After the bidirectional models and rollout policies are well trained, we utilize them to generate imaginary trajectories, while conducting double check and admitting high-confidence transitions simultaneously.

We take the novel perspective of incorporating offline RL algorithms as subroutines of tabula rasa online RL. This is feasible because an online learning agent can repurpose its historical interactions as offline dataset. We formalize this idea into a framework that accommodates several variants of offline RL incorporation such as final policy recommendation and online fine-tuning. We further introduce convenient techniques to improve its effectiveness in enhancing online learning efficiency. Our extensive and systematic empirical analyses show that 1) the effectiveness of the proposed framework depends strongly on the nature of the task, 2) our proposed techniques greatly enhance its effectiveness, and 3) existing online fine-tuning methods are overall ineffective, calling for more research therein.

The field of Offline Reinforcement Learning (RL) aims to derive effective policies from pre-collected datasets without active environment interaction. While traditional offline RL algorithms like Conservative Q-Learning (CQL) and Implicit Q-Learning (IQL) have shown promise, they often face challenges in balancing exploration and exploitation, especially in environments with varying reward densities. The recently proposed Decision Transformer (DT) approach, which reframes offline RL as a sequence modelling problem, has demonstrated impressive results across various benchmarks. This paper presents a comparative study evaluating the performance of DT against traditional offline RL algorithms in dense and sparse reward settings for the ANT con-tinous control environment. Our research investigates how these algorithms perform when faced with different reward structures, examining their ability to learn effective policies and generalize across varying levels of feedback. Through empirical analysis in the ANT environment, we found that DTs showed less sensitivity to varying reward density compared to other methods and particularly excelled with medium-expert datasets in sparse reward scenarios. In contrast, traditional value-based methods like IQL showed improved performance in dense reward settings with high-quality data, while CQL offered balanced performance across different data qualities. Additionally, DTs exhibited lower variance in performance but required significantly more computational resources compared to traditional approaches. These findings suggest that sequence modelling approaches may be more suitable for scenarios with uncertain reward structures or mixed-quality data, while value-based methods remain competitive in settings with dense rewards and high-quality demonstrations.

Reinforcement learning (RL) with limited samples is common in real-world applications. However, offline RL performance under this constraint is often suboptimal. We consider an alternative approach to dealing with limited samples by introducing the Quantum Metric Encoder (QME). In this methodology, instead of applying the RL framework directly on the original states and rewards, we embed the states into a more compact and meaningful representation, where the structure of the encoding is inspired by quantum circuits. For classical data, QME is a classically simulable, trainable unitary embedding and thus serves as a quantum-inspired module, on a classical device. For quantum data in the form of quantum states, QME can be implemented directly on quantum hardware, allowing for training without measurement or re-encoding. We evaluated QME on three datasets, each limited to 100 samples. We use Soft-Actor-Critic (SAC) and Implicit-Q-Learning (IQL), two well-known RL algorithms, to demonstrate the effectiveness of our approach. From the experimental results, we find that training offline RL agents on QME-embedded states with decoded rewards yields significantly better performance than training on the original states and rewards. On average across the three datasets, for maximum reward performance, we achieve a 116.2% improvement for SAC and 117.6% for IQL. We further investigate the $Δ$-hyperbolicity of our framework, a geometric property of the state space known to be important for the RL training efficacy. The QME-embedded states exhibit low $Δ$-hyperbolicity, suggesting that the improvement after embedding arises from the modified geometry of the state space induced by QME. Thus, the low $Δ$-hyperbolicity and the corresponding effectiveness of QME could provide valuable information for developing efficient offline RL methods under limited-sample conditions.

Q and P represent the number of gradient descent iterations and the number of MCMC samplings, respectively. For the other symbols, see the main text.

Visualizations of the real and learned state-values of IA VI, IQL and DIQL can be found in Figure 7.Figure 7: Visualization of state-values for different numbers of trajectories in Objectworld. Table 2: Comparison between online and offline estimation of state-action visitations for the Ob-jectworld environment, given a data set with an action distribution equivalent to the true optimal Boltzmann distribution. The pseudocode of the tabular variant of Constrained Inverse Q-learning can be found in Algorithm 4. See [4] for further details of Constrained Q-learning.Algorithm 4: Tabular Model-free Constrained Inverse Q-learning The pseudocode of Deep Constrained Inverse Q-learning can be found in Algorithm 5. The lower row shows the EVD. 3 For DIQL, the parameters were optimized in the range of Hence, it can only increase.

Reward Generation via Large Vision-Language Model in Offline Reinforcement Learning Y ounghwan Lee Electrical Engineering KAIST Daejeon, South Korea youngh2@kaist.ac.kr Chang D. Y oo Electrical Engineering KAIST Daejeon, South Korea cd yoo@kaist.ac.kr Abstract --In offline reinforcement learning (RL), learning from fixed datasets presents a promising solution for domains where real-time interaction with the environment is expensive or risky. However, designing dense reward signals for offline dataset requires significant human effort and domain expertise. Reinforcement learning with human feedback (RLHF) has emerged as an alternative, but it remains costly due to the human-in-the-loop process, prompting interest in automated reward generation models. T o address this, we propose Reward Generation via Large Vision-Language Models (RG-VLM), which leverages the reasoning capabilities of L VLMs to generate rewards from offline data without human involvement.