On the power of small-depth threshold circuits

We investigate the power of threshold circuits of small depth. In particular, we give functions that require exponential size unweighted threshold circuits of depth 3 when we restrict the bottom fanin. We also prove that there are monotone functionsf _k that can be computed in depthk and linear size , -circuits but require exponential size to compute by a depthk–1 monotone weighted threshold circuit.     

Learning a circuit by injecting values

We propose a new model for exact learning of acyclic circuits using experiments in which chosen values may be assigned to an arbitrary subset of wires internal to the circuit, but only the value of the circuit's single output wire may be observed. We give polynomial time algorithms to learn (1) arbitrary circuits with logarithmic depth and constant fan-in and (2) Boolean circuits of constant depth and unbounded fan-in over AND, OR, and NOT gates. Thus, both AC0 and NC1 circuits are learnable in polynomial time in this model. Negative results show that some restrictions on depth, fan-in and gate types are necessary: exponentially many experiments are required to learn AND/OR circuits of unbounded depth and fan-in; it is NP-hard to learn AND/OR circuits of unbounded depth and fan-in 2; and it is NP-hard to learn circuits of constant depth and unbounded fan-in over AND, OR, and threshold gates, even when the target circuit is known to contain at most one threshold gate and that threshold gate has threshold 2. We also consider the effect of adding an oracle for behavioral equivalence. In this case there are polynomial-time algorithms to learn arbitrary circuits of constant fan-in and unbounded depth and to learn Boolean circuits with arbitrary fan-in and unbounded depth over AND, OR, and NOT gates. A corollary is that these two classes are PAC-learnable if experiments are available. Finally, we consider an extension of the model called the synchronous model. We show that an even more general class of circuits are learnable in this model. In particular, we are able to learn circuits with cycles.     

Exponential lower bounds for depth three Boolean circuits

We consider the class of unbounded fan-in depth three Boolean circuits, for which the bottom fan-in is limited by k and the top gate is an OR. It is known that the smallest such circuit computing the parity function has gates (for k = O(n ^1/2)) for some , and this was the best lower bound known for explicit (P-time computable) functions. In this paper, for k = 2, we exhibit functions in uniform NC ¹ that require size depth 3 circuits. The main tool is a theorem that shows that any circuit on n variables that accepts a inputs and has size s must be constant on a projection (subset defined by equations of the form x _i = 0, x _i = 1, x _i = x _j or x _i = ) of dimension at least log(a/s)log n. Received: April 1, 1997.     

Circuits and multi-party protocols

Using multi-party communication techniques, we prove that depth-3 circuits with a threshold gate at the top, arbitrary symmetric gates at the next level, and fan-in k MOD _m gates at the bottom need exponential size to compute the k-wise inner product function of Babai, Nisan and Szegedy, where m is an odd positive integer satisfying . This is one of the rare lower-bound results involving MOD _m gates with non-prime power moduli.?Exponential gap theorems are also presented between the multi-party communication complexities of closely related functions. Received: January 5, 1994     

Complex polynomials and circuit lower bounds for modular counting

We study the power of constant-depth circuits containing negation gates, unbounded fan-in AND and OR gates, and a small number of MAJORITY gates. It is easy to show that a depth 2 circuit of sizeO(n) (wheren is the number of inputs) containingO(n) MAJORITY gates can determine whether the sum of the input bits is divisible byk, for any fixedk>1, whereas it is known that this requires exponentialsize circuits if we have no MAJORITY gates. Our main result is that a constant-depth circuit of size containingn ^o(1) MAJORITY gates cannot determine if the sum of the input bits is divisible byk; moreover, such a circuit must give the wrong answer on a constant fraction of the inputs. This result was previously known only fork=2. We prove this by obtaining an approximate representation of the behavior of constant-depth circuits by multivariate complex polynomials.     

Perceptrons of large weight

A threshold gate is a linear combination of input variables with integer coefficients (weights). It outputs 1 if the sum is positive. The maximum absolute value of the coefficients of a threshold gate is called its weight. A degree-d perceptron is a Boolean circuit of depth 2 with a threshold gate at the top and any Boolean elements of fan-in at most d at the bottom level. The weight of a perceptron is the weight of its threshold gate.For any constant d ≥ 2 independent of the number of input variables n, we construct a degree-d perceptron that requires weights of at least \(n^{\Omega (n^d )} \); i.e., the weight of any degree-d perceptron that computes the same Boolean function must be at least \(n^{\Omega (n^d )} \). This bound is tight: any degree-d perceptron is equivalent to a degree-d perceptron of weight \(n^{O(n^d )} \). For the case of threshold gates (i.e., d = 1), the result was proved by Håstad in [2]; we use Håstad’s technique.     

Randomness-Efficient Sampling within NC1

We construct a randomness-efficient averaging sampler that is computable by uniform constant-depth circuits with parity gates (i.e., in uniform AC ⁰[⊕]). Our sampler matches the parameters achieved by random walks on constant-degree expander graphs, allowing us to apply a variety expander-based techniques within NC ¹. For example, we obtain the following results:

Monotone circuits for monotone weighted threshold functions

Weighted threshold functions with positive weights are a natural generalization of unweighted threshold functions. These functions are clearly monotone. However, the naive way of computing them is adding the weights of the satisfied variables and checking if the sum is greater than the threshold; this algorithm is inherently non-monotone since addition is a non-monotone function. In this work we by-pass this addition step and construct a polynomial size logarithmic depth unbounded fan-in monotone circuit for every weighted threshold function, i.e., we show that weighted threshold functions are in mAC¹. (To the best of our knowledge, prior to our work no polynomial monotone circuits were known for weighted threshold functions.)Our monotone circuits are applicable for the cryptographic tool of secret sharing schemes. Using general results for compiling monotone circuits (Yao, 1989) and monotone formulae (Benaloh and Leichter, 1990) into secret sharing schemes, we get secret sharing schemes for every weighted threshold access structure. Specifically, we get: (1) information-theoretic secret sharing schemes where the size of each share is quasi-polynomial in the number of users, and (2) computational secret sharing schemes where the size of each share is polynomial in the number of users.     

On Approximate Majority and Probabilistic Time

We prove new results on the circuit complexity of approximate majority, which is the problem of computing the majority of a given bit string whose fraction of 1’s is bounded away from 1/2 (by a constant). We then apply these results to obtain new relationships between probabilistic time, BPTime (t), and alternating time, ∑_O(1)Time (t). Our main results are the following:

Introductory paper

This special section is devoted to a selection of papers that originally appeared in two thematic sessions on high-level testing of complex systems at IDPT 2002 and 2003, the 6th and 7th World Conference on Integrated Design and Process Technology, which took place in Pasadena, CA in June 2002 and in Austin, TX in December 2003, respectively.This collection of papers spans a wide panoramic view on the development of testing and validation technology along several dimmensions. It touches on issues such as

AI Bridges and Dreams

This paper is a modified version of my acceptance lecture for the 1986 SPL-Insight Award. It turned into something of a personal credo -describing my view of

BPP has subexponential time simulations unlessEXPTIME has publishable proofs

We show thatBPP can be simulated in subexponential time for infinitely many input lengths unless exponential time

Complexities of efficient solutions of rectilinear polygon cover problems

The rectilinear polygon cover problem is one in which a certain class of features of a rectilinear polygon ofn vertices has to be covered with the minimum number of rectangles included in the polygon. In particular, we consider covering the entire interior, the boundary, and the set of corners of the polygon. These problems have important applications in storing images and in the manufacture of integrated circuits. Unfortunately, most of these problems are known to be NP-complete. Hence it is necessary to develop efficient heuristics for these problems or to show that the design of efficient heuristics is impossible. In this paper we show:

A Network Game with Attackers and a Defender

Consider an information network with threats called attackers; each attacker uses a probability distribution to choose a node of the network to damage. Opponent to the attackers is a protector entity called defender; the defender scans and cleans from attacks some part of the network (in particular, a link), which it chooses independently using its own probability distribution. Each attacker wishes to maximize the probability of escaping its cleaning by the defender; towards a conflicting objective, the defender aims at maximizing the expected number of attackers it catches. We model this network security scenario as a non-cooperative strategic game on graphs. We are interested in its associated Nash equilibria, where no network entity can unilaterally increase its local objective. We obtain the following results:

Quantum Algorithms for Matching Problems

We present quantum algorithms for the following matching problems in unweighted and weighted graphs with n vertices and m edges:

Logic-based processing of semantically complex natural language discourse

A logic-based system of knowledge representation for natural language discourse has three primary advantages: