Open-world missions often rely on repeated formulas, yet designers lack systematic ways to examine pacing, variation, and experiential balance across large portfolios. We introduce the Mission Action Quality Vector (MAQV), a six-dimensional framework—covering combat, exploration, narrative, emotion, problem-solving, and uniqueness—paired with an action block grammar representing missions as gameplay sequences. Using about 2200 missions from 20 AAA titles, we apply LLM-assisted parsing to convert community walkthroughs into structured action sequences and score them with MAQV. An interactive dashboard enables designers to reveal underlying mission formulas. In a mixed-methods study with experienced players and designers, we validate the pipeline’s fidelity and the tool’s usability, and use thematic analysis to identify recurring design trade-offs, pacing grammars, and systematic differences by quest type and franchise evolution. Our work offers a reproducible analytical workflow, a data-driven visualization tool, and reflective insights to support more balanced, varied mission design at scale.

Diffusion models excel at modeling complex and multi-modal trajectory distributions for decision-making and control. Reward-gradient guided denoising has been recently proposed to generate trajectories that maximize both a differentiable reward function and the likelihood under the data distribution captured by a diffusion model. Reward-gradient guided denoising requires a differentiable reward function fitted to both clean and noised samples, limiting its applicability as a general trajectory optimizer. In this paper, we propose Diffusion-ES, a method that combines gradient-free optimization with trajectory denoising to optimize black-box non-differentiable objectives while staying in the data manifold. Diffusion-ES samples trajectories during evolutionary search from a diffusion model and scores them using a black-box reward function. It mutates high-scoring trajecto-ries using a truncated diffusion process that applies a small number of noising and denoising steps, allowing for much more efficient exploration of the solution space. We show that Diffusion-Ex achieves state-of-the-art performance on nuPlan, an established closed-loop planning benchmark for autonomous driving. Diffusion-ES outperforms existing sampling-based planners, reactive deterministic or diffusion-based policies, and reward-gradient guidance. Additionally, we show that unlike prior guidance methods, our method can optimize non-differentiable language-shaped reward functions generated by few-shot LLM prompting. When guided by a human teacher that issues instructions to follow, our method can generate novel, highly complex behaviors, such as aggressive lane weaving, which are not present in the training data. This allows us to solve the hardest nuPlan scenarios which are beyond the capabilities of existing tra-jectory optimization methods and driving policies.11Project page: diffusion-es.github.io

Adversarial perturbations in object recognition and image classification tasks impact a learning model's ability to perform accurately and increase the safety risks of deployed machine learning models. During the adversarial example generation process, adversaries approach areas most prone to model uncertainty. Identifying partially occluded items, especially without understanding general object shapes, contributes to significant model uncertainty since object boundaries are not inherently at the forefront of the feature generalization process in deep learning models. Thus, this work aims to reduce model uncertainty surrounding partially occluded boundaries and increase adversarial robustness by augmenting the training dataset with amodal segmentation boundary masks. By observing performance degradation, robust sensitivity, and loss sensitivity, we show how including these masks during training impacts an adversary's ability to generate effective adversarial examples on the versatile MS COCO dataset. Lastly, we observe how including these masks during training influences the performance of adversarial training.

Significant progress has been made in training multimodal trajectory forecasting models for autonomous driving. However, effectively integrating these models with downstream planners and model-based control approaches is still an open problem. Although these models have conventionally been evaluated for open-loop prediction, we show that they can be used to parameterize autoregressive closed-loop models without retraining. We consider recent trajectory prediction approaches which leverage learned anchor embeddings to predict multiple trajectories, finding that these anchor embeddings can parameterize discrete and distinct modes representing high-level driving behaviors. We propose to perform fully reactive closed-loop planning over these discrete latent modes, allowing us to tractably model the causal interactions between agents at each step. We validate our approach on a suite of more dynamic merging scenarios, finding that our approach avoids the frozen robot problem which is pervasive in conventional planners. Our approach also outperforms the previous state-of-the-art in CARLA on challenging dense traffic scenarios when evaluated at realistic speeds.

Researchers have proposed that contagion processes govern how information and behavior itself spreads through social networks. Empirical evidence for such contagion often makes unjustified, but implicit assumptions about the mechanisms underlying contagion. Here, we present an approximate Bayesian method that uses empirical data to draw inferences about the underlying mechanisms. We provide initial validation of our approach in three simulation experiments, each investigating how a real-world factor (e.g., noise) impacts inferential accuracy.