Traditional approaches to secure computation begin by representing the function f being computed as a circuit. For any function f that depends on each of its inputs, this implies a protocol with complexity at least linear in the input size. In fact, linear running time is inherent for secure computation of non-trivial functions, since each party must “touch” every bit of their input lest inform...

Turing machines which use a single tape both for their input and as their only working space (known as 1TTMs) appear to be extremely slow for many problems compared with multi-tape Turing machines and Random Access Machines (RAMS). However, apart from a few simple problems which require quadratic time on a 1TTM but have simple linear time solutions on a multi-tape Turing machine, there have bee...

In this paper, we propose and investigate a new neural network architecture called Neural Random Access Machine. It can manipulate and dereference pointers to an external variable-size random-access memory. The model is trained from pure input-output examples using backpropagation. We evaluate the new model on a number of simple algorithmic tasks whose solutions require pointer manipulation and...

The Rambus standardization skullduggery saga continues. As related in the May/June, July/August, and September/October 2001 Micro Law columns, SDRAM technology licensor Rambus sued chipmaker Infineon for patent infringement because Infineon refused to take a license under Rambus’ patents. (SDRAMs are synchronous dynamic random-access memory chips. Instead of running asynchronously (like ordinar...

I present a parallel algorithm for exact prob­ abilistic inference in Bayesian networks. For polytree networks with n variables, the worst­ case time complexity is O(log n) on a CREW PRAM (concurrent-read, exclusive-write paral­ lel random-access machine) with n processors, for any constant number of evidence variables. For arbitrary networks, the time complexity is O( r3w log n) for n processo...

Theorem 1 (Main Theorem). Every non-deterministic RAM computation in time t can be simulated by a non-deterministic oblivious two-tape TM in time t polylog(t). We prove this theorem in two steps. First, we show how to simulate non-deterministic RAM machines using non-deterministic multi-tape TMs with polylogarithmic overhead. Then, we show how to simulate multi-tape TMs using oblivious two-tape...

It is shown that the time needed by a concurrent read concurrent write parallel random access machine CRAM to check if an input has a certain property is the same as the minimal depth of a rst order inductive de nition of the property This in turn is equal to the number of iterations of a rst order sentence needed to express the property The second contribution of this paper is the introduction...

It is known that a parallel computer can solve problems that are impossible to be solved sequentially. That is, any general purpose sequential model of computation, such as the Turing machine or the random access machine (RAM), cannot simulate certain computations, for example solutions to real-time problems, that are carried out by a specific parallel computer. This paper extends the scope of ...

We prove a time-space tradeoff lower bound of T = Ω ( n log( n S ) log log( n S ) ) for randomized oblivious branching programs to compute 1GAP , also known as the pointer jumping problem, a problem for which there is a simple deterministic time n and space O(log n) RAM (random access machine) algorithm. Ajtai [3, 4] and Damgård, Meldgaard, and Nielsen [11] recently derived simulations of gener...

The parallel random access machine (PRAM) is the most commonly used general-purpose machine model for describing parallel computations. Unfortunately the PRAM model is not physically realizable, since on large machines a parallel shared memory access can only be accomplished at the cost of a signiicant time delay. A number of PRAM simulation algorithms have been presented in the literature. The...