not. We then give an example that shows importance s...

Sorting molecules with DNA robots Single-stranded DNA robots can move over the surface of a DNA origami sheet and sort molecular cargoes. Thubagere et al. developed a simple algorithm for recognizing two types of molecular cargoes and their drop-off destinations on the surface (see the Perspective by Reif). The DNA robot, which has three modular fu...

The gold standard for identifying more effective pedagogical approaches is to perform an experiment. Unfortunately, frequently a hypothesized alternate way of teaching does not yield an improved effect. Given the expense and logistics of each experiment, and the enormous space of potential ways to improve teaching, it would be highly preferable if...

Many interesting real world domains involve reinforcement learning (RL) in partially observable environments. Efficient learning in such domains is important, but existing sample complexity bounds for partially observable RL are at least exponential in the episode length. We give, to our knowledge, the first partially observable RL algorithm with a...

We explore how crowdworkers can be trained to tackle complex crowdsourcing tasks. We are particularly interested in training novice workers to perform well on solving tasks in situations where the space of strategies is large and workers need to discover and try different strategies to be successful. In a first experiment, we perform a comparison o...