arXiv subjects

Seoyeon Choi

Publications and source records attributed to Seoyeon Choi.

6 recordsLinked to original sources

arXiv2026

A Guiding Framework for K-12 Teachers in Creating AI-powered Learning Technologies through Vibe Coding

Large language models generate code from natural language prompts, enabling "vibe coding," which allows non-programmers to develop computational solutions. Vibe coding for teachers amplifies the value of teachers-as-designers, improving technology integration while fostering AI literacy. However, structured guidance on supporting this process is lacking. We propose GAIDE (A Guiding Framework for AI-Integrated Design for Educators), a framework that supports K-12 teachers in creating AI-powered learning technologies through vibe coding. The initial framework, built on Design Thinking and INTERACT, was validated through a CORDTRA interaction analysis of three teachers and four faculty mentors in an eight-week workshop to derive the final framework. Additionally, the qualitative analysis of pre- and post-interviews found an enhancement of teachers' AI literacy. Findings highlight the potential of learning-by-creating for professional development.

cs.CY↗

arXiv2026

Overthinking Loops in Agents: A Structural Risk via MCP Tools

Tool-using LLM agents increasingly coordinate real workloads by selecting and chaining third-party tools based on text-visible metadata such as tool names, descriptions, and return messages. We show that this convenience creates a supply-chain attack surface: a malicious MCP tool server can be co-registered alongside normal tools and induce overthinking loops, where individually trivial or plausible tool calls compose into cyclic trajectories that inflate end-to-end tokens and latency without any single step looking abnormal. We formalize this as a structural overthinking attack, distinguishable from token-level verbosity, and implement 14 malicious tools across three servers that trigger repetition, forced refinement, and distraction. Across heterogeneous registries and multiple tool-capable models, the attack causes severe resource amplification (up to $142.4\times$ tokens) and can degrade task outcomes. Finally, we find that decoding-time concision controls do not reliably prevent loop induction, suggesting defenses should reason about tool-call structure rather than tokens alone.

cs.CL↗

arXiv2025

MIMIC-D: Multi-modal Imitation for MultI-agent Coordination with Decentralized Diffusion Policies

As robots become more integrated in society, their ability to coordinate with other robots and humans on multi-modal tasks (those with multiple valid solutions) is crucial. Such behaviors can be learned from expert demonstrations via imitation learning (IL), but when expert demonstrations are multi-modal, standard IL approaches usually average across modes or collapse to a single mode, preventing effective coordination. Being inspired by diffusion models' ability to capture complex multi-modal trajectory distributions in single-agent settings, we develop a diffusion-based framework for coordinated multi-modal behavior in multi-agent systems. However, existing multi-agent diffusion approaches typically require a centralized planner or explicit communication among agents. This assumption can fail in real-world scenarios where robots must operate independently or with agents like humans that they cannot directly communicate with. Therefore, we propose MIMIC-D, a joint training with decentralized execution paradigm for multi-modal multi-agent IL via diffusion. We jointly train all agents' policies with only local information to achieve implicit coordination. In simulation and hardware experiments, our method exhibits robust multi-modal coordination behavior in various tasks and environments, improving upon state-of-the-art baselines.

cs.RO↗

arXiv2025

CRAFT: Coaching Reinforcement Learning Autonomously using Foundation Models for Multi-Robot Coordination Tasks

Multi-Agent Reinforcement Learning (MARL) provides a powerful framework for learning coordination in multi-agent systems. However, applying MARL to robotics remains challenging due to their high-dimensional continuous joint action spaces, complex reward design, and non-stationarity from concurrently learning agents. On the other hand, humans often learn complex coordination with the help of coaches, who guide learning through carefully designed curricula and detailed feedback. Building on the reasoning capabilities of foundation models, we argue that these models can similarly coach robots to learn coordination. Motivated by this, we propose CRAFT: Coaching Reinforcement learning Autonomously using Foundation models for learning coordination Tasks, a framework that leverages foundation models to act as a "coach" for multi-robot coordination. CRAFT automatically decomposes long-horizon coordination tasks into sequences of subtasks using the planning capability of Large Language Models (LLMs). Then, CRAFT trains each subtask using LLM-generated reward functions, and refines them through a Vision Language Model (VLM)-guided reward-refinement loop. We evaluate CRAFT on multi-quadruped navigation and bimanual manipulation tasks, and demonstrate its capability to learn complex coordination behaviors. In addition, in a multi-quadruped navigation setting, we show that our learned policies transfer to the real world. Project website is https://iconlab.negarmehr.com/CRAFT/

cs.RO↗

arXiv2025

EquiContact: A Hierarchical SE(3) Vision-to-Force Equivariant Policy for Spatially Generalizable Contact-rich Tasks

This paper presents a framework for learning vision-based robotic policies for contact-rich manipulation tasks that generalize spatially across task configurations. We focus on achieving robust spatial generalization of the policy for the peg-in-hole (PiH) task trained from a small number of demonstrations. We propose EquiContact, a hierarchical policy composed of a high-level vision planner (Diffusion Equivariant Descriptor Field, Diff-EDF) and a novel low-level compliant visuomotor policy (Geometric Compliant ACT, G-CompACT). G-CompACT operates using only localized observations (geometrically consistent error vectors (GCEV), force-torque readings, and wrist-mounted RGB images) and produces actions defined in the end-effector frame. Through these design choices, we show that the entire EquiContact pipeline is SE(3)-equivariant, from perception to force control. We also outline three key components for spatially generalizable contact-rich policies: compliance, localized policies, and induced equivariance. Real-world experiments on PiH, screwing, and surface wiping tasks demonstrate a near-perfect success rate and robust generalization to unseen spatial configurations, validating the proposed framework and principles. The experimental videos and more details can be found on the project website: https://equicontact.github.io/EquiContact-website/

cs.RO↗

arXiv2019

Intensive Monitoring Survey of Nearby Galaxies (IMSNG)

Intensive Monitoring Survey of Nearby Galaxies (IMSNG) is a high cadence observation program monitoring nearby galaxies with high probabilities of hosting supernovae (SNe). IMSNG aims to constrain the SN explosion mechanism by inferring sizes of SN progenitor systems through the detection of the shock-heated emission that lasts less than a few days after the SN explosion. To catch the signal, IMSNG utilizes a network of 0.5-m to 1-m class telescopes around the world and monitors the images of 60 nearby galaxies at distances D < 50 Mpc to a cadence as short as a few hours. The target galaxies are bright in near-ultraviolet (NUV) with M_NUV < -18.4 AB mag and have high probabilities of hosting SNe (0.06 SN/yr per galaxy). With this strategy, we expect to detect the early light curves of 3.4 SNe per year to a depth of R ~ 19.5 mag, enabling us to detect the shock-heated emission from a progenitor star with a radius as small as 0.1 R_sun. The accumulated data will be also useful for studying faint features around the target galaxies and other science projects. So far, 18 SNe have occurred in our target fields (16 in IMSNG galaxies) over 5 years, confirming our SN rate estimate of 0.06 SN/yr per galaxy.

astro-ph.GA↗