We consider the inverse reinforcement learning (IRL) problem, where an unknown reward function of some Markov decision process is estimated based on observed expert demonstrations. In most existing approaches, IRL is formulated and solved as a nonconvex optimization problem, posing challenges in scenarios where robustness and reproducibility are critical. We discuss a convex formulation of the IRL problem (CIRL) initially proposed by Ng and Russel, and reformulate the problem such that the domain-specific language CVXPY can be applied directly to specify and solve the convex problem. We also extend the CIRL problem to scenarios where the expert policy is not given analytically but by trajectory as state-action pairs, which can be strongly inconsistent with optimality, by augmenting some of the constraints. Theoretical analysis and practical implementation for hyperparameter auto-selection are introduced. This note helps the users to easily apply CIRL for their problems, without background knowledge on convex optimization.

"First, do no harm" faces a fundamental challenge in artificial intelligence: how can we specify what constitutes harm? While prior work treats harm specification as a technical hurdle to be overcome through better algorithms or more data, we argue this assumption is unsound. Drawing on information theory, we demonstrate that complete harm specification is fundamentally impossible for any system where harm is defined external to its specifications. This impossibility arises from an inescapable information-theoretic gap: the entropy of harm H(O) always exceeds the mutual information I(O;I) between ground truth harm O and a system's specifications I. We introduce two novel metrics: semantic entropy H(S) and the safety-capability ratio I(O;I)/H(O), to quantify these limitations. Through a progression of increasingly sophisticated specification attempts, we show why each approach must fail and why the resulting gaps are not mere engineering challenges but fundamental constraints akin to the halting problem. These results suggest a paradigm shift: rather than pursuing complete specifications, AI alignment research should focus on developing systems that can operate safely despite irreducible specification uncertainty.

This work introduces a toolchain for applying Reinforcement Learning (RL), specifically the Deep Deterministic Policy Gradient (DDPG) algorithm, in safety-critical real-world environments. As an exemplary application, transient load control is demonstrated on a single-cylinder internal combustion engine testbench in Homogeneous Charge Compression Ignition (HCCI) mode, that offers high thermal efficiency and low emissions. However, HCCI poses challenges for traditional control methods due to its nonlinear, autoregressive, and stochastic nature. RL provides a viable solution, however, safety concerns, such as excessive pressure rise rates, must be addressed when applying to HCCI. A single unsuitable control input can severely damage the engine or cause misfiring and shut down. Additionally, operating limits are not known a priori and must be determined experimentally. To mitigate these risks, real-time safety monitoring based on the k-nearest neighbor algorithm is implemented, enabling safe interaction with the testbench. The feasibility of this approach is demonstrated as the RL agent learns a control policy through interaction with the testbench. A root mean square error of 0.1374 bar is achieved for the indicated mean effective pressure, comparable to neural network-based controllers from the literature. The toolchain's flexibility is further demonstrated by adapting the agent's policy to increase ethanol energy shares, promoting renewable fuel use while maintaining safety. This RL approach addresses the longstanding challenge of applying RL to safety-critical real-world environments. The developed toolchain, with its adaptability and safety mechanisms, paves the way for future applicability of RL in engine testbenches and other safety-critical settings.

Reinforcement learning (RL) has demonstrated success in automating insulin dosing in simulated type 1 diabetes (T1D) patients but is currently unable to incorporate patient expertise and preference. This work introduces PAINT (Preference Adaptation for INsulin control in T1D), an original RL framework for learning flexible insulin dosing policies from patient records. PAINT employs a sketch-based approach for reward learning, where past data is annotated with a continuous reward signal to reflect patient's desired outcomes. Labelled data trains a reward model, informing the actions of a novel safety-constrained offline RL algorithm, designed to restrict actions to a safe strategy and enable preference tuning via a sliding scale. In-silico evaluation shows PAINT achieves common glucose goals through simple labelling of desired states, reducing glycaemic risk by 15% over a commercial benchmark. Action labelling can also be used to incorporate patient expertise, demonstrating an ability to pre-empt meals (+10% time-in-range post-meal) and address certain device errors (-1.6% variance post-error) with patient guidance. These results hold under realistic conditions, including limited samples, labelling errors, and intra-patient variability. This work illustrates PAINT's potential in real-world T1D management and more broadly any tasks requiring rapid and precise preference learning under safety constraints.

Deep reinforcement learning has achieved remarkable success in learning control policies from pixels across a wide range of tasks, yet its application remains hindered by low sample efficiency, requiring significantly more environment interactions than humans to reach comparable performance. Model-based reinforcement learning (MBRL) offers a solution by leveraging learnt world models to generate simulated experience, thereby improving sample efficiency. However, in visually complex environments, small or dynamic elements can be critical for decision-making. Yet, traditional MBRL methods in pixel-based environments typically rely on auto-encoding with an $L_2$ loss, which is dominated by large areas and often fails to capture decision-relevant details. To address these limitations, we propose an object-centric MBRL pipeline, which integrates recent advances in computer vision to allow agents to focus on key decision-related elements. Our approach consists of four main steps: (1) annotating key objects related to rewards and goals with segmentation masks, (2) extracting object features using a pre-trained, frozen foundation vision model, (3) incorporating these object features with the raw observations to predict environmental dynamics, and (4) training the policy using imagined trajectories generated by this object-centric world model. Building on the efficient MBRL algorithm STORM, we call this pipeline OC-STORM. We demonstrate OC-STORM's practical value in overcoming the limitations of conventional MBRL approaches on both Atari games and the visually complex game Hollow Knight.

We address the problem of quantum reinforcement learning (QRL) under model-free settings with quantum oracle access to the Markov Decision Process (MDP). This paper introduces a Quantum Natural Policy Gradient (QNPG) algorithm, which replaces the random sampling used in classical Natural Policy Gradient (NPG) estimators with a deterministic gradient estimation approach, enabling seamless integration into quantum systems. While this modification introduces a bounded bias in the estimator, the bias decays exponentially with increasing truncation levels. This paper demonstrates that the proposed QNPG algorithm achieves a sample complexity of $\tilde{\mathcal{O}}(\epsilon^{-1.5})$ for queries to the quantum oracle, significantly improving the classical lower bound of $\tilde{\mathcal{O}}(\epsilon^{-2})$ for queries to the MDP.

Summary and Contributions: The paper proposed a learning system for applying deep reinforcement learning algorithms to an optical interferometer aligning task. To achieve successful learning, the authors first developed a simulation tool for simulating interferometric patterns captured by the camera, which is used to train the DRL agents using dueling ddqn. To enable the controller to transfer to the real-world system, they apply domain randomization during training. They demonstrated successful learning of an agent that can perform interferometer aligning tasks and outperforms human operators. Tha main contribution of the paper is that it demonstrates successful application to a new domain of tasks and achieves human-level performance.

Summary: This paper proposes a new algorithm that help stabilize off-policy Q-learning. The idea is to introduce approximate Bellman updates that are based on constraint actions sampled only from the support of the training data distribution. The paper shows the main source of instability is the boostrapping error. The boostrapping process might use actions that do not lie in the training data distribution. This work shows a way to mitigate this issue.

The reviewers were in consensus about the merits of this paper, in particular the value of the proposed approach and the theoretical analysis. Some concerns were raised about the experimental validation but these have been alleviated by the new results and baselines added during rebuttal. Some concerns remain regarding the clarity of the paper. The authors claim to have revised the text but we are not able to see it to validate that it has improved in this respect. The authors are strongly encouraged to put some real effort into improving the clarity of the final version.

I am on the fence for this paper. I like that the approach is simple. I like that it was tested in different domains. I like that several different views of performance are included. Nevertheless I have many questions, that need resolution before I can increase my score.