170 results about "Link weight" patented technology

An integrated circuit comprising a plurality of processing modules (M, S; IP) and a network (N) arranged for coupling said modules (M, S; IP) is provided. Said integrated circuit further comprises a plurality of network interfaces (NI) each being coupled between one of said processing modules (M, S; IP) and said network (N). Said network (N) comprises a plurality of routers (R) coupled via network links (L) to adjacent routers (R). Said processing modules (M, S; IP) communicate between each other over connections using connection paths (C1-C12) through the network (N), wherein each of said connection paths (C1-C12) employ at least one network link (L) for a required number of time slots. At least one time slot allocating unit (SA) is provided for computing a link weight factor for at least one network link (L) in said connection path (C1-C12) as a function of at least one connection requirement for said at least one network link (L), for computing a connection path weight factor for at least one connection path (C1-C12) as a function of the computed link weight factor of at least one network link (L) in said connection path (C1-C12), and for allocating time slots to said network links (L) according to the computed connection path weight factors.

A method, system, and computer program product for hypertext link analysis that includes independently employing non-normalized backward and forward operators to obtain two independent weights for each document in a hypertext-linked graph, for subsequent ranking and analysis.

A route computation algorithm, a load-balancing scheme inside a router for making a hop-by-hop routing network (such as the Internet) nonblocking are described in this patent. The output of the route computation algorithm includes a set of link weights that determine the paths of the hop-by-hop routing network. The route computation algorithm and the load-balancing scheme also determine the ingress and egress traffic constraints at each edge router such that as long as the traffic entering and leaving the network does not exceed the constraints, none of the internal links will ever have traversing traffic more than its link capacity. The network is thus non-blocking internally. This greatly simplifies flow admission control and allows hard QoS to be supported inside the network.

A control apparatus which is capable of compensating for a control error properly and quickly even under a condition where the control error is temporarily increased e.g. by degradation of reliability of the detection results of reference parameters other than controlled variables, thereby making it possible to ensure a high accuracy of control. An air-fuel ratio controller of the control apparatus calculates modified errors by multiplying e.g. an air-fuel ratio error estimated value by link weight functions, calculates basic local correction values such that the modified errors become equal to 0; calculates local correction values by multiplying the basic local correction values and the like by the link weight functions; calculates corrected valve lift by adding a lift correction value, which is the total sum of the local correction values, to a value of valve lift; calculates a first estimated intake air amount for feedforward control of an air-fuel ratio, based on the corrected valve lift; calculates an air-fuel ratio correction coefficient for feedback control of the air-fuel ratio; and calculates a fuel injection amount based on these.

The present invention extends to methods, systems, and computer program products for recognizing input gestures. A neural network is trained using example inputs and backpropagation to recognize specified input patterns. Input gesture data is representative of movements in contact on a multi-touch input display surface relative to one or more axes over time. Example inputs used for training the neural network to recognize a specified input pattern can be created from sampling input gesture data for example input gestures known to represent the specified input pattern. Trained neural networks can subsequently be used to recognize input gestures that are similar to known input gestures as the specified input pattern corresponding to the known input gestures.

One embodiment of the present invention provides a system that learns a generative model for textual documents. During operation, the system receives a current model, which contains terminal nodes representing random variables for words and cluster nodes representing clusters of conceptually related words. Within the current model, nodes are coupled together by weighted links, so that if a cluster node in the probabilistic model fires, a weighted link from the cluster node to another node causes the other node to fire with a probability proportionate to the link weight. The system also receives a set of training documents, wherein each training document contains a set of words. Next, the system applies the set of training documents to the current model to produce a new model.