Reinforced Knowledge Distillation: Multi-class Imbalanced Classifier Based on Policy Gradient Reinforcement Learning

2021 ◽  
Author(s):  
Saite Fan ◽  
Xinmin Zhang ◽  
Zhihuan Song
Keyword(s):  
Reinforcement Learning ◽  
Policy Gradient ◽  
Knowledge Distillation

scholarly journals Formula-E race strategy development using distributed policy gradient reinforcement learning

2021 ◽  
Vol 216 ◽  
pp. 106781
Author(s):  
Xuze Liu ◽  
Abbas Fotouhi ◽  
Daniel J. Auger
Keyword(s):  
Reinforcement Learning ◽  
Strategy Development ◽  
Policy Gradient

Preceding vehicle following algorithm with human driving characteristics

2021 ◽  
pp. 095440702098154
Author(s):  
Feng Pan ◽  
Hong Bao
Keyword(s):  
Reinforcement Learning ◽  
Weight Vector ◽  
Gradient Algorithm ◽  
Inner Product ◽  
Inverse Reinforcement Learning ◽  
Reward Function ◽  
Human Driver ◽  
Policy Gradient ◽  
Preceding Vehicle ◽  
Action Spaces

This paper proposes a new approach of using reinforcement learning (RL) to train an agent to perform the task of vehicle following with human driving characteristics. We refer to the ideal of inverse reinforcement learning to design the reward function of the RL model. The factors that need to be weighed in vehicle following were vectorized into reward vectors, and the reward function was defined as the inner product of the reward vector and weights. Driving data of human drivers was collected and analyzed to obtain the true reward function. The RL model was trained with the deterministic policy gradient algorithm because the state and action spaces are continuous. We adjusted the weight vector of the reward function so that the value vector of the RL model could continuously approach that of a human driver. After dozens of rounds of training, we selected the policy with the nearest value vector to that of a human driver and tested it in the PanoSim simulation environment. The results showed the desired performance for the task of an agent following the preceding vehicle safely and smoothly.

The synthesis method of regulators for multichannel systems using neural networks

2020 ◽  
pp. 111-118
Author(s):  
Александр Александрович Воевода ◽  
Дмитрий Олегович Романников
Keyword(s):  
Neural Networks ◽  
Reinforcement Learning ◽  
Automatic Control ◽  
Control Systems ◽  
Synthesis Method ◽  
Neural Net ◽  
Initial State ◽  
Set Point ◽  
Automatic Control Systems ◽  
Policy Gradient

Синтез регуляторов для многоканальных систем - актуальная и сложная задача. Одним из возможных способов синтеза является применение нейронных сетей. Нейронный регулятор либо обучают на предварительно рассчитанных данных, либо используют для настройки параметров ПИД-регулятора из начального устойчивого положения замкнутой системы. Предложено использовать нейронные сети для регулирования двухканального объекта, при этом обучение будет выполняться из неустойчивого (произвольного) начального положения с применением методов обучения нейронных сетей с подкреплением. Предложена структура нейронной сети и замкнутой системы, в которой уставка задается при помощи входного параметра нейронной сети регулятора The problem for synthesis of automatic control systems is hard, especially for multichannel objects. One of the approaches is the use of neural networks. For the approaches that are based on the use of reinforcement learning, there is an additional issue - supporting of range of values for the set points. The method of synthesis of automatic control systems using neural networks and the process of its learning with reinforcement learning that allows neural networks learning for supporting regulation is proposed in the predefined range of set points. The main steps of the method are 1) to form a neural net input as a state of the object and system set point; 2) to perform modelling of the system with a set of randomly generated set points from the desired range; 3) to perform a one-step of the learning using the Deterministic Policy Gradient method. The originality of the proposed method is that, in contrast to existing methods of using a neural network to synthesize a controller, the proposed method allows training a controller from an unstable initial state in a closed system and set of a range of set points. The method was applied to the problem of stabilizing the outputs of a two-channel object, for which stabilization both outputs and the first near the input set point is required

scholarly journals Deep reinforcement learning based control for Autonomous Vehicles in CARLA

2022 ◽  
Author(s):  
Óscar Pérez-Gil ◽  
Rafael Barea ◽  
Elena López-Guillén ◽  
Luis M. Bergasa ◽  
Carlos Gómez-Huélamo ◽  
...  
Keyword(s):  
Reinforcement Learning ◽  
Autonomous Vehicles ◽  
Autonomous Vehicle ◽  
Vehicle Control ◽  
Data Sources ◽  
Simulation Environment ◽  
Urban Simulation ◽  
Policy Gradient ◽  
Almost All ◽  
Control Layer

AbstractNowadays, Artificial Intelligence (AI) is growing by leaps and bounds in almost all fields of technology, and Autonomous Vehicles (AV) research is one more of them. This paper proposes the using of algorithms based on Deep Learning (DL) in the control layer of an autonomous vehicle. More specifically, Deep Reinforcement Learning (DRL) algorithms such as Deep Q-Network (DQN) and Deep Deterministic Policy Gradient (DDPG) are implemented in order to compare results between them. The aim of this work is to obtain a trained model, applying a DRL algorithm, able of sending control commands to the vehicle to navigate properly and efficiently following a determined route. In addition, for each of the algorithms, several agents are presented as a solution, so that each of these agents uses different data sources to achieve the vehicle control commands. For this purpose, an open-source simulator such as CARLA is used, providing to the system with the ability to perform a multitude of tests without any risk into an hyper-realistic urban simulation environment, something that is unthinkable in the real world. The results obtained show that both DQN and DDPG reach the goal, but DDPG obtains a better performance. DDPG perfoms trajectories very similar to classic controller as LQR. In both cases RMSE is lower than 0.1m following trajectories with a range 180-700m. To conclude, some conclusions and future works are commented.

scholarly journals Deterministic Value-Policy Gradients

2020 ◽  
Vol 34 (04) ◽  
pp. 3316-3323
Author(s):  
Qingpeng Cai ◽  
Ling Pan ◽  
Pingzhong Tang
Keyword(s):  
Reinforcement Learning ◽  
State Of The Art ◽  
Learning Algorithms ◽  
Infinite Horizon ◽  
Gradient Algorithm ◽  
Continuous Control ◽  
Model Bias ◽  
Model Free ◽  
Policy Gradient ◽  
Analytical Gradients

Reinforcement learning algorithms such as the deep deterministic policy gradient algorithm (DDPG) has been widely used in continuous control tasks. However, the model-free DDPG algorithm suffers from high sample complexity. In this paper we consider the deterministic value gradients to improve the sample efficiency of deep reinforcement learning algorithms. Previous works consider deterministic value gradients with the finite horizon, but it is too myopic compared with infinite horizon. We firstly give a theoretical guarantee of the existence of the value gradients in this infinite setting. Based on this theoretical guarantee, we propose a class of the deterministic value gradient algorithm (DVG) with infinite horizon, and different rollout steps of the analytical gradients by the learned model trade off between the variance of the value gradients and the model bias. Furthermore, to better combine the model-based deterministic value gradient estimators with the model-free deterministic policy gradient estimator, we propose the deterministic value-policy gradient (DVPG) algorithm. We finally conduct extensive experiments comparing DVPG with state-of-the-art methods on several standard continuous control benchmarks. Results demonstrate that DVPG substantially outperforms other baselines.

Sign in / Sign up

Export Citation Format

Share Document