Drawing from hats by noise-based logic

We utilize the asymmetric random telegraph wave-based instantaneous noise-base logic scheme to represent the problem of drawing numbers from a hat, and we consider two identical hats with the first 2^N integer numbers. In the first problem, Alice secretly draws an arbitrary number from one of the hats, and Bob must find out which hat is missing a number. In the second problem, Alice removes a known number from one of the hats and another known number from the other hat, and Bob must identify these hats. We show that, when the preparation of the hats with the numbers is accounted for, the noise-based logic scheme always provides an exponential speed-up and/or it requires exponentially smaller computational complexity than deterministic alternatives. Both the stochasticity and the ability to superpose numbers are essential components of the exponential improvement.

💡 Research Summary

The paper introduces a novel computational paradigm called noise‑based logic (NBL) and demonstrates how it can solve a class of combinatorial problems with exponential speed‑up compared to conventional deterministic algorithms. The authors build on an asymmetric random telegraph wave (RTW) scheme that generates stochastic binary signals which switch between two voltage levels with a certain probability. Each RTW acts as a “noise bit” and a set of N such bits can be combined to encode an N‑bit integer as a product (tensor‑like) of the corresponding RTW signals. Because the RTW signals are independent and superposable, a single physical channel can simultaneously represent all 2^N possible integers in a superposed “number wave”.

The concrete problem used as a testbed is the “hat‑drawing” scenario. Two identical hats each contain the first 2^N integer numbers, encoded as number waves. In the first variant, Alice secretly removes a single, arbitrarily chosen integer from one of the hats; Bob must determine which hat is missing a number. In the second variant, Alice removes two known integers, one from each hat, and Bob must identify the two hats that have been altered.

In a deterministic setting, solving the first variant requires scanning the contents of both hats, which entails O(2^N) time or, with sophisticated data structures, at best O(N) per query after an O(2^N) preprocessing step. The second variant is even more costly because two numbers must be located independently, leading to O(N·log 2^N) or higher per query. By contrast, NBL allows Bob to measure the superposed signals of both hats simultaneously, compute the difference signal, and read out the missing bits directly from the amplitude and sign of the difference wave. This operation is performed in a single logical step, giving a time complexity of O(N) regardless of the size of the hats.

The authors carefully account for the cost of preparing the hats—that is, converting each integer into its corresponding number wave. This preparation requires O(N·2^N) operations, but it is a one‑time cost. Once the hats are initialized, an arbitrary number of queries can be answered with only O(N) work each, yielding an exponential amortized speed‑up. The analysis shows that the stochastic nature of the RTW signals (providing true parallelism) and the ability to superpose many numbers in a single physical medium are both essential for the observed advantage.

Beyond the algorithmic analysis, the paper discusses practical implementation. The required hardware consists of RTW generators and simple logic gates, both of which can be realized in standard CMOS technology with low power consumption. Unlike quantum computing, NBL does not need cryogenic temperatures or elaborate error‑correction schemes, yet it achieves a form of parallelism comparable to quantum superposition.

In summary, the study demonstrates that noise‑based logic, when applied to problems that can be expressed as superpositions of binary states, can provide exponential reductions in both runtime and computational complexity. The hat‑drawing examples illustrate how a stochastic, superposable representation enables immediate identification of missing elements, a capability that would be infeasible with conventional deterministic circuits. The results suggest that NBL could become a powerful tool for high‑speed data retrieval, cryptanalysis, combinatorial optimization, and other domains where massive parallel comparison is required.