Autonomous agents are overtaking the work force of the future.What are they exactly ?  A quick guide for beginners .



Before we can even begin to understand what an autonomous agent is we need to understand what an Agent is . 

An agent is simply a person or a program which performs certain tasks to achieve it’s goals in the most optimal way possible

We as humans often times get distracted and are bound to our limbic system needs , an autonomous agent does not have such limitations and therefore are excellent agents and can be applied to any industry to achieve most optimal results all the time !!


An Autonomous Agent is a computer program that has been designed to carry out tasks autonomously with no human intervention. Thus, an autonomous agent is free from any direct human control, and is able to react creatively and independently within its own environment. Autonomous agents have In-built attributes such as utilities , actions , skills . It can learn skills autonomously without owners approval to gain a tactical advantage in completing it’s task .





If you are thinking it seems somewhat like an AI you are right to think that upto some extent ,Autonomous Agent = task completion and AI = problem-solving. Main difference being Autonomous agents are designed for use in predictable environments to complete tasks within a specific, usually pre-planned, environment whereas an AI could be used in any environment .



Autonomous agents also have a special feature which makes them really worth talking about , having the ability to use deception . So what is deception ? Why do these agents even need to use deception ?

Deception is the act of taking advantage of a situation by utilising the uncertainty created due to absence of complete information therefore the situation can be moulded to gain an upper hand during negotiation . Humans often use deception to raise the probability of success when negotiating. For example – Consider you are going out to buy groceries and you do not know the price of the vegetables , Now this information can be used by the salesman to ask you for a higher price point than what might be an acceptable one . On the other hand , If you know the exact price you can use that to negotiate and thus pay a lower price . This is how information leads to handling with deception and absence of it creates the opportunity for it .  





And why do these agents even need to use Deception ? This is because agents usually work in a multi agent environment and are mostly one of these two –

  • competitive system  
  • cooperative system

A competitive system would require agents to maximize their utility even if it is to the peril of it’s competitor whereas a cooperative system requires maximizing the utility of the overall system / environment .


As per the wikipedia:
A multi-agent system (MAS or “self-organized system”) is a computerized system composed of multiple interacting intelligent agents. Multi-agent systems can solve problems that are difficult or impossible for an individual agent or a monolithic system to solve. Intelligence may include methodic, functional, procedural approaches, algorithmic search or reinforcement learning. In a MAS we can oversee how each agent is doing and which agents are producing the best results and thus we can replicate those same actions and optimize our process even more !!  


In a multi-agent competitive scenario, agents usually use the following deception mechanisms :-
i.hidden utilities 
ii. hidden actions, 
iii. decoy actions.

Hidden utility mechanisms are often used by agents such that its utility regarding their actions is not shared . If an agent does not share its utility associated with each action, the other will be picking actions (at least initially) grounded solely in it’s own utility. In context to a similar situation , hidden and decoy actions are often used to show that the agent can perform a certain action while they cannot or vice versa in order to not let the competitor gain an intel advantage over them and be able to use deception to increase utility . 

Similarly , in contexts where agents want to minimise the overall cost (maximise the overall utility), typically it does not make sense to use deception mechanisms, as they comprise extra difficulties on the task. In such scenario agents work cooperatively to gain the maximum possible utility .

WHERE CAN WE USE AUTONOMOUS AGENTS ??

So , now that we have learned what autonomous agents are and how do they work, where exactly  can we exactly use these agents in a real world scenario ? Well , one can say that we can use these agents to relieve humans from any type of mundane routine computational tasks !!  Tasks such as data mining or even buying stuff from the internet at the cheapest price possible !! We can also program them such that they act as our personal bots which can then do anything ranging from playing games against us ,personalising news according to our own palette and sorting e-mail according to our priorities . The possibilities are endless !!

Another great example is what fetch.ai is developing – 

“ Autonomous Economic Agent “ . These are autonomous agents who are mainly responsible for economic gain for it’s owner . It can be used in any industry ranging from supply chain to executing micro transaction trades for maximum profit of the owner . AEA can also connect to the blockchain to settle transactions . We will write an extensive blog on AEA as well very soon so stay tuned for that !!

CONCLUSION 

 
So , Autonomous agents are self learning programs which act in a specific domain to produce the most optimal results , but they do have limitations as well , like they are really difficult to set up and you need a really high level of understanding to even begin to utilize them . Although we are making great strides in the direction of making it more and more mainstream like what fetch.ai is doing, it won’t be too far in the future where we will see Autonomous agents doing everything that we can and much more !!


Now that you’ve read this post, you know exactly what an autonomous agent  is and what are they used for. Stay tuned for Blog about Autonomous economic agent COMING SOON !!


References :

Fetch.ai
https://towardsdatascience.com/autonomous-agents-and-multi-agent-systems-101-agents-and-deception-f4da3401f92a


Written by: Ashwini kumar at Hashtrust Technologies

That’s it for now! I hope you find the article helpful. Thanks for reading.

Leave a Comment

Your email address will not be published.