Tag
A comprehensive survey of 86 recent studies on the security of deep reinforcement learning in autonomous driving, systematically categorizing attacks and defenses by threat models in single- and multi-agent settings and assessing their applicability to autonomous driving. Accepted for publication in IEEE Transactions on Intelligent Transportation Systems.
该论文通过四个基于Modelica的可复现办公楼案例研究,评估深度强化学习在HVAC监督控制中的实际收益,发现无储能场景下DRL相对良好规则基线的边际极小(0.18%-1.3%),而在储能场景中优势更明显但仍远未达到完美预见最优。
PF-Exec 提出两层决策架构(锚定潜在方向决策器 + 每维增益学习执行器),用于大规模无人机中继网络的负载均衡,联合优化关联、带宽、功率与悬停位置,并在高压力场景下保持接近 1 的负载公平性。
该论文提出一种将情感分析深度融入Actor-Critic类深度强化学习智能体(A2C、PPO、DDPG、TD3、SAC)的股票交易策略,发现结合Q学习的策略优化方法在融合情感信息后能显著提升盈利与风险规避能力。
UC Berkeley 教授、Physical Intelligence 联合创始人 Sergey Levine 将其春季深度强化学习课程(CS 185/285)全部 25 讲加客座讲座共约 36 小时的录像与作业代码免费公开在 YouTube 和 GitHub,内容涵盖模仿学习、策略梯度、离线 RL、RL for LLMs 等主题,与 ALOHA、π0 等机器人学习方法及当下推理模型训练直接相关。
The paper proposes PX-UAP, a method using probabilistic robustness and explainable AI to generate universal adversarial perturbations against deep reinforcement learning-based intrusion detection systems, demonstrating improved attack effectiveness in experiments.
This paper introduces a graph-based deep reinforcement learning framework for the one-dimensional bin packing problem, reducing optimality gaps compared to existing methods and enabling zero-shot generalization across instance sizes.
RefinePPO introduces iterative action refinement for continuous control policies, achieving performance comparable or better than standard PPO with faster convergence in benchmark tasks.
This paper proposes a composite-gradient learning method that integrates deep reinforcement learning and model predictive control for shared control authority in autonomous systems, with evaluations on traffic networks showing modest benefits under strong interaction.
The paper proposes a Decision Transformer-based approach for optimizing UAV-mounted RIS-assisted dynamic D2D communications, demonstrating cross-scenario generalization and efficient zero-shot transfer.
This paper conducts a systematic review of AI for data center energy optimization, identifies research gaps, and proposes the CLEAR-DC framework to close the optimizer-load loop by integrating energy, carbon, and water metrics.
This paper explores the impact of network topology and opponent information on the emergence of cooperation in multi-agent reinforcement learning systems, specifically in the Iterated Prisoner's Dilemma, finding that graph structure and information availability significantly influence cooperative strategies.
This research paper proposes a simulation framework using deep reinforcement learning for controlling connected and automated vehicle platoon joining in mixed traffic, showing that PPO achieves high success rates while highlighting trade-offs between safety and efficiency.
This paper presents a modified JAMPR deep reinforcement learning model to solve the Pickup and Delivery problem with Capacity and Time Window constraints (CPDPTW), offering fast optimal solutions for small to medium-sized instances and suboptimal solutions for larger ones.
This paper introduces the Travelling Thief Problem with Drone (TTP-D), which jointly optimizes ground routing, drone synchronization, and item selection using mixed-integer programming, metaheuristics, and attention-based deep reinforcement learning.
Berkeley's Deep RL class is now fully available online on YouTube for free, as announced by Sergey Levine.
The lectures for the Deep Reinforcement Learning course CS185/285 from UC Berkeley are now available online for public viewing.
This paper proposes a proxemics-based reward formulation for deep reinforcement learning social navigation, modeling human personal space as Gaussian-mixture fields to improve social compliance while maintaining navigation efficiency.
This paper presents a deep reinforcement learning approach for solving vehicle routing problems, demonstrated through three industrial truck planning case studies. The proposed method achieves over 10% cost reduction compared to baseline results and discusses generalization to more VRP variants.
This paper proposes PLAN, a lightweight parallel liquid-inspired approximation network for efficient representation learning in flexible job shop scheduling, achieving better makespan and lower inference latency with fewer parameters than state-of-the-art baselines.