Rahul Dandekar

Non-equilibrium statistical physics: fluctuating hydrodynamics, self-organized criticality

[Rahul Dandekar @rahuldphys]

Looking for a 1BHK in North Bangalore: Hebbal, Nagawara, RT Nagar, and other areas along Manyata and along outer ring road. Move in: July onwards.

Please reply if you have any pointers and RT for reach!

@BangaloreRoomi
@fmrbangalore
@MIBangalore
@flatshareblr
@GruhamBot

[Rahul Dandekar @rahuldphys]

@plain_simon Hmm, I will have to think about that.

[Rahul Dandekar @rahuldphys]

@plain_simon Are you looking for a model which would not be SOC without disorder but is SOC with disorder?

[Rahul Dandekar @rahuldphys]

@plain_simon Oh! I see what you meant now by "pushing" the system towards a critical point. There's so many models of SOC. I'm most familiar with sandpile type models: some have disorder built in (Oslo model), some have been studied after adding extra disorder. Eg. https://t.co/Y7Fu2QubGG

[Rahul Dandekar @rahuldphys]

@plain_simon In percolation the structure of the local noise changes the critical point, eg bond vs site percolation, but not the universality class. To change that, there will be scaling relations for how long ranged or fat tailed one needs to be, eg here https://t.co/BjcUX0sLmD

[Rahul Dandekar @rahuldphys]

@plain_simon I'd say percolation theory is the cleanest place to look for effects of correlated noise on phase transitions. Directed polymers perhaps a good place to look for effects of Levy noise.

[Rahul Dandekar @rahuldphys]

@plain_simon Do you mean there's a phase transition that wasn't there before, or the critical point is shifted, or the character of the phase transition is changed? Generally it'd need to be long ranged correlated or local but levy distributed noise to change the character of the transition.

[Rahul Dandekar @rahuldphys]

@drmtgr Indeed. I feel it's Onsager's regression hypothesis. Then it's natural that the "entropy" of an atypical evolving state is the KL divergence from the equil measure, since the log-probability of seeing an atypical state is proportional to the KL divergence from the typical state

[Rahul Dandekar @rahuldphys]

My reply to this question seems to be put under 'probable spam'. Great feature, Elon.

Anyway, so I'm gonna post it here. 🧵

[François Fleuret @francoisfleuret]

Since my "questions" are generally well received: What is the relation between physics' entropy and information theory entropy?

No hand wavy explanation, I want something clear and exact.

[Rahul Dandekar @rahuldphys]

Similarly, you can extract work from a thermal system if you measure its current state (gaining info about it), but that you haven't lowered its temperature! What you've done is put it in a non-equilibrium state where you can catch it before it thermalises again.

[Rahul Dandekar @rahuldphys]

Just to clarify, "they have the same formula" != "they are the same thing". For one thing, a tape whose information you don't know isn't at a higher temperature than one whose info you know. So information can be an additional source of work than entropy.

[Rahul Dandekar @rahuldphys]

@francoisfleuret This seems to be showing as 'probable spam'. Brilliant.

[Rahul Dandekar @rahuldphys]

@francoisfleuret How entropy generalises to non-equilibrium situations and whether this can be meaningfully related to information depends on how useful information is in a nonequilibrium situation. This is not a priori known and can change from system to system.