Reinforcement learning from human feedback
In machine learning, reinforcement learning from human feedback is a technique to align an intelligent agent with human preferences. It involves training a reward model to represent preferences, which can then be used to train other models through reinforcement learning.
In classical reinforcement learning, an intelligent agent's goal is to learn a function that guides its behavior, called a policy. The function is iteratively optimized to increase the reward signal derived from the agent's task performance. However, explicitly defining a reward function that accurately approximates human preferences is challenging. Therefore, RLHF seeks to train a "reward model" directly from human feedback. The reward model is first trained in a supervised manner to predict if a response to a given prompt is good or bad based on ranking data collected from human annotators. This model then serves as a reward function to improve an agent's policy through an optimization algorithm like proximal policy optimization.
RLHF has applications in various domains in machine learning, including natural language processing tasks such as text summarization and conversational agents, computer vision tasks like text-to-image models, and the development of video game bots. While RLHF is an effective method of training models to act better in accordance with human preferences, it also faces challenges due to the way the human preference data is collected. Though RLHF does not require massive amounts of data to improve performance, sourcing high-quality preference data is still an expensive process. Furthermore, if the data is not carefully collected from a representative sample, the resulting model may exhibit unwanted biases.
Background and motivation
Optimizing a model based on human feedback is desirable when a task is difficult to specify yet easy to judge. For example, one may want to train a model to generate safe text that is both helpful and harmless. Asking humans to manually create examples of harmless and harmful text would be difficult and time-consuming. However, humans are adept at swiftly assessing and comparing the harmfulness of different AI-generated text. Therefore, a more practical objective would be to allow the model to use this type of human feedback to improve its text generation.Despite the clear benefits of incorporating human feedback in training models, prior efforts—including some that leverage reinforcement learning —have encountered significant challenges. Most attempts were either narrow and difficult to generalize, breaking down on more complex tasks, or they faced difficulties learning from sparse or noisy reward functions.
RLHF was not the first successful method of using human feedback for reinforcement learning, but it is one of the most widely used. The foundation for RLHF was introduced as an attempt to create a general algorithm for learning from a practical amount of human feedback. The algorithm as used today was introduced by OpenAI in a paper on enhancing text continuation or summarization based on human feedback, and it began to gain popularity when the same method was reused in their paper on InstructGPT. RLHF has also been shown to improve the robustness of RL agents and their capacity for exploration, which results in an optimization process more adept at handling uncertainty and efficiently exploring its environment in search of the highest reward.
Collecting human feedback
Human feedback is commonly collected by prompting humans to rank instances of the agent's behavior. These rankings can then be used to score outputs, for example, using the Elo rating system, which is an algorithm for calculating the relative skill levels of players in a game based only on the outcome of each game. While ranking outputs is the most widely adopted form of feedback, recent research has explored other forms, such as numerical feedback, natural language feedback, and prompting for direct edits to the model's output.One initial motivation of RLHF was that it requires relatively small amounts of comparison data to be effective. It has been shown that a small amount of data can lead to comparable results to a larger amount. In addition, increasing the amount of data tends to be less effective than proportionally increasing the size of the reward model. Nevertheless, a larger and more diverse amount of data can be crucial for tasks where it is important to avoid bias from a partially representative group of annotators.
When learning from human feedback through pairwise comparison under the Bradley–Terry–Luce model, the maximum likelihood estimator for linear reward functions has been shown to converge if the comparison data is generated under a well-specified linear model. This implies that, under certain conditions, if a model is trained to decide which choices people would prefer between pairs of choices, it will necessarily improve at predicting future preferences. This improvement is expected as long as the comparisons it learns from are based on a consistent and simple rule.
Both offline data collection models, where the model is learning by interacting with a static dataset and updating its policy in batches, as well as online data collection models, where the model directly interacts with the dynamic environment and updates its policy immediately, have been mathematically studied proving sample complexity bounds for RLHF under different feedback models.
In the offline data collection model, when the objective is policy training, a pessimistic MLE that incorporates a lower confidence bound as the reward estimate is most effective. Moreover, when applicable, it has been shown that considering K-wise comparisons directly is asymptotically more efficient than converting them into pairwise comparisons for prediction purposes.
In the online scenario, when human feedback is collected through pairwise comparisons under the Bradley–Terry–Luce model and the objective is to minimize the algorithm's regret, it has been shown that an optimistic MLE that incorporates an upper confidence bound as the reward estimate can be used to design sample efficient algorithms. A key challenge in RLHF when learning from pairwise comparisons is associated with the non-Markovian nature of its optimal policies. Unlike simpler scenarios where the optimal strategy does not require memory of past actions, in RLHF, the best course of action often depends on previous events and decisions, making the strategy inherently memory-dependent.
Applications
RLHF has been applied to various domains of natural language processing, such as conversational agents, text summarization, and natural language understanding. Ordinary reinforcement learning, in which agents learn from their actions based on a predefined "reward function", is difficult to apply to NLP tasks because the rewards tend to be difficult to define or measure, especially when dealing with complex tasks that involve human values or preferences. RLHF can steer NLP models, in particular language models, to provide answers that align with human preferences with regard to such tasks by capturing their preferences beforehand in the reward model. This results in a model capable of generating more relevant responses and rejecting inappropriate or irrelevant queries. Some notable examples of RLHF-trained language models are OpenAI's ChatGPT, DeepMind's Sparrow, Google's Gemini, and Anthropic's Claude.In computer vision, RLHF has also been used to align text-to-image models. Studies that successfully used RLHF for this goal have noted that the use of KL regularization in RLHF, which aims to prevent the learned policy from straying too far from the unaligned model, helped to stabilize the training process by reducing overfitting to the reward model. The final image outputs from models trained with KL regularization were noted to be of significantly higher quality than those trained without. Other methods tried to incorporate the feedback through more direct training—based on maximizing the reward without the use of reinforcement learning—but conceded that an RLHF-based approach would likely perform better due to the online sample generation used in RLHF during updates as well as the aforementioned KL regularization over the prior model, which mitigates overfitting to the reward function.
RLHF was initially applied to other areas, such as the development of video game bots and tasks in simulated robotics. For example, OpenAI and DeepMind trained agents to play Atari games based on human preferences. In classical RL-based training of such bots, the reward function is simply correlated to how well the agent is performing in the game, usually using metrics like the in-game score. In comparison, in RLHF, a human is periodically presented with two clips of the agent's behavior in the game and must decide which one looks better. This approach can teach agents to perform at a competitive level without ever having access to their score. In fact, it was shown that RLHF can sometimes lead to superior performance over RL with score metrics because the human's preferences can contain more useful information than performance-based metrics. The agents achieved strong performance in many of the environments tested, often surpassing human performance.
Training
In RLHF, two different models are trained: a reward model and a reinforcement learning policy. The reward model learns to determine what behavior is desirable based on human feedback, while the policy is guided by the reward model to determine the agent's actions. Both models are commonly initialized using a pre-trained autoregressive language model. This model is then customarily trained in a supervised manner on a relatively small dataset of pairs of prompts to an assistant and their accompanying responses, written by human annotators.Reward model
The reward model is a function that takes a string as input, and produces a single number, which is the "reward".It is usually initialized with a pre-trained model, as this initializes it with an understanding of language and focuses training explicitly on learning human preferences. In addition to being used to initialize the reward model and the RL policy, the model is then also used to sample data to be compared by annotators.
The reward model is then trained by replacing the final layer of the previous model with a randomly initialized regression head. This change shifts the model from its original classification task over its vocabulary to simply outputting a number corresponding to the score of any given prompt and response. This model is trained on the human preference comparison data collected earlier from the supervised model. In particular, it is trained to minimize the following cross-entropy loss function:
where is the number of responses the labelers ranked, is the output of the reward model for prompt and completion, is the preferred completion over, denotes the sigmoid function, and denotes the expected value. This can be thought of as a form of logistic regression, where the model predicts the probability that a response is preferred over.
This loss function essentially measures the difference between the reward model's predictions and the decisions made by humans. The goal is to make the model's guesses as close as possible to the humans' preferences by minimizing the difference measured by this equation. In the case of only pairwise comparisons,, so the factor of. In general, all comparisons from each prompt are used for training as a single batch.
After training, the outputs of the model are normalized such that the reference completions have a mean score of 0. That is, for each query and reference pair by calculating the mean reward across the training dataset and setting it as the bias in the reward head.