Abstract
The class FORMULA[s]∘G consists of Boolean functions computable by size-s De Morgan formulas whose leaves are any Boolean functions from a class G. We give lower bounds and (SAT, Learning, and pseudorandom generators (PRGs)) algorithms for FORMULA[n1.99]∘G, for classes G of functions with low communication complexity. Let R(k)G be the maximum k-party number-on-forehead randomized communication complexity of a function in G. Among other results, we show the following:
• | The Generalized Inner Product function GIPkn cannot be computed in FORMULA[s]° G on more than 1/2+ε fraction of inputs for s=o(n2/k⋅4k⋅R(k)(G)⋅log(n/ε)⋅log(1/ε))2). This significantly extends the lower bounds against bipartite formulas obtained by [62]. As a corollary, we get an average-case lower bound for GIPkn against FORMULA[n1.99]∘PTFk−1, i.e., sub-quadratic-size De Morgan formulas with degree-k-1) PTF (polynomial threshold function) gates at the bottom. Previously, it was open whether a super-linear lower bound holds for AND of PTFs. | ||||
• | There is a PRG of seed length n/2+O(s⋅R(2)(G)⋅log(s/ε)⋅log(1/ε)) that ε-fools FORMULA[s]∘G. For the special case of FORMULA[s]∘LTF, i.e., size-s formulas with LTF (linear threshold function) gates at the bottom, we get the better seed length O(n1/2⋅s1/4⋅log(n)⋅log(n/ε)). In particular, this provides the first non-trivial PRG (with seed length o(n)) for intersections of n halfspaces in the regime where ε≤1/n, complementing a recent result of [45]. | ||||
• | There exists a randomized 2n-t #SAT algorithm for FORMULA[s]∘G, where t=Ω(n\√s⋅log2(s)⋅R(2)(G))/1/2. In particular, this implies a nontrivial #SAT algorithm for FORMULA[n1.99]∘LTF. | ||||
• | The Minimum Circuit Size Problem is not in FORMULA[n1.99]∘XOR; thereby making progress on hardness magnification, in connection with results from [14, 46]. On the algorithmic side, we show that the concept class FORMULA[n1.99]∘XOR can be PAC-learned in time 2O(n/log n). | ||||
- Amir Abboud and Karl Bringmann. 2018. Tighter connections between formula-SAT and shaving logs. In ICALP. 8:1–8:18.Google Scholar
- Amir Abboud, Aviad Rubinstein, and R. Ryan Williams. 2017. Distributed PCP theorems for hardness of approximation in P. In FOCS. 25–36.Google Scholar
- Amir Abboud, Richard Ryan Williams, and Huacheng Yu. 2015. More applications of the polynomial method to algorithm design. In SODA. 218–230.Google Scholar
- Josh Alman, Timothy M. Chan, and R. Ryan Williams. 2016. Polynomial representations of threshold functions and algorithmic applications. In FOCS. 467–476.Google Scholar
- Noga Alon, Oded Goldreich, Johan Håstad, and René Peralta. 1992. Simple construction of almost k-wise independent random variables. Random Structures & Algorithms 3, 3 (1992), 289–304.Google Scholar
Cross Ref
- Andris Ambainis, Andrew M. Childs, Ben Reichardt, Robert Spalek, and Shengyu Zhang. 2010. Any AND-OR formula of size can be evaluated in time on a quantum computer. SIAM Journal on Computing 39, 6 (2010), 2513–2530.Google Scholar
Digital Library
- Alexander E. Andreev. 1987. On a method for obtaining more than quadratic effective lower bounds for the complexity of -schemes. Moscow Univeristy Mathematics Bulletin 42, 1 (1987), 63–66.Google Scholar
- Roy Armoni, Michael E. Saks, Avi Wigderson, and Shiyu Zhou. 1996. Discrepancy Sets and Pseudorandom generators for combinatorial rectangles. In FOCS. 412–421.Google Scholar
- László Babai, Noam Nisan, and Mario Szegedy. 1992. Multiparty protocols, pseudorandom generators for logspace, and time-space trade-offs. Journal of Computer and System Sciences 45, 2 (1992), 204–232.Google Scholar
Digital Library
- Robert Beals, Harry Buhrman, Richard Cleve, Michele Mosca, and Ronald de Wolf. 2001. Quantum lower bounds by polynomials. Journal of the ACM 48, 4 (2001), 778–797.Google Scholar
Digital Library
- Dan Boneh, Yuval Ishai, Alain Passelègue, Amit Sahai, and David J. Wu. 2018. Exploring crypto dark matter: New simple PRF candidates and their applications. In TCC. 699–729.Google Scholar
- Harry Buhrman, Ilan Newman, Hein Röhrig, and Ronald de Wolf. 2007. Robust polynomials and quantum algorithms. Theory of Computing Systems 40, 4 (2007), 379–395.Google Scholar
Digital Library
- Lijie Chen. 2018. On the hardness of approximate and exact (Bichromatic) maximum inner product. In CCC. 14:1–14:45.Google Scholar
- Lijie Chen, Ce Jin, and Ryan Williams. 2019. Hardness magnification for all sparse NP languages. In FOCS. 1240–1255.Google Scholar
- Lijie Chen and Ruosong Wang. 2019. Classical algorithms from quantum and Arthur-Merlin communication protocols. In ITCS. 23:1–23:20.Google Scholar
- Ruiwen Chen, Valentine Kabanets, Antonina Kolokolova, Ronen Shaltiel, and David Zuckerman. 2015. Mining circuit lower bound proofs for meta-algorithms. Computational Complexity 24, 2 (2015), 333–392.Google Scholar
Digital Library
- Mahdi Cheraghchi, Valentine Kabanets, Zhenjian Lu, and Dimitrios Myrisiotis. 2019. Circuit lower bounds for MCSP from local pseudorandom generators. In ICALP. 39:1–39:14.Google Scholar
- Don Coppersmith. 1982. Rapid multiplication of rectangular matrices. SIAM Journal on Computing 11, 3 (1982), 467–471.Google Scholar
Cross Ref
- Evgeny Dantsin and Edward A. Hirsch. 2009. Worst-case upper bounds.Handbook of Satisfiability, Vol. 185 (2009), 403–424.Google Scholar
- Irit Dinur and Or Meir. 2018. Toward the KRW composition conjecture: Cubic formula lower bounds via communication complexity. Computational Complexity 27, 3 (2018), 375–462.Google Scholar
Digital Library
- Edward Farhi, Jeffrey Goldstone, and Sam Gutmann. 2008. A quantum algorithm for the Hamiltonian NAND tree. Theory of Computing 4, 1 (2008), 169–190.Google Scholar
Cross Ref
- Bill Fefferman, Ronen Shaltiel, Christopher Umans, and Emanuele Viola. 2013. On beating the hybrid argument. Theory of Computing 9 (2013), 809–843.Google Scholar
Cross Ref
- Yoav Freund. 1990. Boosting a weak learning algorithm by majority. In COLT. 202–216.Google Scholar
- Mika Göös, Toniann Pitassi, and Thomas Watson. 2018. The landscape of communication complexity classes. Computational Complexity 27, 2 (2018), 245–304.Google Scholar
Digital Library
- Parikshit Gopalan, Ryan O'Donnell, Yi Wu, and David Zuckerman. 2010. Fooling functions of halfspaces under product distributions. In CCC. 223–234.Google Scholar
- Lov K. Grover. 1996. A fast quantum mechanical algorithm for database search. In STOC. 212–219.Google Scholar
- Prahladh Harsha and Srikanth Srinivasan. 2019. On polynomial approximations to AC. Random Structures & Algorithms 54, 2 (2019), 289–303.Google Scholar
Cross Ref
- Johan Håstad. 1998. The shrinkage exponent of De Morgan formulas is 2. SIAM Journal on Computing 27, 1 (1998), 48–64.Google Scholar
Cross Ref
- Peter Høyer, Troy Lee, and Robert Spalek. 2007. Negative weights make adversaries stronger. In STOC. 526–535.Google Scholar
- Russell Impagliazzo, Raghu Meka, and David Zuckerman. 2012. Pseudorandomness from Shrinkage. In FOCS. 111–119.Google Scholar
- Russell Impagliazzo and Noam Nisan. 1993. The effect of random restrictions on formula size. Random Structures & Algorithms 4, 2 (1993), 121–134.Google Scholar
Cross Ref
- Russell Impagliazzo, Noam Nisan, and Avi Wigderson. 1994. Pseudorandomness for network algorithms. In STOC. 356–364.Google Scholar
- Stasys Jukna. 2012. Boolean Function Complexity: Advances and Frontiers. Algorithms and Combinatorics, Vol. 27. Springer.Google Scholar
Digital Library
- Valentine Kabanets. 2002. Derandomization: A brief overview. Current Trends in Theoretical Computer Science 1 (2002), 165–188.Google Scholar
- Valentine Kabanets and Zhenjian Lu. 2018. Satisfiability and derandomization for small polynomial threshold circuits. In APPROX/RANDOM. 46:1–46:19.Google Scholar
- Adam Tauman Kalai, Adam R. Klivans, Yishay Mansour, and Rocco A. Servedio. 2008. Agnostically learning halfspaces. SIAM Journal on Computing 37, 6 (2008), 1777–1805.Google Scholar
Digital Library
- Adam Tauman Kalai, Yishay Mansour, and Elad Verbin. 2008. On agnostic boosting and parity learning. In STOC. 629–638.Google Scholar
- Michael J. Kearns and Umesh V. Vazirani. 1994. An Introduction to Computational Learning Theory. MIT Press.Google Scholar
Digital Library
- Valeriy M. Khrapchenko. 1971. Method of determining lower bounds for the complexity of -schemes. Mathematical Notes 10, 1 (1971), 474–479.Google Scholar
Cross Ref
- Eyal Kushilevitz and Noam Nisan. 1997. Communication Complexity. Cambridge University Press.Google Scholar
Digital Library
- Sophie Laplante, Troy Lee, and Mario Szegedy. 2006. The quantum adversary method and classical formula size lower bounds. Computational Complexity 15, 2 (2006), 163–196.Google Scholar
Digital Library
- Saburo Muroga, Iwao Toda, and Satoru Takasu. 1961. Theory of majority decision elements. Journal of the Franklin Institute 271 (1961), 376–418.Google Scholar
Cross Ref
- E. I. Nechiporuk. 1966. On a Boolean function. Doklady Akademii Nauk SSSR 169, 4 (1966), 765–766. English translation in Soviet Mathematics Doklady.Google Scholar
- Noam Nisan. 1994. The communication complexity of threshold gates. In Proceedings of “Combinatorics, Paul Erdos is Eighty”. 301–315.Google Scholar
- Ryan O'Donnell, Rocco A. Servedio, and Li-Yang Tan. 2019. Fooling polytopes. In STOC. 614–625.Google Scholar
- Igor Carboni Oliveira, Ján Pich, and Rahul Santhanam. 2019. Hardness magnification near state-of-the-art lower bounds. In CCC. 27:1–27:29.Google Scholar
- Igor Carboni Oliveira and Rahul Santhanam. 2017. Conspiracies between learning algorithms, circuit lower bounds, and pseudorandomness. In CCC. 18:1–18:49.Google Scholar
- Mike Paterson and Uri Zwick. 1993. Shrinkage of De Morgan formulae under restriction. Random Structures & Algorithms 4, 2 (1993), 135–150.Google Scholar
Cross Ref
- Mihai Patrascu and Ryan Williams. 2010. On the possibility of faster SAT algorithms. In SODA. 1065–1075.Google Scholar
- Pavel Pudlák, Vojtech Rödl, and Petr Savický. 1988. Graph complexity. Acta Informatica 25, 5 (1988), 515–535. https://doi.org/10.1007/BF00279952Google Scholar
Digital Library
- Ben Reichardt. 2009. Span programs and quantum query complexity: The general adversary bound is nearly tight for every Boolean function. In FOCS. 544–551.Google Scholar
- Ben Reichardt. 2011. Faster quantum algorithm for evaluating game trees. In SODA. 546–559.Google Scholar
- Ben Reichardt. 2011. Reflections for quantum query algorithms. In SODA. 560–569.Google Scholar
- Ben Reichardt and Robert Spalek. 2012. Span-program-based quantum algorithm for evaluating formulas. Theory of Computing 8, 1 (2012), 291–319.Google Scholar
Cross Ref
- Igor S. Sergeev. 2017. Upper bounds for the size and the depth of formulae for MOD-functions. Discrete Mathematics and Applications 27, 1 (2017), 15–22.Google Scholar
Cross Ref
- Rocco A. Servedio and Li-Yang Tan. 2017. What circuit classes can be learned with non-trivial savings?. In ITCS. 30:1–30:21.Google Scholar
- Rocco A. Servedio and Li-Yang Tan. 2019. Improved pseudorandom generators from pseudorandom multi-switching lemmas. In APPROX/RANDOM. 45:1–45:23.Google Scholar
- Bella Abramovna Subbotovskaya. 1961. Realization of linear functions by formulas using , &, –. In Doklady Akademii Nauk, Vol. 136. Russian Academy of Sciences, 553–555.Google Scholar
- Avishay Tal. 2014. Shrinkage of De Morgan formulae by spectral techniques. In FOCS. 551–560.Google Scholar
- Avishay Tal. 2015. #SAT algorithms from shrinkage. Electronic Colloquium on Computational Complexity (ECCC) 22 (2015), 114.Google Scholar
- Avishay Tal. 2016. The bipartite formula complexity of inner-product is quadratic. Electronic Colloquium on Computational Complexity (ECCC) 23 (2016), 181.Google Scholar
- Avishay Tal. 2017. Formula lower bounds via the quantum method. In STOC. 1256–1268.Google Scholar
- Salil P. Vadhan. 2012. Pseudorandomness. Foundations and Trends in Theoretical Computer Science 7, 1-3 (2012), 1–336.Google Scholar
Cross Ref
- Leslie G. Valiant. 1984. A theory of the learnable. In STOC. 436–445.Google Scholar
- Emanuele Viola. 2015. The communication complexity of addition. Combinatorica 35, 6 (2015), 703–747.Google Scholar
Digital Library
- Ryan Williams. 2014. Nonuniform ACC circuit lower bounds. Journal of the ACM 61, 1 (2014), 2:1–2:32.Google Scholar
Digital Library
- Ryan Williams. 2016. Strong ETH breaks with Merlin and Arthur: Short non-interactive proofs of batch evaluation. In CCC. 2:1–2:17.Google Scholar
- Andrew Chi-Chih Yao. 1982. Theory and applications of trapdoor functions. In FOCS. 80–91.Google Scholar
Index Terms
Algorithms and Lower Bounds for De Morgan Formulas of Low-Communication Leaf Gates
Recommendations
Algorithms and lower bounds for de morgan formulas of low-communication leaf gates
CCC '20: Proceedings of the 35th Computational Complexity ConferenceThe class FORMULA[s] ο G consists of Boolean functions computable by size-s de Morgan formulas whose leaves are any Boolean functions from a class G. We give lower bounds and (SAT, Learning, and PRG) algorithms for FORMULA[n1.99] ο G, for classes G of ...
Circuit Lower Bounds for MCSP from Local Pseudorandom Generators
The Minimum Circuit Size Problem (MCSP) asks if a given truth table of a Boolean function f can be computed by a Boolean circuit of size at most θ, for a given parameter θ. We improve several circuit lower bounds for MCSP, using pseudorandom generators (...
Shrinkage of De Morgan Formulae by Spectral Techniques
FOCS '14: Proceedings of the 2014 IEEE 55th Annual Symposium on Foundations of Computer ScienceWe give a new and improved proof that the shrinkage exponent of De Morgan formulae is 2. Namely, we show that for any Boolean function f : {0,1}n → {0,1}, setting each variable out of x1,…,xn with probability 1 -- p to a randomly chosen constant, ...






Comments