Small-bias sample space

In theoretical computer science, a small-bias sample space (also known as -biased sample space, -biased generator, or small-bias probability space) is a probability distribution that fools parity functions. In other words, no parity function can distinguish between a small-bias sample space and the uniform distribution with high probability, and hence, small-bias sample spaces naturally give rise to pseudorandom generators for parity functions.

The main useful property of small-bias sample spaces is that they need far fewer truly random bits than the uniform distribution to fool parities. Efficient constructions of small-bias sample spaces have found many applications in computer science, some of which are derandomization, error-correcting codes, and probabilistically checkable proofs. The connection with error-correcting codes is in fact very strong since -biased sample spaces are equivalent to -balanced error-correcting codes.

Definition

edit

Bias

edit

Let   be a probability distribution over  . The bias of   with respect to a set of indices   is defined as[1]

 

where the sum is taken over  , the finite field with two elements. In other words, the sum   equals   if the number of ones in the sample   at the positions defined by   is even, and otherwise, the sum equals  . For  , the empty sum is defined to be zero, and hence  .

ϵ-biased sample space

edit

A probability distribution   over   is called an  -biased sample space if   holds for all non-empty subsets  .

ϵ-biased set

edit

An  -biased sample space   that is generated by picking a uniform element from a multiset   is called  -biased set. The size   of an  -biased set   is the size of the multiset that generates the sample space.

ϵ-biased generator

edit

An  -biased generator   is a function that maps strings of length   to strings of length   such that the multiset   is an  -biased set. The seed length of the generator is the number   and is related to the size of the  -biased set   via the equation  .

Connection with epsilon-balanced error-correcting codes

edit

There is a close connection between  -biased sets and  -balanced linear error-correcting codes. A linear code   of message length   and block length   is  -balanced if the Hamming weight of every nonzero codeword   is between   and  . Since   is a linear code, its generator matrix is an  -matrix   over   with  .

Then it holds that a multiset   is  -biased if and only if the linear code  , whose columns are exactly elements of  , is  -balanced.[2]

Constructions of small epsilon-biased sets

edit

Usually the goal is to find  -biased sets that have a small size   relative to the parameters   and  . This is because a smaller size   means that the amount of randomness needed to pick a random element from the set is smaller, and so the set can be used to fool parities using few random bits.

Theoretical bounds

edit

The probabilistic method gives a non-explicit construction that achieves size  .[2] The construction is non-explicit in the sense that finding the  -biased set requires a lot of true randomness, which does not help towards the goal of reducing the overall randomness. However, this non-explicit construction is useful because it shows that these efficient codes exist. On the other hand, the best known lower bound for the size of  -biased sets is  , that is, in order for a set to be  -biased, it must be at least that big.[2]

Explicit constructions

edit

There are many explicit, i.e., deterministic constructions of  -biased sets with various parameter settings:

  • Naor & Naor (1990) achieve  . The construction makes use of Justesen codes (which is a concatenation of Reed–Solomon codes with the Wozencraft ensemble) as well as expander walk sampling.
  • Alon et al. (1992) achieve  . One of their constructions is the concatenation of Reed–Solomon codes with the Hadamard code; this concatenation turns out to be an  -balanced code, which gives rise to an  -biased sample space via the connection mentioned above.
  • Concatenating Algebraic geometric codes with the Hadamard code gives an  -balanced code with  .[2]
  • Ben-Aroya & Ta-Shma (2009) achieves  .
  • Ta-Shma (2017) achieves   which is almost optimal because of the lower bound.

These bounds are mutually incomparable. In particular, none of these constructions yields the smallest  -biased sets for all settings of   and  .

Application: almost k-wise independence

edit

An important application of small-bias sets lies in the construction of almost k-wise independent sample spaces.

k-wise independent spaces

edit

A random variable   over   is a k-wise independent space if, for all index sets   of size  , the marginal distribution   is exactly equal to the uniform distribution over  . That is, for all such   and all strings  , the distribution   satisfies  .

Constructions and bounds

edit

k-wise independent spaces are fairly well understood.

  • A simple construction by Joffe (1974) achieves size  .
  • Alon, Babai & Itai (1986) construct a k-wise independent space whose size is  .
  • Chor et al. (1985) prove that no k-wise independent space can be significantly smaller than  .

Joffe's construction

edit

Joffe (1974) constructs a  -wise independent space   over the finite field with some prime number   of elements, i.e.,   is a distribution over  . The initial   marginals of the distribution are drawn independently and uniformly at random:

 .

For each   with  , the marginal distribution of   is then defined as

 

where the calculation is done in  . Joffe (1974) proves that the distribution   constructed in this way is  -wise independent as a distribution over  . The distribution   is uniform on its support, and hence, the support of   forms a  -wise independent set. It contains all   strings in   that have been extended to strings of length   using the deterministic rule above.

Almost k-wise independent spaces

edit

A random variable   over   is a  -almost k-wise independent space if, for all index sets   of size  , the restricted distribution   and the uniform distribution   on   are  -close in 1-norm, i.e.,  .

Constructions

edit

Naor & Naor (1990) give a general framework for combining small k-wise independent spaces with small  -biased spaces to obtain  -almost k-wise independent spaces of even smaller size. In particular, let   be a linear mapping that generates a k-wise independent space and let   be a generator of an  -biased set over  . That is, when given a uniformly random input, the output of   is a k-wise independent space, and the output of   is  -biased. Then   with   is a generator of an  -almost  -wise independent space, where  .[3]

As mentioned above, Alon, Babai & Itai (1986) construct a generator   with  , and Naor & Naor (1990) construct a generator   with  . Hence, the concatenation   of   and   has seed length  . In order for   to yield a  -almost k-wise independent space, we need to set  , which leads to a seed length of   and a sample space of total size  .

Notes

edit
  1. ^ cf., e.g., Goldreich (2001)
  2. ^ a b c d cf., e.g., p. 2 of Ben-Aroya & Ta-Shma (2009)
  3. ^ Section 4 in Naor & Naor (1990)

References

edit
  • Alon, Noga; Babai, László; Itai, Alon (1986), "A fast and simple randomized parallel algorithm for the maximal independent set problem" (PDF), Journal of Algorithms, 7 (4): 567–583, doi:10.1016/0196-6774(86)90019-2
  • Alon, Noga; Goldreich, Oded; Håstad, Johan; Peralta, René (1992), "Simple Constructions of Almost k-wise Independent Random Variables" (PDF), Random Structures & Algorithms, 3 (3): 289–304, CiteSeerX 10.1.1.106.6442, doi:10.1002/rsa.3240030308
  • Ben-Aroya, Avraham; Ta-Shma, Amnon (2009). "Constructing Small-Bias Sets from Algebraic-Geometric Codes". 2009 50th Annual IEEE Symposium on Foundations of Computer Science (PDF). pp. 191–197. CiteSeerX 10.1.1.149.9273. doi:10.1109/FOCS.2009.44. ISBN 978-1-4244-5116-6.
  • Chor, Benny; Goldreich, Oded; Håstad, Johan; Freidmann, Joel; Rudich, Steven; Smolensky, Roman (1985). "The bit extraction problem or t-resilient functions". 26th Annual Symposium on Foundations of Computer Science (SFCS 1985). pp. 396–407. CiteSeerX 10.1.1.39.6768. doi:10.1109/SFCS.1985.55. ISBN 978-0-8186-0644-1. S2CID 6968065.
  • Goldreich, Oded (2001), Lecture 7: Small bias sample spaces
  • Joffe, Anatole (1974), "On a Set of Almost Deterministic k-Independent Random Variables", Annals of Probability, 2 (1): 161–162, doi:10.1214/aop/1176996762
  • Naor, Joseph; Naor, Moni (1990), "Small-bias probability spaces: Efficient constructions and applications", Proceedings of the twenty-second annual ACM symposium on Theory of computing - STOC '90, pp. 213–223, CiteSeerX 10.1.1.421.2784, doi:10.1145/100216.100244, ISBN 978-0897913614, S2CID 14031194
  • Ta-Shma, Amnon (2017), "Explicit, almost optimal, epsilon-balanced codes", Proceedings of the 49th Annual ACM SIGACT Symposium on Theory of Computing, pp. 238–251, doi:10.1145/3055399.3055408, ISBN 9781450345286, S2CID 5648543