joint marginals for k-tuples of distant random variables or satisfying assignments of CNFs. A popular solution in these cases is the idea of turning the complex inference task into a sequence of simpler ones by selecting and clamping variables one at a time and running lifted message passing again after

lifting and modified BP iterations. In turn, we can efficiently monitor the true BP messages sent and received in each iteration and group nodes accordingly. As our experiments show, iLBP can yield significantly faster more lifted network while not degrading performance. Above all, we show that i

Lifted message passing algorithms exploit repeated structure within a given graphical model to answer queries efficiently. Given evidence, they construct a lifted network of supernodes and superpotentials corresponding to sets of nodes and potentials that are indistinguishable given the evidence

to be enforced. We develop a generalization of the TRW-S algorithm [9] for this problem, where we use a decomposition into junction chains, monotonic w.r.t. some ordering on the nodes. This generalizes the monotonic chains in [9] in a natural way. We also show how to deal with nested factors in an efficient way

, coherent caching of global data provided by many shared-memory architectures is of crucial importance. Because message-passing mechanisms are much more efficient than shared-memory loads and stores for certain types of interprocessor communication and synchronization operations, however, we argue

Abstract not found

A reduced-complexity scheme using message

approach considers the problem of automatic parallelization and scheduling from sequential programs. The emphasis has been in the development of compilers or software tools that will assist in programming parallel architectures [2, 16, 18, 19]. Since message passing architectures require coarse grain

A week ago I didn’t understand kite power. I didn’t understand what were the best ways to turn kites into power-harvesting machines, and I didn’t know how to describe the theoretical limit for kite-based power-harvesting. Then I talked to Saul Griffith of Makani Power, who gave a great talk in Cambridge, and now it makes a lot more sense to me. Although power may be extracted from kites in ways that bear little resemblance to wind turbines, the fundamental limits for kite power are very similar to those of wind turbines. Kites have two advantages over windmills: they require less stuff per kilowatt of capacity; and they can suck energy from higher altitudes, where the winds are stronger and steadier. Here’s a summary of the conclusions: 1. Appropriately controlled kites can deliver a power per unit vertical area that is very similar to the power per unit vertical area of a standard wind turbine with blades sweeping out the same area. 2. A farm of kites could in principle deliver a power per unit land area that is similar to the power per unit land area of standard wind farm experiencing the same wind-speed. Saul Griffith said he reckons the limit for plausible kite-farms is an average power per unit land area of 6 W/m 2. How would it actually work? How can a high-speed twizzling kite be used to deliver power from the wind? A kite that just twizzles in circles at the end of a tethered string isn’t delivering any useful power, even if it is pulling with a large force on the string: a big force and a big velocity aren’t useful if they are perpendicular to each other. To get power from a force and a velocity, we need the force and the velocity to be aligned with each other – at least, partially aligned – not perpendicular. To explain some ways in which kite power could work, I’ll take an excursion through planes, gliders, and windmills. I’m going to assume that you have already read Technical Chapters B (wind) and C (flight) of Sustainable Energy – without the hot air. Drag forces

Abstract not found