Negative Rewards Research Articles

Penerapan Teori Belajar Behaviorisme dan Teori Belajar Sosial Bandura dalam Pembelajaran

This research focuses on the Application of Behaviorism Learning Theory and Bandura's Social Learning Theory in Learning. The purpose of this research is to reveal behaviorism learning theory and Albert Bandura's social learning theory and implement them in learning. The content of behaviorism learning theory is that learning is a change in behavior as a result of the interaction between the stimulus as a result of the interaction between stimulus and response. The content of social learning theory emphasizes direct experience or observation, because according to Albert Bandura, namely people learn from what they read, hear, and see in the media, and also from other people and their environment. environment. The method used in the research is a qualitative method that intends to understand the object of research with sophisticated efforts through the art of speaking, the art of writing and the art of presenting what has been researched. In the study there are results, namely social learning theory Social learning theory helps understand the occurrence of aggression and psychological deviations and how to modify behavior. And the result of applying behavioristic theory and Bandura's social theory is the formation of a behavior that is aggression. This is the formation of a desired behavior. The desired behavior gets positive reinforcement and inappropriate behavior gets negative rewards and learner behavior. The learning behavior of students is the result of the ability of students to interpret knowledge or information, interpret a model that is not appropriate for them. information, interpret a model that is imitated, then process cognitively and determine actions according to the desired goals. determine the action according to the desired goal.

Peer Incentive Reinforcement Learning for Cooperative Multiagent Games

Social learning, especially social incentives, is extremely important for humans to achieve a high level of coordination. Inspired by this, we introduce this concept into cooperative multi-agent reinforcement learning (MARL), to implicitly address the credit assignment problem and promote the inter-agent direct interactions for cooperations among agents in cooperative multi-agent games. In this paper, we propose a novel Intrinsic Reward method with Peer Incentives (IRPI) based on actor-critic policy gradient. This method can enable agents to incentivize each other for their cooperations through using causal influence among them. Specifically, a novel intrinsic reward mechanism is innovatively designed to empower each agent the ability to give positive or negative rewards to other peer agents' actions through considering the causal influence of the other agents on it. The mechanism is realized by a feed-forward neural network through utilizing causal influence between the agents. The causal influence of one agent on another is inferred via counterfactual reasoning using the joint action-value function in MARL. The quality of the influence is assessed via counterfactual reasoning using the individual value function in MARL. Simulations are carried out on two popular multi-agent game testbeds: Starcraft II Micromanagement and Multi-Agent Particle Environments. Simulational results demonstrate that the proposed IRPI can enhance cooperations among the agents to achieve better performance compared with a number of state-of-the-art MARL methods in a variety of cooperative multi-agent games.

Introduction of Reinforcement Learning and Its Application Across Different Domain

In the modern era of rapid development in Deep Neural Networks, Reinforcement Learning (RL) has evolved into a pivotal and transformative technology. RL, a learning process where these machine agent interacts with several unknown environment through trial and error. The agent, responsive to the learning machine, go through these interaction, and start receiving feedback in the form of positive rewards or negative rewards like penalties from the environment, and constantly refines its behavior. This research paper offers an in-depth introduction to the foundational concepts of RL, focusing on Markov Decision Processes and various RL algorithms.</p> <p>Machine Learning (ML) is a subset of Artificial Intelligence, which deals with ‘‘the question of how to develop software agents (Machine) that improve automatically with experience’’. The basic three categories of Machine Learning are.</p> <ol> <li>Supervised Learning</li> <li>Unsupervised Learning</li> <li>Reinforcement Learning</li> </ol> <p>RL method is that in any situation the agent has to choose between using its acquired knowledge of the environment i.e. using an action already tried or performed previously or exploring actions never tried before in that situation.</p> <p>In this review paper, we will discuss the most used learning algorithms in games robotics and healthcare, autonomous control as well as communication and networking, natural language processing.[1]</p>

Learning Potential in Subgoal-Based Reward Shaping

Human knowledge can reduce the number of iterations required to learn in reinforcement learning. Though the most common approach uses trajectories, it is difficult to acquire them in certain domains. Subgoals, which are intermediate states, have been studied instead of trajectories. Subgoal-based reward shaping is a reward-shaping framework with a sequence of subgoals based on potential-based reward shaping. The potential function which is a component of potential-based reward shaping has a hyperparameter that controls its output. However, it is not easy to select a hyperparameter because its appropriate value depends on the reward function of an environment, and the reward function is unknown. We propose <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">learned potential</i> that parameterizes a hyperparameter and acquires its potential through learning. A value is an expected accumulated reward if an agent follows its policy after the current state and is strongly related to the reward function. With learned potential, we build an abstract state space with a sequence of subgoals and use the value over the abstract states as the potential to accelerate the value learning. N-step TD learning learns the values over the abstract state. We conducted experiments to evaluate the effectiveness of learned potential, and the results indicate its effectiveness compared with a baseline reinforcement learning algorithm and several reward-shaping algorithms. The results also indicate that the participants’ subgoals are superior to subgoals generated randomly with learned potential. We discuss the appropriate number of subgoals for learned potential, that partially ordered subgoal is helpful for learned potential, that learned potential cannot make learning efficient in step penalized rewards, and that learned potential is superior to the non-learned potential in mixed positive and negative rewards.

An Integrated Model for Autonomous Speed and Lane Change Decision-Making Based on Deep Reinforcement Learning

The implementation of autonomous driving is inseparable from developing intelligent driving decision-making models, which are facing high scene complexity, poor decision-making coupling, and the inability to guarantee decision-making safety. This paper starts with the priority and logic of lane change and car-following decision-making, considering driving efficiency, safety, and comfort, then constructs a double-layer decision-making model. This paper uses two deep reinforcement learning algorithms for the upper and lower layers to process large-scale mixed state space and ensure the composite action output of lane-changing decisions and car-following decisions. In the upper layer model, we use the D3QN algorithm to distinguish the potential value of the environment and the value of selecting lane-changing actions when making lane-changing decisions. Different from the traditional mechanisms that only use negative rewards, the lane changing benefit function and dangerous action shielding mechanism are used to eliminate collisions. DDPG algorithm is adopted in the lower layer model to process car-following decisions and output continuous vehicle speed control. Besides, coupled training is taken for the two algorithms to improve the coordination of the double-layer model. This paper selected mixed standard driving cycle conditions to build a highly complex training environment and used NGSIM data to reconstruct scenes to test our model. Simulations in SUMO are presented that the double-layer model can increase the driving speed of the original data by 23.99%, which has higher effectiveness than other models.

Relationships in the Digital Age: Self-disclosure and Communication in Social Networking Sites

With the development of new technologies and rise of social media, the desire to communicate and share information is seen as a primary driver of social networking sites, where the aspect of self-disclosure becomes one of the key characteristics for interpersonal attraction. The aim of the research is to examine how people communicate, self-disclose and maintain relationships on social networking sites and to find out how the use of computer-mediated communication relates to relational development and maintenance. MTurk tool and social media posts were utilized to gather information. Statistical data analysis program SPSS was employed: confirmatory factor analysis (CFA) and scale analysis and construction were performed. The investigation revealed significant differences between males and females self-disclosure practices on social networking site Facebook with outstanding differences depending upon the country of living, relationship status, perceived positive and negative rewards, benefits and satisfaction. The results of the study open path for future research within the framework of communication privacy management theory in line with the social penetration theory with an aim to analyse how self-disclosure inter-correlates with privacy concerns. Keywords: relationship; computer-mediated communication; social network sites; penetration; self-disclosure.

Deviant Men at Work: A Social-Motivational Theory for Change

This paper proposes a motivational theory of deviance from gender norms that recognizes, understands and purposely develops working men´s communal behavior. Combining key streams of analysis that integrate organizational and psychological research, I provide an overarching approach about specific routes to facilitate change from men with testable propositions that inform future research. I explore the critical role of accommodating social and organizational patterns that reinforce men´s communal orientations. These contexts enable a psychological process of discrepancy reduction that allow identity-restructuring processes and facilitate gender deviance by changing meaning and importance of the masculine identity. These contexts also serve as motivational mechanisms that increase perceptions that gender-deviant efforts: (a) are feasible, (b) lead to a fair balance of positive and negative rewards and (c) are personally desirable. Evidence from research corroborates that explicit representations of deviant men and the value of communion at work are necessary steps to assist these psychological processes and allow men engage in fruitful forms of gender deviant behavior. This transformation is critical for both gender equality and organizational functioning.

Implementasi Teori Behavioris-Strukturalis dalam Pembelajaran Bahasa Arab

Many theories about learning have developed from the 19th century to the present. At the beginning of the 19th century learning theory that developed rapidly and made many contributions to psychologists was the theory of learning behavior (behaviorism). Behaviorism learning theory is oriented towards measurable and observable results. Repetition and training are used so that the desired behavior can become a habit. The expected result from the application of this behavioristic theory is the formation of the desired behavior. Desired behavior gets positive reinforcement, inappropriate behavior gets negative rewards, and evaluations or assessments are based on visible behavior.

Reward criteria impact on the performance of reinforcement learning agent for autonomous navigation

In reinforcement learning, an agent takes action at every time step (follows a policy) in an environment to maximize the expected cumulative reward. Therefore, the shaping of a reward function plays a crucial role in an agent’s learning. Designing an optimal reward function is not a trivial task. In this article, we propose a reward criterion using which we develop different reward functions. The reward criterion chosen is based on the percentage of positive and negative rewards received by an agent. This reward criteria further gives rise to three different classes, ‘Balanced Class,’ ‘Skewed Positive Class,’ and ‘Skewed Negative Class.’ We train a Deep Q-Network agent on a point-goal based navigation task using the different reward classes. We also compare the performance of the proposed classes with a benchmark class. Based on the experiments, the skewed negative class outperforms the benchmark class by achieving very less variance. On the other hand, the benchmark class converges relatively faster than the skewed negative class.

Ancillary mechanism for autonomous decision-making process in asymmetric confrontation: a view from Gomoku

ABSTRACT This paper investigates how agents learn and perform efficient strategies by trying different actions in asymmetric confrontation setting. Firstly, we use Gomoku as an example to analyse the causes and impacts of asymmetric confrontation: the first mover gains higher power than the second mover. We find that the first mover learns how to attack quickly while it is difficult for the second mover to learn how to defend since it cannot win the first mover and always receives negative rewards. As such, the game is stuck at a deadlock in which the first mover cannot make further advances to learn how to defend, and the second mover learns nothing. Secondly, we propose an ancillary mechanism (AM) to add two principles to the agent’s actions to overcome this difficulty. AM is a guidance for the agents to reduce the learning difficulty and to improve their behavioural quality. To the best of our knowledge, this is the first study to define asymmetric confrontation in reinforcement learning and propose approaches to tackle such problems. In the numerical tests, we first conduct a simple human vs AI experiment to calibrate the learning process in asymmetric confrontation. Then, an experiment of 15*15 Gomoku game by letting two agents (with AM and without AM) compete is applied to check the potential of AM. Results show that adding AM can make both the first and the second movers become stronger in almost the same amount of calculation.

Strategies used by Heads of Secondary Schools in Managing Students' Discipline for Academic Achievement in Ilemela District, Mwanza, Tanzania

The strategy to maintain student discipline in secondary school is a crucial which guide students for future self-sustainability and the right to learn since a well-disciplined student learn without any distraction. The study examined the strategies used by heads of secondary schools in managing indiscipline problems for students’ academic achievement in Ilemela District in Mwanza Region, Tanzania. The study used convergent parallel mixed research design and applied assertive discipline theory. Data were collected from 113 respondents using questionnaires and interview guide. Qualitative data was analyzed thematically and quantitative data were analyzed through descriptive statistic with the help of the SPSS Version 20, IBM, USA. The study revealed that the strategies used to curb the indiscipline cases include the use of guidance and counseling, cooperation between teachers and parents, provision of positive and negative rewards, and rules and regulations have to be enacted. The study recommended that teachers should have good relationship with both parents and students, parents should play their parenting roles effectively.

Stochastic mixed model sequencing with multiple stations using reinforcement learning and probability quantiles

In this study, we propose a reinforcement learning (RL) approach for minimizing the number of work overload situations in the mixed model sequencing (MMS) problem with stochastic processing times. The learning environment simulates stochastic processing times and penalizes work overloads with negative rewards. To account for the stochastic component of the problem, we implement a state representation that specifies whether work overloads will occur if the processing times are equal to their respective 25%, 50%, and 75% probability quantiles. Thereby, the RL agent is guided toward minimizing the number of overload situations while being provided with statistical information about how fluctuations in processing times affect the solution quality. To the best of our knowledge, this study is the first to consider the stochastic problem variation with a minimization of overload situations.

A Study on an Enhanced Autonomous Driving Simulation Model Based on Reinforcement Learning Using a Collision Prevention Model

This paper set out to revise and improve existing autonomous driving models using reinforcement learning, thus proposing a reinforced autonomous driving prediction model. The paper conducted training for a reinforcement learning model using DQN, a reinforcement learning algorithm. The main aim of this paper was to reduce the time spent on training and improve self-driving performance. Rewards for reinforcement learning agents were developed to mimic human driving behavior as much as possible. High rewards were given for greater distance travelled within lanes and higher speed. Negative rewards were given when a vehicle crossed into other lanes or had a collision. Performance evaluation was carried out in urban environments without pedestrians. The performance test results show that the model with the collision prevention model exhibited faster performance improvement within the same time compared to when the model was not applied. However, vulnerabilities to factors such as pedestrians and vehicles approaching from the side were not addressed, and the lack of stability in the definition of compensation functions and limitations with respect to the excessive use of memory were shown.

Mood and Risk-Taking as Momentum for Creativity.

This study examined the effects of mood and risk-taking on divergent and convergent thinking using a Q-learning computation model. The results revealed that while mood was not significantly related to divergent or convergent thinking (as creative thinking types), risk-taking exerted positive effects on divergent thinking in the face of negative rewards. The results were consistent with the representational change theory in insight problem solving. Although this theory accounts directly for insight, the underlying idea of going beyond current contexts and implicit constrains could be applied to creative thinking as well. The results indeed accounted for the relevance of this theory to divergent thinking. The current study is one of the first empirical studies simultaneously examining the role of mood and risk-taking in creativity. In particular, no related studies exist that took a computational approach to estimate the relevant parameters in the framework of dynamic optimization. Our Q learning model enables to distinguish and identify the different roles of mood and risk-taking in updating Q values and making decisions.

On the design of managerial incentives for sustainability investments in the presence of competitors

In many industries, an increasing number of firm owners tie managers’ incentives to sustainability investments. Positive rewards directly increase a manager’s total pay when that manager makes sustainability investments, whereas negative rewards directly decrease a manager’s pay when those investments are made. Strategic incentive design literature posits that such organizational choices also affect the decisions of a firm’s competitors. This paper uses a game-theoretic framework to analyze the effects of sustainability incentives in a setting with two competing firms. In contrast to the existing literature, in the current paper sustainability investments have a demand-enhancing effect and can increase or decrease the unit cost of production, making the current framework more in line with industrial practice. The results show that a firm invests in sustainability only if the demand-enhancing effects outweigh the cost-increasing effects. More importantly, positively rewarding managers for sustainability investments is done in equilibrium only if the innovation capability of the firm is sufficiently high. However, in terms of profits, those positive rewards lead to a prisoner’s dilemma. When innovation capability is lower, firm owners use negative rewards and raise their profits. Another finding is that rival firms that cooperate in determining their sustainability incentives increase their profits but do so using negative rewards. These results, which have not been reported in the literature, point to some critical trade-offs in terms of sustainability investments and firm profits when sustainability incentives are considered and are both managerially and academically relevant.

Memory retention of appetitive and aversive conditioning in the damselfish Stegastes fuscus

We compared the memory of damselfish Stegastes fuscus in an aversive and appetitive conditioning task. Fish were trained to associate the sides of the tank that corresponded to the presence of a positive (conspecific presence) or negative (electroshock) stimulus. After two conditioning sessions, they were tested for learning. The fish conditioned to the stimulus were then re-tested for memory retention after 5, 10 or 15 days. Both the positive and negative rewards were associated with a specific side of the tank, indicating learning ability. Additionally, in both contexts, S. fuscus stored the information learned and showed similar behavioural patterns after 5, 10 and 15 days, suggesting long-lasting memory. For the ecological context, long lasting memories of social encounters outcomes and negative experiences of threatening situations may confer advantages that ultimately affect fishes' fitness.

Reward Processing in Drug-Naive Youth with Various Levels of Risk for Substance Use Disorders: A Pilot Study.

Background: It has been hypothesized that attention-deficit/hyperactivity disorder (ADHD) and substance use disorders (SUDs) share common neurobiological features. When abnormalities in the mesolimbic reward system are characteristic of children with ADHD, it is unclear whether youth at particularly elevated risk for SUD may exhibit any further disturbances. The objective of this pilot study is to examine possible neurobiological differences among youth with different levels of SUD risk before exposure to any abusable substances. Methods: We recruited 47 drug-naive children aged 8-13 divided into (1) Low Risk (LR)-ADHD only (n = 16); (2) High Risk (HR)-ADHD+familial SUD (n = 17); and (3) healthy controls (HC, n = 14) who underwent one functional magnetic resonance imaging scan while performing a hybrid task. We used the omnibus analysis of covariance model to assess for group differences in brain activation in regions linked to the brain reward and behavioral control systems. Results: Behavioral analysis showed significant Cue and Flanker main effects, but no significant main effect for Group. Whole-brain analysis showed significant differences in widely distributed networks related to both reward processing and behavioral control. Region of interest (ROI) activations showed that the HR group had the highest activation in the right putamen during both expected rewards and unexpected nonreward outcomes and in the anterior cingulate cortex (ACC) during unexpected nonreward outcomes, while LR and HC youth showed similarly low activation during these contrasts. Furthermore, the LR and HR groups showed lower activation than HC in the right ACC and the right caudate during flanker contrasts. Conclusions: These are the first preliminary results to demonstrate that the magnitude of activation during reward notification differs as a function of reward outcome in youth at high versus LR for SUD, such that youth at LR for SUD exhibit the highest activation for positive rewards, whereas those at HR for SUD exhibit the highest activation during negative rewards.

Application of Q-learning based on adaptive greedy considering negative rewards in football match system

Aiming at the problem that the multi-robot task allocation method in soccer system is easy to fall into the problem of local optimal solution and real-time performance, a new multi-robot task allocation method is proposed. First of all, in order to improve the speed and efficiency of finding optimal actions and make better use of the disadvantages that traditional Q-learning can't often propagate negative values, we propose a new way to propagate negative values, that is, Q-learning methods based on negative rewards. Next, in order to adapt to the dynamic external environment, an adaptive ε greedy method of which the mode of operation is judged by the ε value is proposed. This method is based on the classical ε-greedy. In the process of solving problems, ε can be adaptively changed as needed for a better balance of exploration and exploitation in reinforcement learning. Finally, we apply this method to the robot's football game system. It has been experimentally proven that dangerous actions can be avoided effectively by the Q-learning method which can spread negative rewards. The adaptive ε-greedy strategy can be used to adapt to the external environment better and faster so as to improve the speed of convergence.

Application of Q-learning based on adaptive greedy considering negative rewards in football match system

Aiming at the problem that the multi-robot task allocation method in soccer system is easy to fall into the problem of local optimal solution and real-time performance, a new multi-robot task allocation method is proposed. First of all, in order to improve the speed and efficiency of finding optimal actions and make better use of the disadvantages that traditional Q-learning can't often propagate negative values, we propose a new way to propagate negative values, that is, Q-learning methods based on negative rewards. Next, in order to adapt to the dynamic external environment, an adaptive e greedy method of which the mode of operation is judged by the e value is proposed. This method is based on the classical e-greedy. In the process of solving problems, e can be adaptively changed as needed for a better balance of exploration and exploitation in reinforcement learning. Finally, we apply this method to the robot's football game system. It has been experimentally proven that dangerous actions can be avoided effectively by the Q-learning method which can spread negative rewards. The adaptive e-greedy strategy can be used to adapt to the external environment better and faster so as to improve the speed of convergence.

Finding efficient swimming strategies in a three-dimensional chaotic flow by reinforcement learning.

We apply a reinforcement learning algorithm to show how smart particles can learn approximately optimal strategies to navigate in complex flows. In this paper we consider microswimmers in a paradigmatic three-dimensional case given by a stationary superposition of two Arnold-Beltrami-Childress flows with chaotic advection along streamlines. In such a flow, we study the evolution of point-like particles which can decide in which direction to swim, while keeping the velocity amplitude constant. We show that it is sufficient to endow the swimmers with a very restricted set of actions (six fixed swimming directions in our case) to have enough freedom to find efficient strategies to move upward and escape local fluid traps. The key ingredient is the learning-from-experience structure of the algorithm, which assigns positive or negative rewards depending on whether the taken action is, or is not, profitable for the predetermined goal in the long-term horizon. This is another example supporting the efficiency of the reinforcement learning approach to learn how to accomplish difficult tasks in complex fluid environments.