Abstract: A method of reinforcement learning of a neural network device for generating a verification vector for verifying a circuit design comprising a circuit block includes inputting a test vector to the circuit block, generating one or more rewards based on a coverage corresponding to the test vector, the coverage being determined based on a state transition of the circuit block based on the test vector, and applying the one or more rewards to a reinforcement learning.

Abstract: Sealant compositions particularly suitable for solid oxide fuel cell sealant are provided and preferably comprise glass matrix and ceramic fiber, wherein glass matrix and ceramic fiber are mixed in an volume ratio of 25:75-75:25 in the sealant, and the ceramic fibers are preferably uniformly dispersed in the sealant to exhibit an orientation. Methods to manufacture the sealant compositions also are provided. Particularly preferred sealant compositions of the invention can efficiently avoid undesired viscous flow of glass matrix, precisely locate the stack of fuel cell on the region to be sealed, and maintain uniform sealing ability under various changes in size of the fuel cell stack.

Abstract: The present invention relates to a carry look-ahead adder. The carry look-ahead adder is configured in 4-bit units in general. Each 4-bt carry look-ahead adder is combined with a carry look-ahead generating unit to constitute a carry look-ahead adder that can process input signals of larger size. The carry look-ahead adder according to the embodiment of the present invention calculates carry of each bit sequentially not by using any carry generation function or any carry propagation function but by using previous bit when generating an internal carry in the adder, so that propagation delay is allowed a little but the logic gate circuit can be simplified.