Abstract: We propose a novel approximate computing technique for low-power ternary multiplication. A carry-truncated ternary multiplier, error compensation circuits ...