arXiv subjects

Hannah Selder

Publications and source records attributed to Hannah Selder.

6 recordsLinked to original sources

arXiv2026

Visual Bias in Simulated Users: The Impact of Luminance and Contrast on Reinforcement Learning-based Interaction

Reinforcement learning (RL) enables simulations of HCI tasks, yet their validity is questionable when performance is driven by visual rendering artifacts distinct from interaction design. We provide the first systematic analysis of how luminance and contrast affect behavior by training 247 \RV{simulated users using RL} on pointing and tracking tasks. We vary the luminance of task-relevant objects, distractors, and background under no distractor, static distractor, and moving distractor conditions, and evaluate task performance and robustness to unseen luminances. Results show luminance becomes critical with static distractors, substantially degrading performance and robustness, whereas motion cues mitigate this issue. Furthermore, robustness depends on preserving relational ordering between luminances rather than matching absolute values. Extreme luminances, especially black, often yield high performance but poor robustness. Overall, seemingly minor luminance changes can strongly shape learned behavior, revealing critical insights into what RL-driven simulated users actually learn.

cs.HC↗

arXiv2026

MyoInteract: A Framework for Fast Prototyping of Biomechanical HCI Tasks using Reinforcement Learning

Reinforcement learning (RL)-based biomechanical simulations have the potential to revolutionise HCI research and interaction design, but currently lack usability and interpretability. Using the Human Action Cycle as a design lens, we identify key limitations of biomechanical RL frameworks and develop MyoInteract, a novel framework for fast prototyping of biomechanical HCI tasks. MyoInteract allows designers to setup tasks, user models, and training parameters from an easy-to-use GUI within minutes. It trains and evaluates muscle-actuated simulated users within minutes, reducing training times by up to 98%. A workshop study with 12 interaction designers revealed that MyoInteract allowed novices in biomechanical RL to successfully setup, train, and assess goal-directed user movements within a single session. By transforming biomechanical RL from a days-long expert task into an accessible hour-long workflow, this work significantly lowers barriers to entry and accelerates iteration cycles in HCI biomechanics research.

cs.HC↗

arXiv2025

Attention Trajectories as a Diagnostic Axis for Deep Reinforcement Learning

The emergence and evolution of feature reliance in deep reinforcement learning agents remain poorly understood. Here, we introduce a methodological framework for analyzing the learning process through quantitative analysis of saliency maps. This approach aggregates saliency information at the object and modality level into hierarchical attention profiles, quantifying how agents allocate attention over time, thereby forming attention trajectories throughout training. These profiles are then compared across controlled conditions, connected to behavioral measurements and reproduced with different saliency methods to assess the robustness of the findings. Applied to Atari 2600 benchmarks, custom Pong environments, and biomechanical user simulations in visuomotor tasks, this framework uncovers algorithm-specific attention biases, diagnosed unintended reward-driven strategies, and overfitting to redundant sensory channels. These patterns correspond to measurable behavioral differences, demonstrating empirical links between attention profiles, learning dynamics, and agent behavior. The results establish attention trajectories as a promising diagnostic axis for tracing how feature reliance develops during training and for identifying biases and vulnerabilities invisible to performance metrics alone.

cs.LG↗

arXiv2025

Stability and performance of stochastic economic MPC -- Stochastic characterization of the closed-loop asymptotics

Model Predictive Control (MPC) is well understood in the deterministic setting, yet rigorous stability and performance guarantees for stochastic MPC remain limited to the consideration of terminal constraints and penalties. In contrast, this work analyzes stochastic economic MPC with an expected cost criterion and establishes closed-loop guarantees without terminal conditions. Relying on stochastic dissipativity and turnpike properties, we construct closed-loop Lyapunov functions that ensure $P$-practical asymptotic stability of a particular optimal stationary process under different notions of stochastic convergence, such as in distribution or in the $p$-th mean. In addition, we derive tight near-optimal bounds for both averaged and non-averaged performance, thereby extending classical deterministic results to the stochastic domain. Finally, we show that the abstract stochastic MPC scheme requiring distributional knowledge shares the same closed-loop properties as a practically implementable algorithm based only on sampled state information, ensuring applicability of our findings. Our findings are illustrated by a numerical example.

math.OC↗

arXiv2025

Demystifying Reward Design in Reinforcement Learning for Upper Extremity Interaction: Practical Guidelines for Biomechanical Simulations in HCI

Designing effective reward functions is critical for reinforcement learning-based biomechanical simulations, yet HCI researchers and practitioners often waste (computation) time with unintuitive trial-and-error tuning. This paper demystifies reward function design by systematically analyzing the impact of effort minimization, task completion bonuses, and target proximity incentives on typical HCI tasks such as pointing, tracking, and choice reaction. We show that proximity incentives are essential for guiding movement, while completion bonuses ensure task success. Effort terms, though optional, help refine motion regularity when appropriately scaled. We perform an extensive analysis of how sensitive task success and completion time depend on the weights of these three reward components. From these results we derive practical guidelines to create plausible biomechanical simulations without the need for reinforcement learning expertise, which we then validate on remote control and keyboard typing tasks. This paper advances simulation-based interaction design and evaluation in HCI by improving the efficiency and applicability of biomechanical user modeling for real-world interface development.

cs.HC↗

arXiv2025

What Makes a Model Breathe? Understanding Reinforcement Learning Reward Function Design in Biomechanical User Simulation

Biomechanical models allow for diverse simulations of user movements in interaction. Their performance depends critically on the careful design of reward functions, yet the interplay between reward components and emergent behaviours remains poorly understood. We investigate what makes a model "breathe" by systematically analysing the impact of rewarding effort minimisation, task completion, and target proximity on movement trajectories. Using a choice reaction task as a test-bed, we find that a combination of completion bonus and proximity incentives is essential for task success. Effort terms are optional, but can help avoid irregularities if scaled appropriately. Our work offers practical insights for HCI designers to create realistic simulations without needing deep reinforcement learning expertise, advancing the use of simulations as a powerful tool for interaction design and evaluation in HCI.

cs.HC↗