1: \begin{abstract}

2: The well-known DeMillo-Lipton-Schwartz-Zippel lemma says that $n$-variate

3: polynomials of total degree at most $d$ over

4: grids, i.e. sets of the form $A_1 \times A_2 \times \cdots \times A_n$, form

5: error-correcting codes (of distance at least $2^{-d}$ provided $\min_i\{|A_i|\}\geq 2$).

6: In this work we explore their local

7: decodability and (tolerant) local testability. While these aspects have been studied

8: extensively when $A_1 = \cdots = A_n = \F_q$ are the same finite field, the

9: setting when $A_i$'s are not the full field does not seem to have been explored before.

10: 

11: In this work we focus on the case $A_i = \{0,1\}$ for every $i$. We show that for every field

12: (finite or otherwise) there is a test whose query complexity depends only on the

13: degree (and not on the number of variables). In contrast we show that

14: decodability is possible over fields of positive characteristic (with query complexity growing with

15: the degree of the polynomial and the characteristic), but not over the reals, where the query complexity must grow with $n$. As a

16: consequence we get a natural example of a code (one with a transitive group of symmetries) that is locally testable

17: but not locally decodable.

18: 

19: Classical results on local decoding and testing of polynomials have relied on the

20: 2-transitive symmetries of the space of low-degree polynomials (under affine

21: transformations). Grids do not possess this symmetry: So we introduce some new

22: techniques to overcome this handicap and in particular use the hypercontractivity of

23: the (constant weight) noise operator on the Hamming cube.

24: \end{abstract}

25: