Scaling up deep Reinforcement Learning (RL) methods presents a significant challenge. Following developments in generative modelling, model-based RL positions itself as a strong contender. Recent advances in sequence modelling have led to effective transformer-based world models, albeit at the price of heavy computations due to the long sequences of tokens required to accurately simulate environments. In this work, we propose $\Delta$-IRIS, a new agent with a world model architecture composed of a discrete autoencoder that encodes stochastic deltas between time steps and an autoregressive transformer that predicts future deltas by summarizing the current state of the world with continuous tokens. In the Crafter benchmark, $\Delta$-IRIS sets a new state of the art at multiple frame budgets, while being an order of magnitude faster to train than previous attention-based approaches. We release our code and models at https://github.com/vmicheli/delta-iris.

The use of machine learning to investigate grasp affordances has received extensive attention over the past several decades. The existing literature provides a robust basis to build upon, though a number of aspects may be improved. Results commonly work in terms of grasp configuration, with little consideration for the manner in which the grasp may be (re-)produced from a reachability and trajectory planning perspective. In addition, the majority of existing learning approaches focus of producing a single viable grasp, offering little transparency on how the result was reached, or insights on its robustness. We propose a different perspective on grasp affordance learning, explicitly accounting for grasp synthesis; that is, the manner in which manipulator kinematics are used to allow materialization of grasps. The approach allows to explicitly map the grasp policy space in terms of generated grasp types and associated grasp quality. Results of numerical simulations illustrate merit of the method and highlight the manner in which it may promote a greater degree of explainability for otherwise intransparent reinforcement processes.

Recent advances in prompt optimization have notably enhanced the performance of pre-trained language models (PLMs) on downstream tasks. However, the potential of optimized prompts on domain generalization has been under-explored. To explore the nature of prompt generalization on unknown domains, we conduct pilot experiments and find that (i) Prompts gaining more attention weight from PLMs' deep layers are more generalizable and (ii) Prompts with more stable attention distributions in PLMs' deep layers are more generalizable. Thus, we offer a fresh objective towards domain-generalizable prompts optimization named "Concentration", which represents the "lookback" attention from the current decoding token to the prompt tokens, to increase the attention strength on prompts and reduce the fluctuation of attention distribution. We adapt this new objective to popular soft prompt and hard prompt optimization methods, respectively. Extensive experiments demonstrate that our idea improves comparison prompt optimization methods by 1.42% for soft prompt generalization and 2.16% for hard prompt generalization in accuracy on the multi-source domain generalization setting, while maintaining satisfying in-domain performance. The promising results validate the effectiveness of our proposed prompt optimization objective and provide key insights into domain-generalizable prompts.

Auto-bidding plays a crucial role in facilitating online advertising by automatically providing bids for advertisers. Reinforcement learning (RL) has gained popularity for auto-bidding. However, most current RL auto-bidding methods are modeled through the Markovian Decision Process (MDP), which assumes the Markovian state transition. This assumption restricts the ability to perform in long horizon scenarios and makes the model unstable when dealing with highly random online advertising environments. To tackle this issue, this paper introduces AI-Generated Bidding (AIGB), a novel paradigm for auto-bidding through generative modeling. In this paradigm, we propose DiffBid, a conditional diffusion modeling approach for bid generation. DiffBid directly models the correlation between the return and the entire trajectory, effectively avoiding error propagation across time steps in long horizons. Additionally, DiffBid offers a versatile approach for generating trajectories that maximize given targets while adhering to specific constraints. Extensive experiments conducted on the real-world dataset and online A/B test on Alibaba advertising platform demonstrate the effectiveness of DiffBid, achieving 2.81% increase in GMV and 3.36% increase in ROI.

Off-policy evaluation (OPE) is crucial for evaluating a target policy's impact offline before its deployment. However, achieving accurate OPE in large state spaces remains challenging. This paper studies state abstractions - originally designed for policy learning - in the context of OPE. Our contributions are three-fold: (i) We define a set of irrelevance conditions central to learning state abstractions for OPE.

Model-free control strategies such as reinforcement learning have shown the ability to learn control strategies without requiring an accurate model or simulator of the world. While this is appealing due to the lack of modeling requirements, such methods can be sample inefficient, making them impractical in many real-world domains. On the other hand, model-based control techniques leveraging accurate simulators can circumvent these challenges and use a large amount of cheap simulation data to learn controllers that can effectively transfer to the real world. The challenge with such model-based techniques is the requirement for an extremely accurate simulation, requiring both the specification of appropriate simulation assets and physical parameters. This requires considerable human effort to design for every environment being considered. In this work, we propose a learning system that can leverage a small amount of real-world data to autonomously refine a simulation model and then plan an accurate control strategy that can be deployed in the real world. Our approach critically relies on utilizing an initial (possibly inaccurate) simulator to design effective exploration policies that, when deployed in the real world, collect high-quality data. We demonstrate the efficacy of this paradigm in identifying articulation, mass, and other physical parameters in several challenging robotic manipulation tasks, and illustrate that only a small amount of real-world data can allow for effective sim-to-real transfer. Project website at https://weirdlabuw.github.io/asid

Alphas are pivotal in providing signals for quantitative trading. The industry highly values the discovery of formulaic alphas for their interpretability and ease of analysis, compared with the expressive yet overfitting-prone black-box alphas. In this work, we focus on discovering formulaic alphas. Prior studies on automatically generating a collection of formulaic alphas were mostly based on genetic programming (GP), which is known to suffer from the problems of being sensitive to the initial population, converting to local optima, and slow computation speed. Recent efforts employing deep reinforcement learning (DRL) for alpha discovery have not fully addressed key practical considerations such as alpha correlations and validity, which are crucial for their effectiveness. In this work, we propose a novel framework for alpha discovery using DRL by formulating the alpha discovery process as program construction. Our agent, $\text{Alpha}^2$, assembles an alpha program optimized for an evaluation metric. A search algorithm guided by DRL navigates through the search space based on value estimates for potential alpha outcomes. The evaluation metric encourages both the performance and the diversity of alphas for a better final trading strategy. Our formulation of searching alphas also brings the advantage of pre-calculation dimensional analysis, ensuring the logical soundness of alphas, and pruning the vast search space to a large extent. Empirical experiments on real-world stock markets demonstrates $\text{Alpha}^2$'s capability to identify a diverse set of logical and effective alphas, which significantly improves the performance of the final trading strategy. The code of our method is available at https://github.com/x35f/alpha2.

Controlling humanoids in complex physically simulated worlds is a long-standing challenge with numerous applications in gaming, simulation, and visual content creation. In our setup, given a rich and complex 3D scene, the user provides a list of instructions composed of target locations and locomotion types. To solve this task we present PlaMo, a scene-aware path planner and a robust physics-based controller. The path planner produces a sequence of motion paths, considering the various limitations the scene imposes on the motion, such as location, height, and speed. Complementing the planner, our control policy generates rich and realistic physical motion adhering to the plan. We demonstrate how the combination of both modules enables traversing complex landscapes in diverse forms while responding to real-time changes in the environment. Video: https://youtu.be/wWlqSQlRZ9M .

There has been significant progress in deep reinforcement learning (RL) in recent years. Nevertheless, finding suitable hyperparameter configurations and reward functions remains challenging even for experts, and performance heavily relies on these design choices. Also, most RL research is conducted on known benchmarks where knowledge about these choices already exists. However, novel practical applications often pose complex tasks for which no prior knowledge about good hyperparameters and reward functions is available, thus necessitating their derivation from scratch. Prior work has examined automatically tuning either hyperparameters or reward functions individually. We demonstrate empirically that an RL algorithm's hyperparameter configurations and reward function are often mutually dependent, meaning neither can be fully optimised without appropriate values for the other. We then propose a methodology for the combined optimisation of hyperparameters and the reward function. Furthermore, we include a variance penalty as an optimisation objective to improve the stability of learned policies. We conducted extensive experiments using Proximal Policy Optimisation and Soft Actor-Critic on four environments. Our results show that combined optimisation significantly improves over baseline performance in half of the environments and achieves competitive performance in the others, with only a minor increase in computational costs. This suggests that combined optimisation should be best practice.

Hierarchical reinforcement learning (HRL) addresses complex long-horizon tasks by skillfully decomposing them into subgoals. Therefore, the effectiveness of HRL is greatly influenced by subgoal reachability. Typical HRL methods only consider subgoal reachability from the unilateral level, where a dominant level enforces compliance to the subordinate level. However, we observe that when the dominant level becomes trapped in local exploration or generates unattainable subgoals, the subordinate level is negatively affected and cannot follow the dominant level's actions. This can potentially make both levels stuck in local optima, ultimately hindering subsequent subgoal reachability. Allowing real-time bilateral information sharing and error correction would be a natural cure for this issue, which motivates us to propose a mutual response mechanism. Based on this, we propose the Bidirectional-reachable Hierarchical Policy Optimization (BrHPO)--a simple yet effective algorithm that also enjoys computation efficiency. Experiment results on a variety of long-horizon tasks showcase that BrHPO outperforms other state-of-the-art HRL baselines, coupled with a significantly higher exploration efficiency and robustness.