This work proposes a self-supervised training strategy designed for combinatorial problems. An obstacle in applying supervised paradigms to such problems is the need for costly target solutions often produced with exact solvers. Inspired by semi- and self-supervised learning, we show that generative models can be trained by sampling multiple solutions and using the best one according to the problem objective as a pseudo-label. In this way, we iteratively improve the model generation capability by relying only on its self-supervision, eliminating the need for optimality information. We validate this Self-Labeling Improvement Method (SLIM) on the Job Shop Scheduling (JSP), a complex combinatorial problem that is receiving much attention from the neural combinatorial community.

Monte-Carlo tree search (MCTS) is an influential sequential decision-making algorithm notably employed in AlphaZero. Despite its success, the primary challenge in AlphaZero training lies in its prolonged time-to-solution due to the high latency imposed by the sequential MCTS process. To address this challenge, this paper proposes and evaluates an inter-decision parallelization strategy called speculative MCTS, a new type of parallelism in AlphaZero which implements speculative execution. This approach allows for the parallel execution of future moves before the current MCTS computations are completed, thus reducing the latency. Additionally, we analyze factors contributing to the overall speedup by studying the synergistic effects of speculation and neural network caching in MCTS.

Goal-conditioned policies are generally understood to be "feed-forward" circuits, in the form of neural networks that map from the current state and the goal specification to the next action to take. However, under what circumstances such a policy can be learned and how efficient the policy will be are not well understood. In this paper, we present a circuit complexity analysis for relational neural networks (such as graph neural networks and transformers) representing policies for planning problems, by drawing connections with serialized goal regression search (S-GRS). We show that there are three general classes of planning problems, in terms of the growth of circuit width and depth as a function of the number of objects and planning horizon, providing constructive proofs. We also illustrate the utility of this analysis for designing neural networks for policy learning.

Discovering symbolic representations for skills is essential for abstract reasoning and efficient planning in robotics. Previous neuro-symbolic robotic studies mostly focused on discovering perceptual symbolic categories given a pre-defined action repertoire and generating plans with given action symbols. A truly developmental robotic system, on the other hand, should be able to discover all the abstractions required for the planning system with minimal human intervention. In this study, we propose a novel system that is designed to discover symbolic action primitives along with perceptual symbols autonomously. Our system is based on an encoder-decoder structure that takes object and action information as input and predicts the generated effect. To efficiently explore the vast continuous action parameter space, we introduce a Curiosity-Based exploration module that selects the most informative actions -- the ones that maximize the entropy in the predicted effect distribution. The discovered symbolic action primitives are then used to make plans using a symbolic tree search strategy in single- and double-object manipulation tasks. We compare our model with two baselines that use different exploration strategies in different experiments. The results show that our approach can learn a diverse set of symbolic action primitives, which are effective for generating plans in order to achieve given manipulation goals.

Large Language Models (LLMs) have shown remarkable performance in various natural language tasks, but they often struggle with planning problems that require structured reasoning. To address this limitation, the conversion of planning problems into the Planning Domain Definition Language (PDDL) has been proposed as a potential solution, enabling the use of automated planners. However, generating accurate PDDL files typically demands human inputs or correction, which can be time-consuming and costly. In this paper, we propose a novel approach that leverages LLMs and environment feedback to automatically generate PDDL domain and problem description files without the need for human intervention. Our method introduces an iterative refinement process that generates multiple problem PDDL candidates and progressively refines the domain PDDL based on feedback obtained from interacting with the environment.

Humans are autotelic agents who learn by setting and pursuing their own goals. However, the precise mechanisms guiding human goal selection remain unclear. Learning progress, typically measured as the observed change in performance, can provide a valuable signal for goal selection in both humans and artificial agents. We hypothesize that human choices of goals may also be driven by latent learning progress, which humans can estimate through knowledge of their actions and the environment – even without experiencing immediate changes in performance. To test this hypothesis, we designed a hierarchical reinforcement learning task in which human participants (N 175) repeatedly chose their own goals and learned goal-conditioned policies.

Goal-directed planning presents a challenge for classical RL algorithms due to the vastness of the combinatorial state and goal spaces, while humans and animals adapt to complex environments, especially with diverse, non-stationary objectives, often employing intermediate goals for long-horizon tasks.Here, we propose a goal reduction mechanism for effectively deriving subgoals from arbitrary and distant original goals, using a novel loop-removal technique.The product of the method, called goal-reducer, distills high-quality subgoals from a replay buffer, all without the need for prior global environmental knowledge.Simulations show that the goal-reducer can be integrated into RL frameworks like Deep Q-learning and Soft Actor-Critic.It accelerates performance in both discrete and continuous action space tasks, such as grid world navigation and robotic arm manipulation, relative to the corresponding standard RL models.Moreover, the goal-reducer, when combined with a local policy, without iterative training, outperforms its integrated deep RL counterparts in solving a navigation task.This goal reduction mechanism also models human problem-solving.Comparing the model's performance and activation with human behavior and fMRI data in a treasure hunting task, we found matching representational patterns between an goal-reducer agent's components and corresponding human brain areas, particularly the vmPFC and basal ganglia. The results suggest that humans may use a similar computational framework for goal-directed behaviors.

Path planning, the problem of efficiently discovering high-reward trajectories, often requires optimizing a high-dimensional and multimodal reward function. Popular approaches like CEM and CMA-ES greedily focus on promising regions of the search space and may get trapped in local maxima. DOO and VOOT balance exploration and exploitation, but use space partitioning strategies independent of the reward function to be optimized. Recently, LaMCTS empirically learns to partition the search space in a reward-sensitive manner for black-box optimization. In this paper, we develop a novel formal regret analysis for when and why such an adaptive region partitioning scheme works.

This work tackles the problem of knowledge base completion (KBC) by walking through a knowledge graph. The graph walk is treated as an MDP with a known state transition model. MCTS is used to plan a trajectory through the graph, and these trajectories are used to train a value function with Q-learning. The Q-function is used to form a policy prior for the MCTS. A KBC-specific RNN structure is also proposed.

High-speed obstacle avoidance of uncrewed aerial vehicles (UAVs) in cluttered environments is a significant challenge. Existing UAV planning and obstacle avoidance systems can only fly at moderate speeds or at high speeds over empty or sparse fields. In this article, we propose a hyper-efficient perception and planning system for the high-speed obstacle avoidance of UAVs. The system mainly consists of three modules: 1) A novel incremental robocentric mapping method with distance and gradient information, which takes 89.5% less time compared to existing methods. 2) A novel obstacle-aware topological path search method that generates multiple distinct paths. 3) An adaptive gradient-based high-speed trajectory generation method with a novel time pre-allocation algorithm. With these innovations, the system has an excellent real-time performance with only milliseconds latency in each iteration, taking 79.24% less time than existing methods at high speeds (15 m/s in cluttered environments), allowing UAVs to fly swiftly and avoid obstacles in cluttered environments. The planned trajectory of the UAV is close to the global optimum in both temporal and spatial domains. Finally, extensive validations in both simulation and real-world experiments demonstrate the effectiveness of our proposed system for high-speed navigation in cluttered environments.