SGD performs worse than Adam by a significant margin on Transformers, but the reason remains unclear. In this work, we provide an explanation of SGD's failure on Transformers through the lens of Hessian: (i) Transformers are ``heterogeneous'': the Hessian spectrum across parameter blocks vary dramatically, a phenomenon we call ``block heterogeneity"; (ii) Heterogeneity hampers SGD: SGD performs badly on problems with block heterogeneity. To validate that heterogeneity hampers SGD, we check various Transformers, CNNs, MLPs, and quadratic problems, and find that SGD works well on problems without block heterogeneity but performs badly when the heterogeneity exists. Our initial theoretical analysis indicates that SGD fails because it applies one single learning rate for all blocks, which cannot handle the heterogeneity among blocks. The failure could be rescued if we could assign different learning rates across blocks, as designed in Adam.
Aligning intelligent agents with human preferences and values is important. This paper examines two popular alignment methods: Direct Preference Optimization (DPO) and Reward-Model-Based Policy Optimization (RMB-PO). A variant of RMB-PO, referred to as RMB-PO+ is also considered. These methods, either explicitly or implicitly, learn a reward model from preference data and differ in the data used for policy optimization to unlock the generalization ability of the reward model. In particular, compared with DPO, RMB-PO additionally uses policy-generated data, and RMB-PO+ further leverages new, preference-free data. We examine the impact of such out-of-preference data. Our study, conducted through controlled and synthetic experiments, demonstrates that DPO performs poorly, whereas RMB-PO+ performs the best. In particular, even when providing the policy model with a good feature representation, we find that policy optimization with adequate out-of-preference data significantly improves performance by harnessing the reward model's generalization capabilities.
Alignment is of critical importance for training large language models (LLMs). The predominant strategy to address this is through Reinforcement Learning from Human Feedback (RLHF), where PPO serves as the de-facto algorithm. Yet, PPO is known to suffer from computational inefficiency, which is a challenge that this paper aims to address. We identify three important properties in RLHF tasks: fast simulation, deterministic transitions, and trajectory-level rewards, which are not leveraged in PPO. Based on such observations, we develop a new algorithm tailored for RLHF, called ReMax. The algorithm design of ReMax is built on a celebrated algorithm REINFORCE but is equipped with a new variance-reduction technique. Our method has three-fold advantages over PPO: first, ReMax is simple to implement and removes many hyper-parameters in PPO, which are scale-sensitive and laborious to tune. Second, ReMax saves about 50% memory usage in principle. As a result, PPO runs out-of-memory when fine-tuning a Llama2 (7B) model on 8xA100-40GB GPUs, whereas ReMax can afford training. This memory improvement is achieved by removing the value model in PPO. Third, based on our calculations, we find that even assuming PPO can afford the training of Llama2 (7B), it would still run about 2x slower than ReMax. This is due to the computational overhead of the value model, which does not exist in ReMax. Importantly, the above computational improvements do not sacrifice the performance. We hypothesize these advantages can be maintained in larger-scaled models. Our implementation of ReMax is available at https://github.com/liziniu/ReMax
Imitation learning (IL) has proven to be an effective method for learning good policies from expert demonstrations. Adversarial imitation learning (AIL), a subset of IL methods, is particularly promising, but its theoretical foundation in the presence of unknown transitions has yet to be fully developed. This paper explores the theoretical underpinnings of AIL in this context, where the stochastic and uncertain nature of environment transitions presents a challenge. We examine the expert sample complexity and interaction complexity required to recover good policies. To this end, we establish a framework connecting reward-free exploration and AIL, and propose an algorithm, MB-TAIL, that achieves the minimax optimal expert sample complexity of $\widetilde{O} (H^{3/2} |S|/\varepsilon)$ and interaction complexity of $\widetilde{O} (H^{3} |S|^2 |A|/\varepsilon^2)$. Here, $H$ represents the planning horizon, $|S|$ is the state space size, $|A|$ is the action space size, and $\varepsilon$ is the desired imitation gap. MB-TAIL is the first algorithm to achieve this level of expert sample complexity in the unknown transition setting and improves upon the interaction complexity of the best-known algorithm, OAL, by $O(H)$. Additionally, we demonstrate the generalization ability of MB-TAIL by extending it to the function approximation setting and proving that it can achieve expert sample and interaction complexity independent of $|S|$
Reinforcement learning (RL) has shown promise for decision-making tasks in real-world applications. One practical framework involves training parameterized policy models from an offline dataset and subsequently deploying them in an online environment. However, this approach can be risky since the offline training may not be perfect, leading to poor performance of the RL models that may take dangerous actions. To address this issue, we propose an alternative framework that involves a human supervising the RL models and providing additional feedback in the online deployment phase. We formalize this online deployment problem and develop two approaches. The first approach uses model selection and the upper confidence bound algorithm to adaptively select a model to deploy from a candidate set of trained offline RL models. The second approach involves fine-tuning the model in the online deployment phase when a supervision signal arrives. We demonstrate the effectiveness of these approaches for robot locomotion control and traffic light control tasks through empirical validation.
Behavioral cloning (BC) can recover a good policy from abundant expert data, but may fail when expert data is insufficient. This paper considers a situation where, besides the small amount of expert data, a supplementary dataset is available, which can be collected cheaply from sub-optimal policies. Imitation learning with a supplementary dataset is an emergent practical framework, but its theoretical foundation remains under-developed. To advance understanding, we first investigate a direct extension of BC, called NBCU, that learns from the union of all available data. Our analysis shows that, although NBCU suffers an imitation gap that is larger than BC in the worst case, there exist special cases where NBCU performs better than or equally well as BC. This discovery implies that noisy data can also be helpful if utilized elaborately. Therefore, we further introduce a discriminator-based importance sampling technique to re-weight the supplementary data, proposing the WBCU method. With our newly developed landscape-based analysis, we prove that WBCU can outperform BC in mild conditions. Empirical studies show that WBCU simultaneously achieves the best performance on two challenging tasks where prior state-of-the-art methods fail.
Imitation learning learns a policy from expert trajectories. While the expert data is believed to be crucial for imitation quality, it was found that a kind of imitation learning approach, adversarial imitation learning (AIL), can have exceptional performance. With as little as only one expert trajectory, AIL can match the expert performance even in a long horizon, on tasks such as locomotion control. There are two mysterious points in this phenomenon. First, why can AIL perform well with only a few expert trajectories? Second, why does AIL maintain good performance despite the length of the planning horizon? In this paper, we theoretically explore these two questions. For a total-variation-distance-based AIL (called TV-AIL), our analysis shows a horizon-free imitation gap $\mathcal O(\{\min\{1, \sqrt{|\mathcal S|/N} \})$ on a class of instances abstracted from locomotion control tasks. Here $|\mathcal S|$ is the state space size for a tabular Markov decision process, and $N$ is the number of expert trajectories. We emphasize two important features of our bound. First, this bound is meaningful in both small and large sample regimes. Second, this bound suggests that the imitation gap of TV-AIL is at most 1 regardless of the planning horizon. Therefore, this bound can explain the empirical observation. Technically, we leverage the structure of multi-stage policy optimization in TV-AIL and present a new stage-coupled analysis via dynamic programming
Q-learning with function approximation could diverge in the off-policy setting and the target network is a powerful technique to address this issue. In this manuscript, we examine the sample complexity of the associated target Q-learning algorithm in the tabular case with a generative oracle. We point out a misleading claim in [Lee and He, 2020] and establish a tight analysis. In particular, we demonstrate that the sample complexity of the target Q-learning algorithm in [Lee and He, 2020] is $\widetilde{\mathcal O}(|\mathcal S|^2|\mathcal A|^2 (1-\gamma)^{-5}\varepsilon^{-2})$. Furthermore, we show that this sample complexity is improved to $\widetilde{\mathcal O}(|\mathcal S||\mathcal A| (1-\gamma)^{-5}\varepsilon^{-2})$ if we can sequentially update all state-action pairs and $\widetilde{\mathcal O}(|\mathcal S||\mathcal A| (1-\gamma)^{-4}\varepsilon^{-2})$ if $\gamma$ is further in $(1/2, 1)$. Compared with the vanilla Q-learning, our results conclude that the introduction of a periodically-frozen target Q-function does not sacrifice the sample complexity.
Since the introduction of GAIL, adversarial imitation learning (AIL) methods attract lots of research interests. Among these methods, ValueDice has achieved significant improvements: it beats the classical approach Behavioral Cloning (BC) under the offline setting, and it requires fewer interactions than GAIL under the online setting. Are these improvements benefited from more advanced algorithm designs? We answer this question with the following conclusions. First, we show that ValueDice could reduce to BC under the offline setting. Second, we verify that overfitting exists and regularization matters. Specifically, we demonstrate that with weight decay, BC also nearly matches the expert performance as ValueDice does. The first two claims explain the superior offline performance of ValueDice. Third, we establish that ValueDice does not work at all when the expert trajectory is subsampled. Instead, the mentioned success holds when the expert trajectory is complete, in which ValueDice is closely related to BC that performs well as mentioned. Finally, we discuss the implications of our research for imitation learning studies beyond ValueDice.
This paper is dedicated to designing provably efficient adversarial imitation learning (AIL) algorithms that directly optimize policies from expert demonstrations. Firstly, we develop a transition-aware AIL algorithm named TAIL with an expert sample complexity of $\tilde{O}(H^{3/2} |S|/\varepsilon)$ under the known transition setting, where $H$ is the planning horizon, $|S|$ is the state space size and $\varepsilon$ is desired policy value gap. This improves upon the previous best bound of $\tilde{O}(H^2 |S| / \varepsilon^2)$ for AIL methods and matches the lower bound of $\tilde{\Omega} (H^{3/2} |S|/\varepsilon)$ in [Rajaraman et al., 2021] up to a logarithmic factor. The key ingredient of TAIL is a fine-grained estimator for expert state-action distribution, which explicitly utilizes the transition function information. Secondly, considering practical settings where the transition functions are usually unknown but environment interaction is allowed, we accordingly develop a model-based transition-aware AIL algorithm named MB-TAIL. In particular, MB-TAIL builds an empirical transition model by interacting with the environment and performs imitation under the recovered empirical model. The interaction complexity of MB-TAIL is $\tilde{O} (H^3 |S|^2 |A| / \varepsilon^2)$, which improves the best known result of $\tilde{O} (H^4 |S|^2 |A| / \varepsilon^2)$ in [Shani et al., 2021]. Finally, our theoretical results are supported by numerical evaluation and detailed analysis on two challenging MDPs.