As autonomous vehicles (AVs) approach Level 4 and Level 5 operational capability [SAE International, 2018], their on- board decision systems must handle not only safety-critical locomotion but also their subsequent moral weight. This paper details the Ethical Decision Head (EDH), a deep re- inforcement learning (RL) framework that encodes ethical reasoning as a differentiable reward signal, enabling a pol- icy gradient agent to learn morally-aligned driving behavior in scenarios whose state representation is aligned with the CARLA simulation environment [Dosovitskiy et al., 2017]. Two normative frameworks are instantiated and evaluated: a Utilitarian framework minimizing total casualties and a Kan- tian framework enforcing course maintenance as a categori- cal imperative. The EDH is trained via Proximal Policy Op- timization (PPO) [Schulman et al., 2017] against a Bradley- Terry reward model [Bradley and Terry, 1952] learned from pairwise human preference annotations over 200 collision- imminent scenarios. Results reveal an asymmetry in the learnability of normative ethical frameworks under human su- pervision. The Kantian condition, which reduces to a con- stant prediction task under the codebook, serves as a pipeline control: it confirms training stability and rules out infrastruc- ture failure as an explanation for the utilitarian result. The Utilitarian agent learned something more unsettling: human raters rewarded self-sacrifice over casualty minimization, and the model learned that preference faithfully. This divergence between what humans prescribe in theory and what they re- ward in practice suggests that RLHF does not learn ethics as philosophers define it, but as humans live it.
Video generation is central to AI-powered content creation. Aligning generated videos with human preferences is a key criterion for evaluating generation quality. Despite significant progress in visual quality, three key challenges remain. First, the reliability of reward signals is constrained by the quality of human preference data, which is often affected by subjective noise and bias. Second, standard scalar reward models collapse multi-aspect human preferences into a single value, leading to the loss of dynamic trade-offs across multiple preference dimensions. Third, in policy optimization, the widely adopted KL divergence imposes primarily local constraints and may fail to capture the global structure of human preferences. To address these challenges, we propose a unified preference-aware learning framework for video generation. First, we introduce elite-guided filtering to calibrate preference data and construct reliable supervision for reward model training. We then model video quality as a multidimensional reward distribution to capture the uncertainty inherent in human preferences, and use the Wasserstein distance to align the learned reward distribution with the empirical human preference distribution. Finally, we introduce Wasserstein-based distributional alignment into GRPO, guiding policy optimization to better match the global structure of human preferences over videos. Experiments on reward modeling and video generation demonstrate that our approach improves the reliability of reward signals and the perceptual consistency of generated videos. Our code is available at https://github.com/alignhs26/ahs.
Audio-driven 3D facial animation is essential for advancing immersion and interactivity in virtual experiences. Although recent advances have shown promising capabilities, the training and evaluation of existing methods typically rely on ground-truth-based errors, which fall short of aligning with human preferences. To address this, we present a comprehensive framework that learns an automatic perceptual model from human preference data and leverages it to improve and evaluate the perceptual quality of audio-driven 3D facial animation. To begin with, we construct FMPair (Facial Motion Pairwise preference), the first human preference dataset for audio-driven 3D facial animation, which is built through a systematic annotation pipeline and comprises 65,574 annotated 3D facial motion pairs from 8,834 distinct in-the-wild audio clips. Based on the pairwise comparison dataset, we propose a Facial Motion Reward model, termed FMReward, which takes audio and 3D facial motion as inputs and predicts a perceptual quality score aligned with human preferences. Building upon FMReward, we further introduce Facial Motion reward Feedback Learning (FMFL), a direct fine-tuning algorithm that leverages a pretrained reward model to optimize diffusion-based audio-driven 3D facial animation models for better alignment with human preferences. Extensive experiments demonstrate the superiority of FMReward over other metrics in aligning with human preferences and the effectiveness of FMFL in improving the perceptual quality of audio-driven 3D facial animation.
Ivo Verhoeven, Pushkar Mishra, Ekaterina Shutovacs.LG cs.CL
This paper studies what discriminatively trained reward models (RMs) memorize by measuring counterfactual memorization on two human preference datasets. We show that RMs 1) misallocate memorization to easy, high margin preference pairs, 2) memorize dataset-specific shortcuts (e.g., model identity, user sampling strategy), and 3) overgeneralize simple heuristic correlates of human preference (e.g., length, compliance) when confronted with unseen preference pairs. Overall, our findings indicate that discriminative training of RMs from human preference data results in biased RMs not yet capable of judging response quality in context-dependent scenarios.
We introduce Rushes, a dataset and benchmark for studying revealed human engagement preferences in interactive narrative environments. Rushes is collected through a game interface where users interact with AI-generated branching narratives and select one choice from a small, explicit candidate set at each decision point. Each interaction logs the full candidate set, the user's choice, and the evolving narrative context, yielding time-ordered trajectories with persistent user-level identifiers. Rushes contains 44,226 decision events from 8,167 unique users across six games, capturing sequential, personalized engagement behavior rather than static judgments. We show that user choices exhibit structured, non-random patterns, quantified by a low choice entropy relative to a uniform baseline. We position Rushes as a diagnostic benchmark for pluralistic alignment and demonstrate a robust Engagement Gap: state-of-the-art LLMs, including GPT-5, fail to outperform simple baselines. While classical Matrix Factorization (SVD) captures measurable personalized signal (37.7%), frontier LLMs (34.23%) struggle to even match the Popularity Baseline (36.4%) on event-level choice prediction. This gap suggests that single, population-level objectives, like those used in modern RLHF, appear insufficient to capture heterogeneous, context-dependent engagement signals. As a result, even highly capable models default to majority preferences rather than adapting to individual trajectories. We release Rushes to support research into pluralistic alignment and sequential decision-making in generative systems. The full code for the platform and dataset will be available here: https://github.com/microsoft/rushes
Given the increased adoption of Vision Language Models (VLMs) in human-interactive settings, it is important that we evaluate how well these models can adapt to real-time preferences for different users. While an increasing number of vision-language benchmarks have recently been introduced, they focus largely on evaluating static capabilities and generally-held preferences learned from extensive training data. This work introduces a new benchmark for evaluating the ability of VLMs to understand dynamic human-preferences, i.e. preferences that are passed in-context at inference time. We provide an automated pipeline for generating this benchmark with variations on image dependence, a dynamic multi-modal human-preference dataset, and evaluations of state-of-the-art models on the novel benchmark.
Story generation aims to automatically produce coherent, structured, and engaging narratives. Although large language models (LLMs) have significantly advanced text generation, stories generated by LLMs still diverge from human-authored works regarding complex narrative structure and human-aligned preferences. A key reason is the absence of effective modeling of human story preferences, which are inherently subjective and under-explored. In this work, we systematically evaluate the modeling of human story preferences and introduce StoryRMB, the first benchmark for assessing reward models on story preferences. StoryRMB contains $1,133$ high-quality, human-verified instances, each consisting of a prompt, one chosen story, and three rejected stories. We find existing reward models struggle to select human-preferred stories, with the best model achieving only $66.3\%$ accuracy. To address this limitation, we construct roughly $100,000$ high-quality story preference pairs across diverse domains and develop StoryReward, an advanced reward model for story preference trained on this dataset. StoryReward achieves state-of-the-art (SoTA) performance on StoryRMB, outperforming much larger models. We also adopt StoryReward in downstream test-time scaling applications for best-of-n (BoN) story selection and find that it generally chooses stories better aligned with human preferences. We will release our dataset, model, and code to facilitate future research. Related code and data are available at https://github.com/THU-KEG/StoryReward.
Reinforcement learning (RL) systems typically optimize scalar reward functions that assume precise and reliable evaluation of outcomes. However, real-world objectives--especially those derived from human preferences--are often uncertain, context-dependent, and internally inconsistent. This mismatch can lead to alignment failures such as reward hacking, over-optimization, and overconfident behavior. We introduce a dual-source uncertainty-aware reward framework that explicitly models both epistemic uncertainty in value estimation and uncertainty in human preferences. Model uncertainty is captured via ensemble disagreement over value predictions, while preference uncertainty is derived from variability in reward annotations. We combine these signals through a confidence-adjusted Reliability Filter that adaptively modulates action selection, encouraging a balance between exploitation and caution. Empirical results across multiple discrete grid configurations (6x6, 8x8, 10x10) and high-dimensional continuous control environments (Hopper-v4, Walker2d-v4) demonstrate that our approach yields more stable training dynamics and reduces exploitative behaviors under reward ambiguity, achieving a 93.7% reduction in reward-hacking behavior as measured by trap visitation frequency. We demonstrate statistical significance of these improvements and robustness under up to 30% supervisory noise, albeit with a trade-off in peak observed reward compared to unconstrained baselines. By treating uncertainty as a first-class component of the reward signal, this work offers a principled approach toward more reliable and aligned reinforcement learning systems.