reinforcement learning

Successor representation supports structural learning of syllable sequences

Humans efficiently learn the temporal structure of speech, yet the underlying cognitive mechanisms remain unclear. Recent research in visuospatial sequential memory has proposed the successor representation (SR), important in reinforcement learning, which encodes multi-step transitional relationships. To test whether S...

A Concise Framework for AI-Driven Blockchain: Integrating DRL Consensus and GNN Security Auditing

This research introduces a novel, high-performance hybrid framework merging Deep Reinforcement Learning (DRL) for dynamic consensus optimization with Graph Neural Networks (GNN) for advanced smart contract security auditing. Traditional blockchain architectures frequently struggle with balancing scalability and securit...

Am I Still Coupled to My Environment? Task-agnostic Runtime Telemetry from the Reinforcement-learning Interaction Loop

A TSP solving model based on multiscale feature fusion and adaptive context awareness

Transformer-based deep reinforcement learning for the Traveling Salesman Problem (TSP) often struggles to capture spatial topology and avoid local optima. To address this, we propose a novel model featuring a Multi-Scale Grid Attention Encoder (MSGAE) to fuse local and global spatial features, alongside a bottleneck-en...

Adaptive Probabilistic Shielding by Learning MDPs for Safe Reinforcement Learning

Decision-Path Inverse Reconstruction: Observed Decision Outcomes as Structural Constraints on Latent Decision Paths

This working paper proposes Decision-Path Inverse Reconstruction (DPIR), an inverse analytical operation within the Atlas Insight Method (AIM). DPIR treats an observed decision outcome not only as the endpoint of a judgment process, but also as structural information that constrains the set of decision paths capable of...

Architectural Design Decisions for Monitoring Deployed Reinforcement Learning Systems

Hierarchical cooperative signal control for high-demand networks within the MFD framework: A deep reinforcement learning approach

Owing to its powerful modeling and decision-making capabilities in complex urban networks, deep reinforcement learning (DRL) has garnered significant attention in both regional signal control and perimeter control. However, regional control tends to fail under oversaturated conditions, while perimeter control often lea...

PEMBERDAYAAN GURU DALAM PEMBUATAN MODUL PEMBELAJARAN BERBASIS DEEP LEARNING MELALUI KOLABORASI GURU PADA MGMP BAHASA INDONESIA SE-KOTA PALANGKA RAYA

The implementation of the Deep Learning approach within the Kurikulum Merdeka framework requires teachers to design instructional modules that facilitate meaningful, mindful, and joyful learning experiences. However, a preliminary needs assessment among teachers in the Indonesian Language Subject Teachers' Consultation...

ProcessLLM-R1: A hypergraph retrieval-augmented and reinforcement learning-driven assembly process planning for commercial aircraft

Diffusion model-driven multi-objective collaborative optimization for building energy management using graph neural networks

A generative framework driven by conditional diffusion models integrated with graph neural networks integrated with graph neural networks is proposed to solve the high-dimensional nonlinear multi-objective energy optimization in building clusters.

ElasticScale: Elastic Large Language Model Reinforcement Learning Training on Heterogeneous Mobile Edge Clusters

ElasticScale is presented, an elastic orchestration system that organizes heterogeneous accelerators into disaggregated rollout and trainer instances, via a HeterogeneousRayWorkerGroup abstraction that manages non-uniform hardware topologies and a multi-instance Federated Weight Averaging protocol that aggregates updat...