Human-in-the-loop optimization identifies optimal interface designs by iteratively observing user performance. However, it often requires numerous iterations due to the lack of prior information. While recent approaches have accelerated this process by leveraging previous optimization data, collecting user data remains costly and often impractical. We present a conceptual framework, Human-in-the-Loop Optimization with Model-Informed Priors (HOMI), which augments human-in-the-loop optimization with a training phase where the optimizer learns adaptation strategies from diverse, synthetic user data generated with predictive models before deployment. To realize HOMI, we introduce Neural Acquisition Function+ (NAF+), a Bayesian optimization method featuring a neural acquisition function trained with reinforcement learning. NAF+ learns optimization strategies from large-scale synthetic data, improving efficiency in real-time optimization with users. We evaluate HOMI and NAF+ with mid-air keyboard optimization, a representative VR input task. Our work presents a new approach for more efficient interface adaptation by bridging in situ and in silico optimization processes.

Publisher Copyright: © 2026 Copyright held by the owner/author(s).

Selecting a target in a 3D environment is often challenging, especially with small/distant targets or when sensor noise is high. To facilitate selection, target-inference methods must be accurate, fast, and account for noise and motor variability. However, traditional data-free approaches fall short in accuracy since they ignore variability. While data-driven solutions achieve higher accuracy, they rely on extensive human datasets so prove costly, time-consuming, and transfer poorly. In this paper, we propose a novel approach that leverages biomechanical simulation to produce synthetic motion data, capturing a variety of movement-related factors, such as limb configurations and motor noise. Then, an inference model is trained with only the simulated data. Our simulation-based approach improves transfer and lowers cost; variety-rich data can be produced in large quantities for different scenarios. We empirically demonstrate that our method matches the accuracy of human-data-driven approaches using data from seven users. When deployed, the method accurately infers intended targets in challenging 3D pointing conditions within 5–10 milliseconds, reducing users’ target-selection error by 71% and completion time by 35%.

There have been significant advances in simulation models predicting human behavior across various interactive tasks. One issue remains, however: identifying the parameter values that best describe an individual user. These parameters often express personal cognitive and physiological characteristics, and inferring their exact values has significant effects on individual-level predictions. Still, the high complexity of simulation models usually causes parameter inference to consume prohibitively large amounts of time, as much as days per user. We investigated amortized inference for its potential to reduce inference time dramatically, to mere tens of milliseconds. Its principle is to pre-train a neural proxy model for probabilistic inference, using synthetic data simulated from a range of parameter combinations. From examining the efficiency and prediction performance of amortized inference in three challenging cases that involve real-world data (menu search, point-and-click, and touchscreen typing), the paper demonstrates that an amortized-inference approach permits analyzing large-scale datasets by means of simulation models. It also addresses emerging opportunities and challenges in applying amortized inference in HCI.

The simulation of user behavior with deep reinforcement learning agents has shown some recent success. However, the inverse problem, that is, inferring the free parameters of the simulator from observed user behaviors, remains challenging to solve. This is because the optimization of the new action policy of the simulated agent, which is required whenever the model parameters change, is computationally impractical. In this study, we introduce a network modulation technique that can obtain a generalized policy that immediately adapts to the given model parameters. Further, we demonstrate that the proposed technique improves the efficiency of user simulator-based inference by eliminating the need to obtain an action policy for novel model parameters. We validated our approach using the latest user simulator for point-and-click behavior. Consequently, we succeeded in inferring the user’s cognitive parameters and intrinsic reward settings with less than 1/1000 computational power to those of existing methods.

The recently advanced robotics technology enables robots to assist users in their daily lives. Haptic guidance (HG) improves users’ task performance through physical interaction between robots and users. It can be classified into optimal action-based HG (OAHG), which assists users with an optimal action, and user prediction-based HG (UPHG), which assists users with their next predicted action. This study aims to understand the difference between OAHG and UPHG and propose a combined HG (CombHG) that achieves optimal performance by complementing each HG type, which has important implications for HG design. We propose implementation methods for each HG type using deep learning-based approaches. A user study (n=20) in a haptic task environment indicated that UPHG induces better subjective evaluations, such as naturalness and comfort, than OAHG. In addition, the CombHG that we proposed further decreases the disagreement between the user intention and HG, without reducing the objective and subjective scores.