Aligning large language models (LLMs) with human preferences is critical to recent advances in generative artificial intelligence. Reinforcement learning from human feedback (RLHF) is widely applied to achieve this objective. A key step in RLHF is to learn the reward function from human feedback. However, human feedback is costly and time-consuming, making it essential to collect high-quality conversation data for human teachers to label. Additionally, different human teachers have different levels of expertise. It is thus critical to query the most appropriate teacher for their opinions. In this paper, we use offline reinforcement learning (RL) to formulate the alignment problem. Motivated by the idea of $D$-optimal design, we first propose a dual active reward learning algorithm for the simultaneous selection of conversations and teachers. Next, we apply pessimistic RL to solve the alignment problem, based on the learned reward estimator. Theoretically, we show that the reward estimator obtained through our proposed adaptive selection strategy achieves minimal generalized variance asymptotically, and prove that the sub-optimality of our pessimistic policy scales as $O(1/\sqrt{T})$ with a given sample budget $T$. Through simulations and experiments on LLMs, we demonstrate the effectiveness of our algorithm and its superiority over state-of-the-arts.

We introduce Teacher2Task, a novel framework for multi-teacher learning that eliminates the need for manual aggregation heuristics. Existing multi-teacher methods typically rely on such heuristics to combine predictions from multiple teachers, often resulting in sub-optimal aggregated labels and the propagation of aggregation errors. Teacher2Task addresses these limitations by introducing teacher-specific input tokens and reformulating the training process. Instead of relying on aggregated labels, the framework transforms the training data, consisting of ground truth labels and annotations from N teachers, into N+1 distinct tasks: N auxiliary tasks that predict the labeling styles of the N individual teachers, and one primary task that focuses on the ground truth labels. This approach, drawing upon principles from multiple learning paradigms, demonstrates strong empirical results across a range of architectures, modalities, and tasks.

Specifying objective functions for machine learning systems is challenging, and misspecified objectives can be hacked [1, 2] or incentivise degenerate behavior [3, 4, 5]. Techniques such as reinforcement learning from human feedback (RLHF) enable ML systems to instead learn appropriate objectives from human feedback [6, 7, 8]. These techniques are widely used to finetune large language models [9, 10, 11, 12] and to train reinforcement learning agents to perform complex maneuvers in continuous control environments [6, 7]. However, while RLHF is relied upon to ensure that these systems are safe, helpful, and harmless [13], it still faces many limitations and unsolved challenges [14]. In particular, RLHF systems typically rely on the assumption that all feedback comes from a single human teacher, despite gathering feedback from a range of teachers with varying levels of rationality and expertise. For example, Stiennon et al. [8], Bai et al. [13] and Ouyang et al. [15] assume that all feedback comes from a single teacher, but find that annotators and researchers actually disagree 23% to 37% of the time. Reward learning has been shown to be highly sensitive to incorrect assumptions about the process that generates feedback [16, 17, 18, 19], so this single-teacher assumption exposes these systems to dangerous failures [20]. Ideally, RLHF systems should consider the differences between each teacher to improve their safety and reliability. To leverage multiple teachers in RLHF, we introduce a novel problem called a Hidden Utility Bandit (HUB), illustrated in Figure 1.

Reward learning algorithms utilize human feedback to infer a reward function, which is then used to train an AI system. This human feedback is often a preference comparison, in which the human teacher compares several samples of AI behavior and chooses which they believe best accomplishes the objective. While reward learning typically assumes that all feedback comes from a single teacher, in practice these systems often query multiple teachers to gather sufficient training data. In this paper, we investigate this disparity, and find that algorithmic evaluation of these different sources of feedback facilitates more accurate and efficient reward learning. We formally analyze the value of information (VOI) when reward learning from teachers with varying levels of rationality, and define and evaluate an algorithm that utilizes this VOI to actively select teachers to query for feedback. Surprisingly, we find that it is often more informative to query comparatively irrational teachers. By formalizing this problem and deriving an analytical solution, we hope to facilitate improvement in reward learning approaches to aligning AI behavior with human values.

Reinforcement learning from human feedback (RLHF) is a powerful technique for training agents to perform difficult-to-specify tasks. However, human feedback can be noisy, particularly when human teachers lack relevant knowledge or experience. Levels of expertise vary across teachers, and a given teacher may have differing levels of expertise for different components of a task. RLHF algorithms that learn from multiple teachers therefore face an expertise problem: the reliability of a given piece of feedback depends both on the teacher that it comes from and how specialized that teacher is on relevant components of the task. Existing state-of-the-art RLHF algorithms assume that all evaluations come from the same distribution, obscuring this inter- and intra-human variance, and preventing them from accounting for or taking advantage of variations in expertise. We formalize this problem, implement it as an extension of an existing RLHF benchmark, evaluate the performance of a state-of-the-art RLHF algorithm, and explore techniques to improve query and teacher selection. Our key contribution is to demonstrate and characterize the expertise problem, and to provide an open-source implementation for testing future solutions.