State (computer science)

About: State (computer science) is a research topic. Over the lifetime, 24436 publications have been published within this topic receiving 225733 citations.

Extracting Automata from Recurrent Neural Networks Using Queries and Counterexamples

Abstract: We present a novel algorithm that uses exact learning and abstraction to extract a deterministic finite automaton describing the state dynamics of a given trained RNN. We do this using Angluin's L* algorithm as a learner and the trained RNN as an oracle. Our technique efficiently extracts accurate automata from trained RNNs, even when the state vectors are large and require fine differentiation.

Representing a computer system state to a user

Abstract: Operations to represent a computer system state to a user include maintaining a model (14) in a computer system (10). The model (14) is updated at times and represents a current state of the computer system (10). The model (14) uses formal-language statements to associate each of several predefined goals (16) with at least one of several predefined actions (18) that can be performed in the computer system (10) to accomplish the associated predefined goal. The operations comprise providing an output to a user regarding the current state of the computer system (10), the output comprising a natural-language statement generated using at least one of the formal-language statements. A statement generating module (20) may include a text planner (202), a grammar (206), a lexicon (208), a translator (204) or a text post-processor (210).

Computer processor with distributed pipeline control that allows functional units to complete operations out of order while maintaining precise interrupts

Abstract: A pipeline control system is distributed over the functional units (15, 17, 20, 25) in a processor (10). Decoder logic (12) issues operations, each with an associated tag, to the functional units, with up to n operations allowed to be outstanding. The units execute the operations and report termination information back to the decoder logic, but do not irrevocably change the state of the machine. Based on the termination information, the decoder logic retires normally terminated operations in order. If an operation terminates abnormally, the decoder logic instructs the units to back out of those operations that include and are later than the operation that terminated abnormally.

Automated power management system for a network of computers

Abstract: A computer system comprising a plurality of computing entities includes automatic power management logic that automatically transitions the system to a state in which less power is consumed when appropriate. The determination as to when this transition should occur is based on determining when demand for the processing abilities of the system are reduced. Once the decision has been made to transition to a reduced power state, the system's power management logic makes this transition in such a way to preferably minimize or at least reduce the performance impact on the system. Also, rather than altering the power state of one of the computing entities in the system, the entity can be deployed as part of another computing system.

Chemical implementation of finite-state machines

Abstract: With methods developed in a prior article on the chemical kinetic implementation of a McCulloch-Pitts neuron, connections among neurons, logic gates, and a clocking mechanism, we construct examples of clocked finite-state machines. These machines include a binary decoder, a binary adder, and a stack memory. An example of the operation of the binary adder is given, and the chemical concentrations corresponding to the state of each chemical neuron are followed in time. Using these methods, we can, in principle, construct a universal Turing machine, and these chemical networks inherit the halting problem