Safe Reinforcement Learning on the Constraint Manifold: Theory and Applications

Integrating learning-based techniques, especially reinforcement learning, into robotics is promising for solving complex problems in unstructured environments. However, most existing approaches are trained in well-tuned simulators and subsequently deployed on real robots without online fine-tuning. In this setting, the simulation's realism seriously impacts the deployment's success rate. Instead, learning with real-world interaction data offers a promising alternative: not only eliminates the need for a fine-tuned simulator but also applies to a broader range of tasks where accurate modeling is unfeasible. One major problem for on-robot reinforcement learning is ensuring safety, as uncontrolled exploration can cause catastrophic damage to the robot or the environment. Indeed, safety specifications, often represented as constraints, can be complex and non-linear, making safety challenging to guarantee in learning systems. In this paper, we show how we can impose complex safety constraints on learning-based robotics systems in a principled manner, both from theoretical and practical points of view. Our approach is based on the concept of the Constraint Manifold, representing the set of safe robot configurations. Exploiting differential geometry techniques, i.e., the tangent space, we can construct a safe action space, allowing learning agents to sample arbitrary actions while ensuring safety. We demonstrate the method's effectiveness in a real-world Robot Air Hockey task, showing that our method can handle high-dimensional tasks with complex constraints. Videos of the real robot experiments are available on the project website (https://puzeliu.github.io/TRO-ATACOM).

ZSL-RPPO: Zero-Shot Learning for Quadrupedal Locomotion in Challenging Terrains using Recurrent Proximal Policy Optimization

We present ZSL-RPPO, an improved zero-shot learning architecture that overcomes the limitations of teacher-student neural networks and enables generating robust, reliable, and versatile locomotion for quadrupedal robots in challenging terrains. We propose a new algorithm RPPO (Recurrent Proximal Policy Optimization) that directly trains recurrent neural network in partially observable environments and results in more robust training using domain randomization. Our locomotion controller supports extensive perturbation across simulation-to-reality transfer for both intrinsic and extrinsic physical parameters without further fine-tuning. This can avoid the significant decline of student's performance during simulation-to-reality transfer and therefore enhance the robustness and generalization of the locomotion controller. We deployed our controller on the Unitree A1 and Aliengo robots in real environment and exteroceptive perception is provided by either a solid-state Lidar or a depth camera. Our locomotion controller was tested in various challenging terrains like slippery surfaces, Grassy Terrain, and stairs. Our experiment results and comparison show that our approach significantly outperforms the state-of-the-art.

Bayesian Reward Models for LLM Alignment

To ensure that large language model (LLM) responses are helpful and non-toxic, we usually fine-tune a reward model on human preference data. We then select policy responses with high rewards (best-of-n sampling) or further optimize the policy to produce responses with high rewards (reinforcement learning from human feedback). However, this process is vulnerable to reward overoptimization or hacking, in which the responses selected have high rewards due to errors in the reward model rather than a genuine preference. This is especially problematic as the prompt or response diverges from the training data. It should be possible to mitigate these issues by training a Bayesian reward model, which signals higher uncertainty further from the training data distribution. Therefore, we trained Bayesian reward models using Laplace-LoRA (Yang et al., 2024) and found that the resulting uncertainty estimates can successfully mitigate reward overoptimization in best-of-n sampling.

Pangu-Agent: A Fine-Tunable Generalist Agent with Structured Reasoning

A key method for creating Artificial Intelligence (AI) agents is Reinforcement Learning (RL). However, constructing a standalone RL policy that maps perception to action directly encounters severe problems, chief among them being its lack of generality across multiple tasks and the need for a large amount of training data. The leading cause is that it cannot effectively integrate prior information into the perception-action cycle when devising the policy. Large language models (LLMs) emerged as a fundamental way to incorporate cross-domain knowledge into AI agents but lack crucial learning and adaptation toward specific decision problems. This paper presents a general framework model for integrating and learning structured reasoning into AI agents' policies. Our methodology is motivated by the modularity found in the human brain. The framework utilises the construction of intrinsic and extrinsic functions to add previous understandings of reasoning structures. It also provides the adaptive ability to learn models inside every module or function, consistent with the modular structure of cognitive processes. We describe the framework in-depth and compare it with other AI pipelines and existing frameworks. The paper explores practical applications, covering experiments that show the effectiveness of our method. Our results indicate that AI agents perform and adapt far better when organised reasoning and prior knowledge are embedded. This opens the door to more resilient and general AI agent systems.

Why Can Large Language Models Generate Correct Chain-of-Thoughts?

This paper delves into the capabilities of large language models (LLMs), specifically focusing on advancing the theoretical comprehension of chain-of-thought prompting. We investigate how LLMs can be effectively induced to generate a coherent chain of thoughts. To achieve this, we introduce a two-level hierarchical graphical model tailored for natural language generation. Within this framework, we establish a compelling geometrical convergence rate that gauges the likelihood of an LLM-generated chain of thoughts compared to those originating from the true language. Our findings provide a theoretical justification for the ability of LLMs to produce the correct sequence of thoughts (potentially) explaining performance gains in tasks demanding reasoning skills.

Are Random Decompositions all we need in High Dimensional Bayesian Optimisation?

Learning decompositions of expensive-to-evaluate black-box functions promises to scale Bayesian optimisation (BO) to high-dimensional problems. However, the success of these techniques depends on finding proper decompositions that accurately represent the black-box. While previous works learn those decompositions based on data, we investigate data-independent decomposition sampling rules in this paper. We find that data-driven learners of decompositions can be easily misled towards local decompositions that do not hold globally across the search space. Then, we formally show that a random tree-based decomposition sampler exhibits favourable theoretical guarantees that effectively trade off maximal information gain and functional mismatch between the actual black-box and its surrogate as provided by the decomposition. Those results motivate the development of the random decomposition upper-confidence bound algorithm (RDUCB) that is straightforward to implement - (almost) plug-and-play - and, surprisingly, yields significant empirical gains compared to the previous state-of-the-art on a comprehensive set of benchmarks. We also confirm the plug-and-play nature of our modelling component by integrating our method with HEBO, showing improved practical gains in the highest dimensional tasks from Bayesmark.

Contextual Causal Bayesian Optimisation

Causal Bayesian optimisation (CaBO) combines causality with Bayesian optimisation (BO) and shows that there are situations where the optimal reward is not achievable if causal knowledge is ignored. While CaBO exploits causal relations to determine the set of controllable variables to intervene on, it does not exploit purely observational variables and marginalises them. We show that, in general, utilising a subset of observational variables as a context to choose the values of interventional variables leads to lower cumulative regrets. We propose a general framework of contextual causal Bayesian optimisation that efficiently searches through combinations of controlled and contextual variables, known as policy scopes, and identifies the one yielding the optimum. We highlight the difficulties arising from the application of the causal acquisition function currently used in CaBO to select the policy scope in contextual settings and propose a multi-armed bandits based selection mechanism. We analytically show that well-established methods, such as contextual BO (CoBO) or CaBO, are not able to achieve the optimum in some cases, and empirically show that the proposed method achieves sub-linear regret in various environments and under different configurations.

Fast Kinodynamic Planning on the Constraint Manifold with Deep Neural Networks

Motion planning is a mature area of research in robotics with many well-established methods based on optimization or sampling the state space, suitable for solving kinematic motion planning. However, when dynamic motions under constraints are needed and computation time is limited, fast kinodynamic planning on the constraint manifold is indispensable. In recent years, learning-based solutions have become alternatives to classical approaches, but they still lack comprehensive handling of complex constraints, such as planning on a lower-dimensional manifold of the task space while considering the robot's dynamics. This paper introduces a novel learning-to-plan framework that exploits the concept of constraint manifold, including dynamics, and neural planning methods. Our approach generates plans satisfying an arbitrary set of constraints and computes them in a short constant time, namely the inference time of a neural network. This allows the robot to plan and replan reactively, making our approach suitable for dynamic environments. We validate our approach on two simulated tasks and in a demanding real-world scenario, where we use a Kuka LBR Iiwa 14 robotic arm to perform the hitting movement in robotic Air Hockey.

AntBO: Towards Real-World Automated Antibody Design with Combinatorial Bayesian Optimisation

Antibodies are canonically Y-shaped multimeric proteins capable of highly specific molecular recognition. The CDRH3 region located at the tip of variable chains of an antibody dominates antigen-binding specificity. Therefore, it is a priority to design optimal antigen-specific CDRH3 regions to develop therapeutic antibodies to combat harmful pathogens. However, the combinatorial nature of CDRH3 sequence space makes it impossible to search for an optimal binding sequence exhaustively and efficiently, especially not experimentally. Here, we present AntBO: a Combinatorial Bayesian Optimisation framework enabling efficient in silico design of the CDRH3 region. Ideally, antibodies should bind to their target antigen and be free from any harmful outcomes. Therefore, we introduce the CDRH3 trust region that restricts the search to sequences with feasible developability scores. To benchmark AntBO, we use the Absolut! software suite as a black-box oracle because it can score the target specificity and affinity of designed antibodies in silico in an unconstrained fashion. The results across 188 antigens demonstrate the benefit of AntBO in designing CDRH3 regions with diverse biophysical properties. In under 200 protein designs, AntBO can suggest antibody sequences that outperform the best binding sequence drawn from 6.9 million experimentally obtained CDRH3s and a commonly used genetic algorithm baseline. Additionally, AntBO finds very-high affinity CDRH3 sequences in only 38 protein designs whilst requiring no domain knowledge. We conclude AntBO brings automated antibody design methods closer to what is practically viable for in vitro experimentation.