Abstract: Educational simulations that allow learners to explore open-ended, real-world complex system problems creatively, are being prolifically used in formal and informal learning settings. These open-ended problems are characterized by loosely defined reward functions that allow a range of: solutions states, strategies, and actions, many of which are often unknown a priori. Additionally, the dynamic emergent states in complex systems demand that the learners co-evolve their solutions with the nonlinear system states (as path dependencies prevent backtracking). Providing assessment based guidance to learners is counterintuitive when used with such problems, as the methods oversimplify the complexity and “close” the open-endedness of the problems, preventing the explorations we wish to support. This talk will discuss a fundamentally new approach that relies on structural causal models (SCM) to generate conceptually novel “exploration support” for learners exploring complex problems and demonstrate this in the context of a museum exhibit that engages learners collaboratively while managing a complex, open-ended simulated ecosystem. We discuss the potential of this method for exploring the new avenue of constructing “just in time” causal explanations relevant to the learner’s current situation that could serve as novel “exploration support” and conclude with future plans for generating tools based on exploration support.

Short Bio: Aditi Mallavarapu is a Postdoctoral Research Associate at the INVITE AI institute at University of Illinois, Urbana Champaign. Previously, she received postdoctoral training at Digital Promise Global with CIRCLS, which is a NSF hub for emergent educational technology research and served as a visiting scholar at the University of Pittsburgh. She has earned the M.S and PhD degree in Computer Science at the University of Illinois, Chicago. She specializes in innovating and implementing human-centered computational techniques in open-ended learning settings that support exploration-based learning (e.g.,in museum exhibits and serious games) and (2) investigating interdisciplinary research communities using network science for studying intentional collaboration (an effort applauded by the evaluation team at the NSF). One of her early works published in the Journal of Educational Data Mining was regarded as one of the first to apply educational data mining techniques to reveal patterns of learner explorations in open-ended learning environments. Her interdisciplinary work contributes intellectually across Learning Analytics, AIED, Human-Computer Interaction, Network Science and Complex Systems Education. Working alongside interdisciplinary researchers, educators, learners and other practitioners, with the technical data-driven approaches, Dr. Mallavarapu is driven to improve and aid learning in open-ended learning environments designed to engage learners in real-world complex problems.