RFP Overview    |    FAQ    |    Examples    |    Application Form    |    Timeline

Bookmark and Share

Information as Fuel
An International Request for Proposals

Topics that might be addressed by successful proposals

The following are examples of appropriate topics for research under this program. This is a non-exclusive list but should serve to indicate the types of work the program aims to support. It should also be noted that simply addressing information in general does not make an project germane to this program.

  • Autonomous Maxwell demons
    Maxwell’s famous thought experiment involved a hypothetical being who would sort fast-moving from slow-moving molecules. Lord Kelvin subsequently designated the being as a demon as it devilishly circumvented the Second Law. The demon must store the information that it gathers in a memory register. The resulting increase in the memory’s Shannon entropy then offsets the decrease of the Clausius entropy that is generated by the demon in the first place, thereby preventing a violation of the second law. Can a practical device that does this be constructed?

  • Computing beyond the Landauer limit
    Rolf Landauer argued that the cost of resetting one bit of information is at least kBT ln(2) energy dissipation, usually rendered as heat. It is widely understood that this means that an irreversible Boolean operation must dissipate at least that much energy, because the two inputs cannot be deduced from the single output, and so effectively one of the inputs is reset to zero. The traditional way to overcome this has been to consider reversible computing using Toffoli or Fredkin gates. Experiments suggest that computation which is physically reversible, albeit logically irreversible, is not subject to the Landauer constraint. Is it possible to build and demonstrate a practical working computer where the thermodynamic cost of a Boolean operation is less than kBT ln(2)?

  • Information engines
    Fundamental aspects of thermodynamics and information theory suggest that, under the right conditions, information may function as an entropic or thermodynamic resource. (Information here is to be understood in the technical sense. It need not be digital, but if it is it may be encoded in zeros and ones.) That is, information may in certain circumstances function like gasoline does in a car: when it reaches the engine, it serves as an energy source to power the engine. Is information as real as gasoline? In what ways can information do something akin to fuel? Is it possible to build and demonstrate an engine that harvests work from information? As an alternative to using information as fuel, it should be possible to satisfy the second law of thermodynamics by passing entropy as information to a memory register rather than as heat to a cold reservoir. Can an engine which operates by dumping entropy to an information register be demonstrated?

  • Experimental Landauer erasure of a quantum state
    It remains to be determined how the energy cost of processing quantum information relates to the energy cost of processing classical information. This may depend on how the processing is done. A quantum annealer which seeks to find the minimum of a given Hamiltonian has different thermodynamic constraints from measurement-based quantum computing. Even for circuit-based quantum information processing there is no agreed theory of energy dissipation. Can practical experiments evaluate competing models? Could they provide a new test for the arrow of time in experimental quantum mechanics?

  • Quantum Szilard engine
    The classical Szilard engine originated as a thought experiment to clarify the relationship between information and thermodynamic quantities such as entropy and temperature. Nanotechnology has advanced to the point where it has become possible to realize something close to Szilard’s original conception. One is led naturally to the concept of a quantum Szilard engine, in which a particle (e.g. an electron or an atom) is trapped in a confined system and used to extract energy from a thermal reservoir. How should work be defined in a quantum system? Is it possible to build a quantum version of Szilard’s engine?

  • Information as fuel in biology
    If information can be used as fuel, then a good place to look for practical implementations will be in molecular and cellular biology, to discover what has emerged through evolutionary optimisation. Do biological nanoengines harvest random fluctuations through a chemical Maxwellian angel to do mechanical work? How close are living systems to the thermodynamic limits the information processing that they require? What information about spatial organization—especially molecular and hierarchical—is detected and monitored in biological organisms? Do the information thermodynamics of biological organisms involve information about the environment (such as its randomness and structure) that reflects niche construction and the correlations between organism and niche that emerge? Could such information be used to develop a dynamical theory of an adaptive agent that builds a model of its environment, makes decisions based on that knowledge, and takes actions that affect the environment? Is biology required for this at all? Can abiotic adaptive learning agents be designed using modern nanoscale technologies?

  • Insights into fundamental philosophical aspects of entropy and information using the (in)distinguishability of particles
    Is it more correct to think of entropy as kinetically generated, with a value related to the amount of phase space that the system will eventually fill, or should one think of entropy from an information theoretic perspective, more along the lines of Shannon? In some cases, the treatment of distinguishability from the kinetic perspective appears to lead to nonsensical conclusions. On the other hand subtly different particles should have an entropy differing from fundamentally indistinguishable ones, as is known from experiments on the entropy of mixing. This dichotomy between views of entropy needs to be better understood, and experimental work, in vivo and in vitro would be of great interest.

Please enter your e-mail address:

Note: Joining the FQXi mailing list does not give you a login account or constitute membership in the organization.