humans are now heading towards building trust on Robots

Article Edited by | Jhon N |


The building of confidence in robots by explaining what they are doing is an important step to the cooperation between people and robots.

On a normal day, people often meet Artificial Intelligence ( AI). Artificial intelligence has in many ways become an integral part of human routine. In smartphone, home appliances and technology, they come into our lives in our cars. They help people in various ways, from appointments to diagnose disease.

Given that people are on the brink of acceptance of robotics into society, the issue is: Can robots be trusted? 'Human beings have a mythical example of robots becoming aggressive as soon as they have all the features of human beings. We are pushed through films, soap operas and dramas to such conclusions. To keep away all these perplexing thoughts, we should focus on how the relationship between humans and robots can reach the height of trust.

Explainable artificial intelligence is a branch that studies how an artificial agent may be represented in a more transparent and reliable way. Building and leading human-robot relationships in the technology world is very important. When people start to work with robots, it is a mandatory move. Robots are trying to build systems which attract human attention while performing well to accomplish this task.

A team at the University of California 's Center for Vision, Cognition , Learning and Autonomy is researching the factors that make machines more reliable. It also looks at various algorithms that build human trust. The laboratory utilises a model representation of knowledge that can interpret the environment and make decisions easily understood by people. This obviously enhances human confidence by citing clear explanation and transparency.

Recent research has shown that a robot can explain its action to a human observer in a variety of ways that make people feel more confident about robots. Strange, the way in which confidence improved was explained does not correspond to the best performing learning algorithms. The performance and explaining of the robot is not proportional. But it is not a good idea to optimise one of them. The team thus concentrates on building robots that take good job performance and reliable explanation into account.

To educate a robot to perform a task: The team conducted a study to learn how robots perform specific tasks and how people react to robot explanation.

The team has taught a robot how to open a medicine flask with a safety lock from human demonstrations. A person was wearing a tactile glove, showing poses, movements, and forces involved in opening a bottle to make the robot scan and understand the movements. The data was transmitted to the robot. Two ways were understood by the robot: symbolic and haptic. Symbolic understanding means understanding of robot 's image in human action. During the moment, haptic has to do with body posture and motions.

The robot learned a symbolic model as a first step, encrypting the string of steps to complete the task of starting the flask. The robot then learned a hectic model with the robot "imagination" itself as a human protester. This step encourages the robot to predict the action a person would undertake when a bottle is opened. The robot also analyses the positions and the strength to complete the task.

Of note, with the combination of the symbolic and the haptic process, the robot could achieve its best performance. The robot could use its recorded knowledge with symbolic representation in order to perform the tasks and to sensor in real time, which the robot did by predicting the outcome.

Get human confidence by explaining: The robot now knows what action it needs to take to perform a task. The next step would be to explain what it did to people to clarify and build confidence.

The robot can draw on its decisions and behavior internally. The step by step description of robotic action by means of symbolic models and the feelings and predictions by means of the haptic model. The team allowed the provision to make the robot write a text on its actions to make the experiment more vivid and clear. You wanted to know whether the summary text works as well as the rest.

It consists of 150 participants from all four groups to watch the robot open the medicine bottle. This program includes 150 people. The types of explanation were divided by giving the groups of individuals symbolic, step by step, haptic expression of the arms position and motion, summary text or symbolic and haptic together. The robot opened the medicine bottle without any explication was seen by one of the groups.

The result was that people who learn the robot to explain the haptic and symbolic process have more confidence in it. The other group, which was explained by a text summary, felt that the robot was doing its job. Both of these processes seemed to foster little human confidence.

Take away for the future. Research has shown that people do not trust robots that only act without explanation. The robots therefore need symbolic as well as haptic components to calm people.

The research has provided an insight into the future of artificial intelligence and encourages research that focuses equally on robots and their ability to explain their actions. Performance and explanation are both necessary for the construction of an artificial intelligence system. To move it further, it should be possible to shape the relationship between man and machine.

It is not a new idea to make robots part of daily life. But people must put their faith in robots to accomplish the move. Confident robots are an important step towards enabling people and robots to a further degree by explaining their actions.