Bacterial Prisoner's Dilemma and Game Theory

dumuzi writes "Scientists studying how bacteria under stress collectively weigh and initiate different survival strategies say they have gained new insights into how humans make strategic decisions that affect their health, wealth and the fate of others in society. The authors of the new study are theoretical physicists and chemists at the University of California, San Diego's Center for Theoretical Biological Physics. In nature, bacteria live in large colonies whose numbers may reach up to 100 times the number of people on earth. Many bacteria respond to extreme stress — such as starvation, poisoning and irradiation — by creating spores. Alternately the bacteria may 'choose' to enter a state called competence where they are able to absorb the nutrients from their newly deceased comrades. 'Each bacterium in the colony communicates via chemical messages and performs a sophisticated decision making process using a specialized network of genes and proteins. Modeling this complex interplay of genes and proteins by the bacteria enabled the scientists to assess the pros and cons of different choices in game theory. It pays for the individual cell to take the risk and escape into competence only if it notices that the majority of the cells decide to sporulate,' explained Onuchic. 'But if this is the case, it should not take this chance because most of the other cells might reach the same conclusion and escape from sporulation.'"

Comments

Bah. nothing new

[Anonymous Coward]

"Alternately the bacteria may 'choose' to enter a state called competence where they are able to absorb the nutrients from their newly deceased comrades."
The vultures on Wallstreet do this all the time.

Re:Bah. nothing new

[Calydor]

I'm pretty sure none of the Wall Street vultures have EVER entered a state called 'competence'.

Thinking Bacteria

[icebike]

Each bacterium in the colony communicates via chemical messages and performs a sophisticated decision making process...

I'm sorry, but that stretches the meaning of "sophisticated" and "decision" beyond all reason.

One might just as well argue that water flowing down hill has made a sophisticated decision.

Re:Thinking Bacteria

[selven]

A rock is a computer 10 billion times more powerful than all of our computers on the planet combined. The hard problem is putting that power to good use,

Re:Thinking Bacteria

[fuzzyfuzzyfungus]

Obviously the bacteria aren't "thinking" in any way that could usefully be crammed into the usual definition of the word; but I don't think that either "sophisticated" or "decision" are being distorted at all.

It is quite common, for instance, to refer machines that have a fair number of parts and are good at what they do as "sophisticated"("a sophisticated inertial navigation mechanism"). Even unicellular procaryotes have a fair amount going on inside, so they could easily fall under this definition.

As for "decision", that certainly can imply a process of rational, reflective cogitation; but it is also quite commonly applied to fairly simple, entirely mechanistic, things. "Decision Algorithms", for instance, are explicitly designed to be mechanistic and, as their name suggests, make decisions. The idea that the process whereby a cell enters either stateA or stateB depending on certain inputs is a "decision process" seems wholly reasonable to me.

Re:Thinking Bacteria

[Anonymous Coward]

> A rock is a computer 10 billion times more powerful than all of our computers on the planet combined.

Is that including the other rocks?

Re:Thinking Bacteria

[BigSlowTarget]

Bingo, there is some major over generalization going on in this article. The chemical reactions of bacteria to a chemical threat, even honed by millions of years of evolution, are not directly comparable to human reactions to information or threat. Even with billions of members a colony of bacteria has less chemical and informational content than a much smaller number of humans.

"Everyone knows the need to try to postpone important decisions until the last moment but apparently there are simple creatures that do it well and therefore can really teach us -- the bacteria," Really? And if postponing the decision has an impact on the possibility of implementing the selected solution? When a politician delays making a decision he can appear weak and indecisive which is certainly not a benefit - IF he has the data and can make the correct decision earlier. Similarly delaying one decision can have a direct impact on later decisions even when you don't know what those decisions are.

In defense of the article the true value could be in the calculations for weighing the probability of the optimum solution given perfect information that are derived from the bacteria. - a situation never to occur in human history but useful for reference and as a base for future theory.

Re:Thinking Bacteria

[hazem]

One might just as well argue that water flowing down hill has made a sophisticated decision.

Actually, the behaviors and communication of groups of bacteria are much more complex than water flowing downhill. Consider that when you get a bacterial infection, the bacteria will typically work in a "growth phase" where they are multiplying but not doing being virulent. When the bacteria reach a certain population size (or density), they all switch on their virulence. Individuals are making decisions that actually manifest as a group decision. Water molecules do not do this.

A very interesting lecture on this is at:
http://www.ted.com/talks/bonnie_bassler_on_how_bacteria_communicate.html

Re:Thinking Bacteria

[Anonymous Coward]

> A rock is a computer 10 billion times more powerful than all of our computers on the planet combined.

Is that including the other rocks?

Yes. It's only that particular rock that's so powerful. All of the rest of them are dumb as, well, rocks.

Re:Thinking Bacteria

[TapeCutter]

The water had no choice, gravity decided for it. Besides the human brian is basically a colony of single celled automata that communicate via chemical messages and perform a sophisticated decision making process (well sometimes anyway).

Re:Thinking Bacteria

[TapeCutter]

Please explain how the chemical process in your head do not lead to your decision that there is no decision.

Re:Thinking Bacteria

[Cassius Corodes]

Why do you need intelligence to do decision-making? This seems an arbitrary requirement - and the literature on decision-making doesn't use intelligence as a prerequisite, rather it refers to cognition - which is more rigorously defined and can be automated.

Your example of grass and dogs seems quite confused - grass lacks the ability to move in real time so the notion of it being unable to choose to lay down seems bizarre, unless I am missing something. According to this one could argue that a failure of a dog to teleport out of the way of a bus shows their inability to make decisions! Likewise chemicals are following externally set rules in physics, there is no internal selection of outcome.

Perhaps the difference in our thinking relies on the definition of decision-making. The one I use is the one used in most decision-making literature, which roughly states is "a cognitive process of selecting among several different options based on external and internal factors." According to this the bacteria is engaging in decision making as it is selecting for several different courses of action. Its use of cognition is pretty basic, but more that enough to satisfy that criteria - its processing information and applying internal biases to select a course of action.

Also your statement about bacteria being sacks of chemicals and micro structures and its reaction is just as applicable to humans and dogs as to bacteria. Remember that we are made up of individual cells too.

Re:Thinking Bacteria

[DriedClexler]

Not true. Once, I introduced a colony of water molecules onto a table. As is typical, they work in a "spread phase" where increase the area-to-water-stack-height ratio. Once they've detected the edge of the table, they begin "burrow mode" and start propagating a message for other water molecules to replaced the ones that started burrowing.

strategy sounds familiar

[v1]

a case of where the individuals are all trying to make decisions that are selfish, but if everyone is selfish, no one wins, so some have to be selfish and some have to fold, for any to survive. I seem to remember playing games like that as a kid, where it was basically a game of chicken, where no one could do anything until everyone was generous, and so everyone then starts building up, and whoever managed to switch back to greedy first won. Also reminiscent of the stock market during a bubble, eh?

Re:strategy sounds familiar

[TubeSteak]

I seem to remember playing games like that as a kid, where it was basically a game of chicken, where no one could do anything until everyone was generous, and so everyone then starts building up, and whoever managed to switch back to greedy first won. Also reminiscent of the stock market during a bubble, eh?

Stock markets do not allow for equal access to information.
That inequality seriously skews any game theory in favor of the well connected.

Re:High Scientific Goals

[Sulphur]

I for one welcome our new cell automaton overlords.

Re:Pfft... predicting social behaviour...

[Wonko the Sane]

First, assume the world's population is an ideal gas in a frictionless vacuum...

So the population of the world = 0? No wonder it's frictionless.

Re:Bacteria analogous to human beings?

[wellingj]

I don't think we will like the implications if they say yes.

James Cagney as a Bacterium...

[theNAM666]

Now look here guys, see, I'll spore as soon as each of you spore, but if any one of you display any signs of competence, it's...

Good Bacteria interaction overview (good watching)

[MonsterMasher]

    Very strange - I just finished watching this lecture video this morning. I've all so seen her talk in TED.com

      http://microbeworld.libsyn.com/index.php?post_id=516458&utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+asm+(MicrobeWorld+Video)#

Cool Stuff!

Re:Bacteria analogous to human beings?

[easyTree]

I'm more interested in how various strategies used by scientists when making the "latest wild claim" (tm) affects their level of success within the game of scientist-gene evolution.

Re:Bacteria analogous to human beings?

[linhares]

human population growth is much more bacteria-like than primate-like.

Re:High Scientific Goals

[JoshuaZ]

Conway's Game of Life wasn't made to simulate life in any meaningful sense. It was designed by Conway because he was investigating simple cellular automata that had non-trivial behavior. http://en.wikipedia.org/wiki/Conway's_Game_of_Life. It happened that the simplest interesting form he found happened to have rules that could be stated with very very rough analogs to living creatures. Some of the rules are very much stretches. For example, while bacteria can die from overcrowding, they cannot die from being lonely. And cells aren't reincarnated or made new from having three neighboring cells (I'm not aware of any species outside science fiction that requires more than two cooperating members to have sex (see for example Asimov's "The Gods Themselves")).

Great Program, Wrong Channel

[DynaSoar]

They've done a bang up job investigating how bacteria adapt, and from the names and departments listed, I can see how they'd be quite able to do so as well as apply it to an expanded game theory scenario.

But applying it to human decision making, strategic or otherwise? Sorry, but they should have included someone on the team from behavioral science that could have pointed out the glaring differences.

They happen on one themselves in saying the bacteria don't lie. The level of stress they're talking about is equivalent to massive drought/starvation. Humans under such conditions do and say all kinds of things, most of it to some degree hiding real intentions.

To extend that, some of human behavior is rational under normal conditions, some isn't (emotionally driven isn't, for instance). With increased stress, less and less is rational. Their very nicely done description of possible decisions at various points based on DNA is entirely rational throughout. Not that the bacteria think, but that the decision is predetermined by being programmed in. There is no irrational result, no off-the-wall craziness drastic behavior resulting in novel solutions. Humans do this. In fact, novel results is a major difference between their work and pretty much any higher organism.

I don't find it particularly instructive that bacteria put off "decision making" until the last moment. As if people don't? It's human nature to constantly refine decisions according to the situation, including attempting top adapt to the situation after a decision has been implemented and the crucial point passed.

The final point they make, where one has to decide based on best guess of others' future behavior, is fairly telling of a major difference between bacteria and humans. Humans can coordinate their decisions so that none obtain an optimal result but all obtain a satisfactory result. That flies in the face of traditional game and economic theory. It also earned John Nash a Nobel. Bacteria can't discuss with predictive insight, they can only wait until the last moment to react.

NOT a prisoner's dilemma

[Main Gauche]

Why is it that anyone who's learned the slightest bit of game theory suddenly thinks everything is a Prisoner's Dilemma?

In a (1-shot) Prisoner's Dilemma, one action is always better for you than another, leaving little to analyze.

In the Bacteria's game, the bacteria are obviously programmed to do what is best to ensure the survival of the species. (FTFA: "bacteria usually do not cheat their friends and inform them by sending chemical messages about their true intensions.") Whether a bacterium should spore or not depends on the proportion of other bacteria doing each action. This is not the structure of a P.D. It's one thing for journalists to make a bad reference, but the physicist himself refers to Prisoner's Dilemma.

Re:NOT a prisoner's dilemma

[zarzu]

Actually, that's not really 100% either. In a single play of a prisoner's dilemma, you still don't know what's best because you don't know what your opponent is going to do; you can only hope that he's going to hold his tongue, but since he won't, you'll both rat each other out no matter what.

no, the whole point of it is that every player has one dominating strategy, meaning no matter what the opponent does, this one strategy is always the best. what your opponent does changes your actual win, but in a one-shot PD it never influences your choice.

Re:Bacteria analogous to human beings?

[Cassius Corodes]

They are - and for good reason. Game theory has been very successful in understanding some of the basic trade-offs involved in individual vs group decision-making. Certain set-ups such as the http://en.wikipedia.org/wiki/Prisoners_dilemma are generic forms of common problems that are encountered both in the human world and the natural world. Having worked in this area I can tell you that solutions found in the natural world often end up as inspiration for real life applications - such as regulation of industry and organisational psychology. At the end of the day one of the most re-occurring problems is how to get selfish people to co-operate as a group - and this problem has been solved so many times by nature in so many ways its basically a handy repository of tried and true solutions just waiting to be discovered.