arXiv:2512.05291v1 Announce Type: new 
Abstract: Actor-critic (AC) methods are a cornerstone of reinforcement learning (RL) but offer limited interpretability. Current explainable RL methods seldom use state attributions to assist training. Rather, they treat all state features equally, thereby neglecting the heterogeneous impacts of individual state dimensions on the reward. We propose RKHS--SHAP-based Advanced Actor--Critic (RSA2C), an attribution-aware, kernelized, two-timescale AC algorithm, including Actor, Value Critic, and Advantage Critic. The Actor is instantiated in a vector-valued reproducing kernel Hilbert space (RKHS) with a Mahalanobis-weighted operator-valued kernel, while the Value Critic and Advantage Critic reside in scalar RKHSs. These RKHS-enhanced components use sparsified dictionaries: the Value Critic maintains its own dictionary, while the Actor and Advantage Critic share one. State attributions, computed from the Value Critic via RKHS--SHAP (kernel mean embedding for on-manifold expectations and conditional mean embedding for off-manifold expectations), are converted into Mahalanobis-gated weights that modulate Actor gradients and Advantage Critic targets. Theoretically, we derive a global, non-asymptotic convergence bound under state perturbations, showing stability through the perturbation-error term and efficiency through the convergence-error term. Empirical results on three standard continuous-control environments show that our algorithm achieves efficiency, stability, and interpretability.

تقدم دراسة جديدة خوارزمية RSA2C المستندة إلى RKHS--SHAP، وهي ابتكار في التعلم المعزز يعزز من قابلية التفسير من خلال استخدام نسب الحالة. تتضمن هذه الطريقة نهجًا قائمًا على النواة داخل فضاءات هيلبرت المستنسخة، مما يسمح بفهم أكثر دقة لكيفية تأثير أبعاد الحالة الفردية على المكافآت.

Un nuevo estudio presenta el algoritmo RSA2C basado en RKHS--SHAP, una innovación en el aprendizaje por refuerzo que mejora la interpretabilidad al utilizar atribuciones de estado. Este método incorpora un enfoque kernelizado dentro de los espacios de Hilbert reproducentes, lo que permite una comprensión más matizada de cómo las dimensiones individuales del estado afectan las recompensas.

Une nouvelle étude présente l'algorithme RSA2C basé sur RKHS--SHAP, une innovation dans l'apprentissage par renforcement qui améliore l'interprétabilité en utilisant des attributions d'état. Cette méthode intègre une approche kernelisée au sein des espaces de Hilbert reproduisants, permettant une compréhension plus nuancée de la manière dont les dimensions d'état individuelles affectent les récompenses.

A new study introduces the RKHS--SHAP-based Advanced Actor--Critic (RSA2C), an innovative algorithm in reinforcement learning that enhances interpretability by utilizing state attributions. This method incorporates a kernelized approach within reproducing kernel Hilbert spaces, allowing for a more nuanced understanding of how individual state dimensions affect rewards.

Bridging Interpretability and Optimization: Provably Attribution-Weighted Actor-Critic in Reproducing Kernel Hilbert Spaces

arXiv:2512.06730v1 Announce Type: new 
Abstract: Patients with motor dysfunction show low subjective engagement in rehabilitation training. Traditional SSVEP-based brain-computer interface (BCI) systems rely heavily on external visual stimulus equipment, limiting their practicality in real-world settings. This study proposes an augmented reality steady-state visually evoked potential (AR-SSVEP) system to address the lack of patient initiative and the high workload on therapists. Firstly, we design four HoloLens 2-based EEG classes and collect EEG data from seven healthy subjects for analysis. Secondly, we build upon the conventional CNN-BiLSTM architecture by integrating a multi-head attention mechanism (MACNN-BiLSTM). We extract ten temporal-spectral EEG features and feed them into a CNN to learn high-level representations. Then, we use BiLSTM to model sequential dependencies and apply a multi-head attention mechanism to highlight motor-intention-related patterns. Finally, the SHAP (SHapley Additive exPlanations) method is applied to visualize EEG feature contributions to the neural network's decision-making process, enhancing the model's interpretability. These findings enhance real-time motor intention recognition and support recovery in patients with motor impairments.

قدمت دراسة حديثة نظامًا لتحفيز بصري ثابت في الواقع المعزز (AR-SSVEP) يهدف إلى تحسين التعرف على نية الحركة من خلال بنية CNN-BiLSTM جديدة وتحليل SHAP على بيانات EEG. تم اختبار هذا النهج باستخدام بيانات EEG تم جمعها من سبعة أشخاص أصحاء، مما يعالج القيود المفروضة على واجهات الدماغ-الكمبيوتر التقليدية التي تعتمد على المحفزات البصرية الخارجية.

Un estudio reciente presentó un sistema de potencial evocado visual en estado estable en realidad aumentada (AR-SSVEP) que busca mejorar el reconocimiento de la intención motora mediante una nueva arquitectura CNN-BiLSTM y análisis SHAP sobre datos EEG. Este enfoque se probó utilizando datos EEG recolectados de siete sujetos sanos, abordando las limitaciones de las interfaces cerebro-computadora (BCI) tradicionales que dependen de estímulos visuales externos.

Une étude récente a introduit un système de potentiel évoqué visuel à état stable en réalité augmentée (AR-SSVEP) visant à améliorer la reconnaissance de l'intention motrice grâce à une nouvelle architecture CNN-BiLSTM et à une analyse SHAP sur des données EEG. Cette approche a été testée à l'aide de données EEG collectées auprès de sept sujets sains, abordant les limitations des interfaces cerveau-ordinateur (BCI) traditionnelles qui dépendent de stimuli visuels externes.

A recent study introduced an augmented reality steady-state visually evoked potential (AR-SSVEP) system aimed at enhancing motor intention recognition through a novel CNN-BiLSTM architecture and SHAP analysis on EEG data. This approach was tested using EEG data collected from seven healthy subjects, addressing the limitations of traditional brain-computer interfaces (BCIs) that rely on external visual stimuli.

Enhancing Interpretability of AR-SSVEP-Based Motor Intention Recognition via CNN-BiLSTM and SHAP Analysis on EEG Data

arXiv:2512.07178v1 Announce Type: cross 
Abstract: Explainable Artificial Intelligence (XAI) has become an increasingly important area of research, particularly as machine learning models are deployed in high-stakes domains. Among various XAI approaches, SHAP (SHapley Additive exPlanations) has gained prominence due to its ability to provide both global and local explanations across different machine learning models. While SHAP effectively visualizes feature importance, it often lacks contextual explanations that are meaningful for end-users, especially those without technical backgrounds. To address this gap, we propose a Python package that extends SHAP by integrating it with a large language model (LLM), specifically OpenAI's GPT, to generate contextualized textual explanations. This integration is guided by user-defined parameters (such as feature aliases, descriptions, and additional background) to tailor the explanation to both the model context and the user perspective. We hypothesize that this enhancement can improve the perceived understandability of SHAP explanations. To evaluate the effectiveness of the proposed package, we applied it in a healthcare-related case study and conducted user evaluations involving real end-users. The results, based on Likert-scale surveys and follow-up interviews, indicate that the generated explanations were perceived as more understandable and contextually appropriate compared to visual-only outputs. While the findings are preliminary, they suggest that combining visualization with contextualized text may support more user-friendly and trustworthy model explanations.

تم اقتراح حزمة بايثون جديدة تُدعى ContextualSHAP لتحسين SHAP (SHapley Additive exPlanations) من خلال دمجها مع GPT من OpenAI، مما يسمح بتوليد تفسيرات نصية سياقية مصممة وفقًا للمعلمات المحددة من قبل المستخدم. يهدف هذا التطور إلى سد الفجوة في توفير تفسيرات ذات مغزى للمستخدمين، وخاصة أولئك الذين يفتقرون إلى الخبرة التقنية.

Se ha propuesto un nuevo paquete de Python llamado ContextualSHAP para mejorar SHAP (SHapley Additive exPlanations) integrándolo con GPT de OpenAI, lo que permite la generación de explicaciones textuales contextualizadas adaptadas a parámetros definidos por el usuario. Este desarrollo busca cerrar la brecha en la provisión de explicaciones significativas para los usuarios, especialmente aquellos sin experiencia técnica.

Un nouveau package Python nommé ContextualSHAP a été proposé pour améliorer SHAP (SHapley Additive exPlanations) en l'intégrant avec GPT d'OpenAI, permettant la génération d'explications textuelles contextualisées adaptées aux paramètres définis par l'utilisateur. Ce développement vise à combler le fossé en fournissant des explications significatives pour les utilisateurs, en particulier ceux manquant d'expertise technique.

A new Python package named ContextualSHAP has been proposed to enhance SHAP (SHapley Additive exPlanations) by integrating it with OpenAI's GPT, allowing for the generation of contextualized textual explanations tailored to user-defined parameters. This development aims to bridge the gap in providing meaningful explanations for end-users, particularly those lacking technical expertise.

Bridging Interpretability and Optimization: Provably Attribution-Weighted Actor-Critic in Reproducing Kernel Hilbert Spaces

Was this article worth reading? Share it

LucidQuery AI

ShareSpeak

HubRE AI