We demonstrate the ability of large language models (LLMs) to perform iterative self-improvement of robot policies. An important insight of this paper is that LLMs have a built-in ability to perform (stochastic) numerical optimization and that this property can be leveraged for explainable robot policy search. Based on this insight, we introduce the SAS Prompt (Summarize, Analyze, Synthesize) – a single prompt that enables iterative learning and adaptation of robot behavior by combining the LLM's ability to retrieve, reason and optimize over previous robot traces in order to synthesize new, unseen behavior. Our approach can be regarded as an early example of a new family of explainable policy search methods that are entirely implemented within an LLM. We evaluate our approach both in simulation and on a real-robot table tennis task. Project website: sites.google.com/asu.edu/sas-llm/

Achieving human-level performance on real world tasks is a north star for the robotics community. We present the first learned robot agent that reaches amateur humanlevel performance in competitive table tennis. Table tennis is a physically demanding sport that takes humans years to master. We contribute (1) a hierarchical and modular policy architecture consisting of (i) low level controllers with their skill descriptors that model their capabilities and (ii) a high level controller that chooses the low level skills, (2) techniques for enabling zero-shot sim-to-real and curriculum building, including an iterative approach (train in sim, deploy in real), and (3) real time adaptation to unseen opponents. Policy performance was assessed through 29 robot vs. human matches of which the robot won 45 % (13/29). All humans were unseen players and their skill level varied from beginner to tournament level. Whilst the robot lost all matches vs. the most advanced players it won 100 % matches vs. beginners and 55 % matches vs. intermediate players, demonstrating solidly amateur humanlevel performance. Videos of the matches can be viewed here1.See sites https://google.com/view/competitive-robot-table-tennis.

Advances in robot learning have enabled robots to generate skills for a variety of tasks. Yet, robot learning is typically sample inefficient, struggles to learn from data sources exhibiting varied behaviors, and does not naturally incorporate constraints. These properties are critical for fast, agile tasks such as playing table tennis. Modern techniques for learning from demonstration improve sample efficiency and scale to diverse data, but are rarely evaluated on agile tasks. In the case of reinforcement learning, achieving good performance requires training on high-fidelity simulators. To overcome these limitations, we develop a novel diffusion modeling approach that is offline, constraint-guided, and expressive of diverse agile behaviors. The key to our approach is a kinematic constraint gradient guidance (KCGG) technique that computes gradients through both the forward kinematics of the robot arm and the diffusion model to direct the sampling process. KCGG minimizes the cost of violating constraints while simultaneously keeping the sampled trajectory in-distribution of the training data. We demonstrate the effectiveness of our approach for time-critical robotic tasks by evaluating KCGG in two challenging domains: simulated air hockey and real table tennis. In simulated air hockey, we achieved a 25.4% increase in block rate, while in table tennis, we achieved a 17.3% increase in success rate compared to imitation learning baselines.

Learning goal conditioned control in the real world is a challenging open problem in robotics. Reinforcement learning systems have the potential to learn autonomously via trial-and-error, but in practice the costs of manual reward design, ensuring safe exploration, and hyperparameter tuning are often enough to preclude real world deployment. Imitation learning approaches, on the other hand, offer a simple way to learn control in the real world, but typically require costly cu-rated demonstration data and lack a mechanism for continuous improvement. Recently, iterative imitation methods have been shown to be effective at relaxing both these constraints, learning goal directed control from undirected demonstration data, and improving continuously via self-supervised goal reaching. These approaches, however, have not yet been shown to scale beyond simple simulated environments. In this work, we present the first evidence that simple iterative imitation learning can scale to goal-directed behavior on a real robot in a dynamic setting: high speed, precision table tennis (e.g. “land the ball on this particular target”). We find that this approach offers a straightforward way to do continuous on-robot learning, without complexities such as reward design, value function learning, or sim-to-real transfer. We also find that this approach is scalable-sample efficient enough to train on a physical robot in just a few hours. In real world evaluations, we find that that the resulting policy can perform on par or better than amateur humans (with players sampled randomly from a robotics lab) at the task of returning the ball to specific targets on the table. Finally, we analyze the effect of an initial undirected bootstrap dataset size on performance, finding that a modest amount of unstructured demonstration data provided up-front drastically speeds up the convergence of a general purpose goal-reaching policy. See supplementary video for examples of the policy on a physical robot.

The impact of game content on the player experience is potentially more critical in casual games than in competitive games because of the diminished role of strategic or tactical diversions. Interestingly, until now procedural content generation (PCG) has nevertheless been investigated almost exclusively in the context of competitive, skills-based gaming. This paper therefore opens a new direction for PCG by placing it at the center of an entirely casual flower-breeding game platform called Petalz. That way, the behavior of players and their reactions to different game mechanics in a casual environment driven by PCG can be investigated. In particular, players in Petalz can: 1) trade their discoveries in a global marketplace; 2) respond to an incentive system that awards diversity; and 3) generate real-world 3-D replicas of their evolved flowers. With over 1900 registered online users and 38 646 unique evolved flowers, Petalz showcases the potential for PCG to enable these kinds of casual game mechanics, thus paving the way for continued innovation with PCG in casual gaming.

Behaviors evolved in simulation are often not robust to variations of their original training environment. Thus often researchers must train explicitly to encourage such robustness. Traditional methods of training for robustness typically apply multiple non-deterministic evaluations with carefully modeled noisy distributions for sensors and effectors. In practice, such training is often computationally expensive and requires crafting accurate models. Taking inspiration from nature, where animals react appropriately to encountered stimuli, this paper introduces a measure called reactivity, i.e. the tendency to seek and react to changes in environmental input, that is applicable in single deterministic trials and can encourage robustness without exposure to noise. The measure is tested in four different maze navigation tasks, where training with reactivity proves more robust than training without noise, and equally or more robust than training with noise when testing with moderate noise levels. In this way, the results demonstrate the counterintuitive fact that sometimes training with no exposure to noise at all can evolve individuals significantly more robust to noise than by explicitly training with noise. The conclusion is that training for reactivity may often be a computationally more efficient means to encouraging robustness in evolved behaviors. © 2012 Massachusetts Institute of Technology.

Multirobot domains are a challenge for learning algorithms because they require robots to learn to cooperate to achieve a common goal. The challenge only becomes greater when robots must perform heterogeneous tasks to reach that goal. Multiagent HyperNEAT is a neuroevolutionary method (i.e. a method that evolves neural networks) that has proven successful in several cooperative multiagent domains by exploiting the concept of policy geometry, which means the policies of team members are learned as a function of how they relate to each other based on canonical starting positions. This paper extends the multiagent HyperNEAT algorithm by introducing situational policy geometry, which allows each agent to encode multiple policies that can be switched depending on the agent's state. This concept is demonstrated both in simulation and in real Khepera III robots in a patrol and return task, where robots must cooperate to cover an area and return home when called. Robot teams that are trained with situational policy geometry are compared to teams that are not and shown to find solutions more consistently that are also able to transfer to the real world.

For domains in which fitness is subjective or difficult to express formally, interactive evolutionary computation (IEC) is a natural choice. It is possible that a collaborative process combining feedback from multiple users can improve the quality and quantity of generated artifacts. Picbreeder, a large-scale online experiment in collaborative interactive evolution (CIE), explores this potential. Picbreeder is an online community in which users can evolve and share images, and most importantly, continue evolving others' images. Through this process of branching from other images, and through continually increasing image complexity made possible by the underlying neuroevolution of augmenting topologies (NEAT) algorithm, evolved images proliferate unlike in any other current IEC system. This paper discusses not only the strengths of the Picbreeder approach, but its challenges and shortcomings as well, in the hope that lessons learned will inform the design of future CIE systems.

Research in neuroevolution-that is, evolving artificial neural networks (ANNs) through evolutionary algorithms-is inspired by the evolution of biological brains, which can contain trillions of connections. Yet while neuroevolution has produced successful results, the scale of natural brains remains far beyond reach. This article presents a method called hypercube-based NeuroEvolution of Augmenting Topologies (HyperNEAT) that aims to narrow this gap. HyperNEAT employs an indirect encoding called connective compositional pattern-producing networks (CPPNs) that can produce connectivity patterns with symmetries and repeating motifs by interpreting spatial patterns generated within a hypercube as connectivity patterns in a lower-dimensional space. This approach can exploit the geometry of the task by mapping its regularities onto the topology of the network, thereby shifting problem difficulty away from dimensionality to the underlying problem structure. Furthermore, connective CPPNs can represent the same connectivity pattern at any resolution, allowing ANNs to scale to new numbers of inputs and outputs without further evolution. HyperNEAT is demonstrated through visual discrimination and food-gathering tasks, including successful visual discrimination networks containing over eight million connections. The main conclusion is that the ability to explore the space of regular connectivity patterns opens up a new class of complex high-dimensional tasks to neuroevolution.

This paper argues that multiagent learning is a potential "killer application" for generative and developmental systems (GDS) because key challenges in learning to coordinate a team of agents are naturally addressed through indirect encodings and information reuse. For example, a significant problem for multiagent learning is that policies learned separately for different agent roles may nevertheless need to share a basic skill set, forcing the learning algorithm to reinvent the wheel for each agent. GDS is a good match for this kind of problem because it specializes in ways to encode patterns of related yet varying motifs. In this paper, to establish the promise of this capability, the Hypercube-based NeuroEvolution of Augmenting Topologies (HyperNEAT) generative approach to evolving neurocontrollers learns a set of coordinated policies encoded by a single genome representing a team of predator agents that work together to capture prey. Experimental results show that it is not only possible, but beneficial to encode a heterogeneous team of agents with an indirect encoding. The main contribution is thus to open up a significant new application domain for GDS.

A significant problem for evolving artificial neural networks is that the physical arrangement of sensors and effectors is invisible to the evolutionary algorithm. For example, in this paper, directional sensors and effectors are placed around the circumference of a robot in analogous arrangements. This configuration ensures that there is a useful geometric correspondence between sensors and effectors. However, if sensors are mapped to a single input layer and the effectors to a single output layer (as is typical), evolution has no means to exploit this fortuitous arrangement. To address this problem, this paper presents a novel generative encoding called connective Compositional Pattern Producing Networks (connective CPPNs) that can effectively detect and capitalize on geometric relationships among sensors and effectors. The key insight is that sensors and effectors with consistent geometric relationships can be exploited by a repeating motif in the neural architecture. Thus, by employing an encoding that can discover such motifs as a function of network geometry, it becomes possible to exploit it. In this paper, a method for evolving connective CPPNs called Hypercube-based Neuroevolution of Augmenting Topologies (HyperNEAT) discovers sensible repeating motifs that take advantage of two different placement schemes, demonstrating the utility of such an approach.