# Tag Info

9

There has been quite a few approaches to achieve such kind of distributed coordination. I present here one of them, for its generality and simplicity (that makes it easy to remember too). But first, the general idea behind these approaches is pretty interesting, around a mechanism called stigmergy. Stigmergy is a behaviour coordination mechanism mediated by ...

8

There are no distinguishable hardware examples for each IA class. The same mobile robot architecture with proper sensors can be implemented to behave as any IA class. The way you can determine the class of an intelligent agent is from the way it processes the percept. Based on chapter 2 of Artificial Intelligent: A Modern Approach I will try to give a ...

6

Swarm intelligence is the term for systems where relatively simple agents work together to solve a complicated problem in a decentralized fashion. In general, distributed computing methods are very important for dealing with problems at scale, and many of them embrace decentralization in a deep way. (Given the reality of hardware failure and the massive ...

3

The game of TIC-TAC-TOE can be modelled as a non-deterministic Markov decision process (MDP) if, and only if: The opponent is considered part of the environment. This is a reasonable approach when the goal is to solve playing against a specific opponent. The opponent is using a stochastic policy. Stochastic policies are a generalisation that include ...

3

Depends on perspective. On one hand, you have an agent playing in an environment with another agent also evolving. This falls under the definition of Multi-Agent Learning, as can be seen with works such as Michael Bowling and Manuela Veloso. Multiagent learning using a variable learning rate. Artificial Intelligence, 136(2):215 – 250, 2002. Michael Bowling....

3

Does it have to do with the reward function? This seems likely to me. You have chosen a reward that is unusual in that it cross-links episodes. It is not really a reinforcement problem to optimise behaviour with respect to results of previous episode behaviour in this way. This might be an option for an evolutionary fitness context, if you have competing ...

3

This is a highly relevant question as market trends have become more emphasized over the fundamentals of individual companies, and algorithmic trading has proven to be quite effective, particularly in areas such as high-frequency micro-trading. This 2013 Forbes article estimated nearly 80% of stock trading volume in the U.S. is conducted by automated ...

2

I think the best example of AI being deployed on the blockchain is SingularityNET. They just had a successful token sell where they sold out of their AGI token which will be able to be used to essentially "pay" for AI tasks to be done for you. Various AI will be put on the network and able to interact and communicate with each other to get various tasks done....

2

A couple of thoughts: Humans can't reliably predict trends in the stock market, so expecting AI's to do so is probably unreasonable. The above would be more true if it were proven that the movement of stock prices is really a random walk, but my understanding is that the current thinking is that stock movements aren't completely random... but just really ...

2

Not particularly sure what you are asking, so the question that I will be answering is this: Can Q learning be used to estimate a value that depends on another value in the Q Learning Matrix even though there is a certain amount of unpredictability involved? The answer is yes! I will use the example of a robotic arm trying to reach a point in space ...

2

A very efficient approach to what you are trying to do is velocity obstacles. Assuming two agents use constant velocity motion vectors, a velocity obstacle models a geometric region in which if the endpoint of the velocity vector for agent 1 falls, it will collide with agent 2 (and vice versa). Hence you can predict what velocity vectors will lead to ...

2

I think there is an intersection. There are problems that are in reinforcement learning and in learning in multi-agent systems. There are problems in reinforcement learning, but not exactly in multi-agent systems. And there is learning in multi-agent systems that is not through reinforcement learning. For sort you can say: multi-agent reinforcement learning. ...

2

it can be either. If you consider the lack of reward as "penalty" then getting 0 reward is bad. if you use a value estimator through a neural network, the range of rewards will dictate the squashing function you use for the output layer

2

The cleanest solution from a theoretical point of view is to switch over to a hierarchical framework, some framework that supports temporal abstraction. My favourite one is the options framework as formalised by Sutton, Precup and Singh. The basic idea is that the things that you consider "actions" for your agents become "options", which are "large actions" ...

2

Sounds like you have several problems with the way your policy is parametrized. You don't have to use the multivariate normal distribution. It can work, and probably others have done it already (if not with AAC, surely with DDPG, as it'll be easier to derive the policy gradient there). I won't explain how to use the multivariate normal with either case as ...

2

I am not an expert in RL. I have been playing Go for some years. Let's quote from AlphaZero's paper first: Aside from komi, the rules of Go are also invariant to colour transposition; this knowledge is exploited by representing the board from the perspective of the current player (see Neural network architecture). In the Game of Go, the difference between ...

2

Let's do the code, so all the details are down. Encoding dictionary: codes, i = {}, 0 for nSquares in range(1,8): for direction in ["N", "NE", "E", "SE", "S", "SW", "W", "NW"]: codes[(nSquares,direction)] = i i += 1 You'll see that the codes dictionary will ...

2

The SARSA update rule looks like: $$Q(S, A) \gets Q(S, A) + \alpha \left[ R + \gamma Q(S', A') \right].$$ Very similar, the $Q$-learning update rule looks like: $$Q(S, A) \gets Q(S, A) + \alpha \left[ R + \gamma \max_{A'} Q(S', A') \right].$$ Both of these update rules are formulated for single-agent Markov Decision Processes. Sometimes you can make them ...

1

I guess it depends on what the goal is. If the goal is a general reward function, this formulation as an MPOMDP could make sense. One way to think about this, is as a way of modeling a general (centralized) POMDP with factored actions and observation spaces. However, it seems that what you are describing might be an active perception problem, where the goal ...

1

This is mostly an implementation architecture problem, and the thing is that basically you can implement anything in the traditional setting. To do so instead of having Env<->Agent1<->Agent2, you should have Agent1<->SuperEnv<->Agent2 where SuperEnv contains Env, and simply uses the reward given to SuperEnv by Agent1 and passes it to ...

1

You might be able to glean what you want from Chapter 13 or Sutton & Barto's Reinforcement Learning: An Introduction, which deals with policy gradient algorithms, and includes pseudocode for a variety of agents based on linear approximation using softmax regression. From your description, you appear to be using - or should consider - softmax regression ...

1

I depends on your overall model architecture (and problem specification). As I understand it, you take the observations of all agents together and feed it into one model, a central controller, which then predicts the action per available agent. I believe that this varying number of applicable observations (depending on the number of currently present agents) ...

1

An agent is a concept, which can have slightly different meanings, abilities or instantiations depending on the context. However, given the purpose of this website, I will use and refer to the definition of agent commonly used in artificial intelligence. An agent is anything that can be viewed as perceiving its environment through sensors and acting upon ...

1

In a sense it seems like I am thinking some kind of "Deep Agent-Based Modeling", where it is okay to have network effects on the lowest layer (which would only evaluate the matching of the generated artifacts and their own preferences); the co-operation problem should be solved by a "Coach" layer, which would attempt to prevent situations, where some agents ...

1

There is quite some research done by Hans-Georg Zimmermann, who has programmed Neural Networks for Siemens since some 20 years in order to predict Stock markets. He wrote some books on it, too, though I don't know if they are any good in English. This article gets to the point a bit faster than the video, I hope it helps. edit: I think this interview gives ...

1

In the context of reinforcement learning, the idea of modeling your goal-oriented problem as a hierarchy of multiple sub-problems is called hierarchical reinforcement learning, which gives rise to concepts such as semi-Markov decision processes and options (aka macro actions). The article The Promise of Hierarchical Reinforcement Learning presents and ...

1

This evening I got inspired by this paper: http://www.dphrygian.com/bin/David-Pittman-GOAP-Masters-Thesis.doc (GOAP paired with the Command and Control Pattern) What do you think about this solution? Each goal has a relevance (that depends on the agent needs) When agent1 working on the "AskAgent2ToTalk" Action, it only sends a goal recommendation to ...

1

In my answer, I have often switched between AGI and ASI for reference. This is fine as an AGI will reach ASI as it is optimizing itself and learning. I think it is not only important by necessary that AGI and ASI are of collaborative nature. Nick Bostrom, in his book Superintelligence: Paths, Dangers, Strategies in Chapter 10 described three ways in which an ...

Only top voted, non community-wiki answers of a minimum length are eligible