Title: The Zigzag Graph Product and ConstantDegree Lossless Expanders
1Expander Graphs The Unbalanced Case
Omer Reingold The Weizmann Institute
2What's in This Talk?
- Expander Graphs an array of definitions.
- Focus on most established notions, and open
problems on explicit constructions. Mainly in the
unbalanced case since this is - What applications often require
- Where constructions are very far from optimal
- Will flash one construction (no details) -
Unbalanced expanders based on Parvaresh-Vardy
Codes Guruswami,Umans,Vadhan 06
3Bipartite Graphs
- As a preparation for the unbalanced case we will
talk of bipartite expanders. - Can also capture undirected expanders
4Vertex Expansion
- Every (not too large) set expands.
5Vertex Expansion
- Goal minimize D (i.e. constant D)
- Degree 3 random graphs are expanders Pin73
6Vertex Expansion
- Also maximize A.
- Trivial upper bound A ? D
- even A ? D-1
- Random graphs A?D-1
72nd Eigenvalue Expansion
- 2nd eigenvalue (in absolute value) of
(normalized) adjacency matrix is bounded away
from 1 - Can be interpreted in terms of Renyi (l2)
entropy
8Expanders Add Entropy
N
Prob. dist. X
Induced dist. X
D
x
- Vertex expansion Support(X) ? A
Support(X) - Some applications rely on less naïve measures
of entropy. - Col(X) PrX(1)X(2) X2
92nd Eigenvalue Expansion
X
X
- Col(X) 1/N ? ?2 (Col(X) 1/N)
- Renyi entropy (log 1/Col(X)) increases as long
as ? lt 1 and Col(X) is not too small
102nd Eigenvalue Expansion
X
X
- Interestingly, vertex expansion and
2nd-eigenvalue expansion are essentially
equivalent for constant degree graphs Tan84,
AM84, Alo86
11Explicit Constructions
- Applications need explicit constructions
- Weakly explicit easy to build the entire graph
(in time poly N). - Strongly explicit
- Given vertex name x and edge label i easy to find
the ith neighbor of x (in time poly log N).
12Explicit constructions 2nd Eigenvalue
- Celebrated sequence of algebraic constructions
Mar73, GG80,JM85,LPS86,AGM87,Mar88,Mor94,.... - Optimal 2nd eigenvalue (Ramanujan graphs)
- Combinatorial constructions Ajt87, RVW00,
BL04. - Open Combinatorial constructions of strongly
explicit Ramanujan (or almost Ramanujan) graphs. - Getting close Ben-Aroya,Ta-Shma 08
13Explicit constructions Vertex Expansion
- Optimal 2nd eigenvalue expansion does not imply
optimal vertex expansion - Exist Ramanujan graphs with vertex expansion ?
D/2 Kah95. - Lossless Expander Expansion gt (1-??) D
- Why should we care?
- Limitation of previous techniques
- Many applications
14Property 1 A Very Strong Unique Neighbor Property
?S, S? K, ?(S) ? 0.9 D S
S
- S has ? 0.8 D S unique neighbors !
- We call graphs where every such S has even a
single unique neighbor unique neighbor
expanders
15Property 2 Incredibly Fault Tolerant
?S, S? K, ?(S) ? 0.9 D S
16Explicit constructions Vertex Expansion
- Open lossless expanders for the undirected case.
- Unique neighbor expanders are known AC02
- For the directed case (expansion only from left
side), lossless expanders are known CRVW02.
Expansion D-O(D?). - Open expansion D-O(1) (even with non-constant
degree).
17Unbalanced Expanders
18Unbalanced Expanders
- Many applications need unbalanced expanders
M
19Array of Definitions
M
X
X
- Many flavors
- How unbalanced.
- Measure of entropy.
- Lossless vs. lossy.
- Is X close to full entropy?
- Lower vs. upper bound on entropy of X.
-
20Vertex Expansion Revisited
M
- Even previously trivial tasks
- require D ??(log N/log M)
- M ltlt N ? Farewell constant degree
21Slightly-Unbalanced Constant-Degree Lossless
Expanders
M? N
?(S) ?(1-?) D S
CRVW02 0lt?,?? 1 constants ? D constant K? (N)
In case someone asks K? (? M/D) D poly(1/?
, log (1/? )) (fully explicit D quasipoly(1/?
, log (1/? )))
22Open More Unbalanced
M
- E.g. MN0.5 and sets of size at most KN0.2
expand. While being greedy - Unique neighbor expanders
- Lossless expanders
- Minimal Degree
23Super-Constant Degree
M
?S, S? K
?(S) ? (1-?)D S
- State of the art GUV06 DPoly(LogN),
MPoly(KD) (w. some tradeoff). - Open MO(KD) (known w. DQuasiPoly(LogN))
- Open D O(LogN)
24Dispersers Sipser 88
N
M
- Bounds
- D 1/? log(N/K)
- DK/M log 1/? -- must be lossy
- Explicit constructions are (comparably) good but
still not optimal
25Increasing Entropy?
M
Prob. dist. X
Induced dist. X
D
x
- Can Renyi entropy increase ?
- Col(X) lt Col(X) ? (essentially) Dgt
minM0.5, N/M
26 Extractors NZ 93
M N
X
X
- (k,?)-extractor if Min-entropy(X) ? k ? X
?-close to uniform - Min-entropy(X) ? k if ?x, Prx ? 2-k
- X and Y are ?-close if maxT PrX?T - PrY?T
½ X-Y1 ? ?
27Equivalently Extractors Mixing
M
- Vertex Expansion Sets on the left have many
neighbors. - Mixing Lemma the neighborhood of S hits any T
with roughly the right proportion.
282-Source Extractors
source of biased correlated bits
EXT
almost uniform output
another independent weak source
- Recently lots of attention and results
- Randomness Extractors are a special case, where
the 2nd source is truly random.
29Explicit Constructs. of Extractors
- Extractors are highly motivated in applications.
As a general rule of thumb Anything expanders
can do, extractors can do better - Lots of progress. Still very far from optimal.
Best in one direction LRVW03, GUV06
DPoly(LogN / ?), M2k(1-?) - Selected open problem M2k with DPoly(LogN / ?)
Interpretation extracting an arbitrary constant
fraction of entropy
Interpretation extracting all the entropy
30A Word About Techniques
- Research on randomness extractors was invigorated
with the discovery of a beautiful and surprising
connection to pseudorandom generators Tre99. - This further led to discoveries of connections
between extractors and error correcting codes
Tre99, RRV99, TZ01, TZS01, SU01. - In particular, GUV06 relies on Parvaresh-Vardy
list-decodable codes
31GUV06 - Basic Construction
- Left vertex f ?? Fqn (poly. of degree n-1 over
Fq) - Edge Label y ? F
- Right vertices Fqm1
- yth neighbor of f
- (y, f(y), (f h mod E)(y), (f h2 mod E)(y), , (f
hm-1 mod E)(y)) - where E(Y) irreducible poly of degree n h
a parameter - Thm This is a (K,A) expander with Khm, A
q-hnm.
32Conclusions
- Many interesting variants of expander graphs
- Constructions in general very far from optimal
- Any clean and useful algebraic characterization?