Abstract: Techniques for classifier generalization in a supervised learning process using input encoding are provided. In one aspect, a method for classification generalization includes: encoding original input features from at least one input sample {right arrow over (x)}S with a uniquely decodable code using an encoder E(?) to produce encoded input features E({right arrow over (x)}S), wherein the at least one input sample {right arrow over (x)}S comprises uncoded input features; feeding the uncoded input features and the encoded input features E({right arrow over (x)}S) to a base model to build an encoded model; and learning a classification function {tilde over (C)}E(?) using the encoded model, wherein the classification function {tilde over (C)}E(?) learned using the encoded model is more general than that learned using the uncoded input features alone.

Abstract: Second machine learning models trained using respective second data sets can be received. The second machine learning models can be run using a first data set used in training a first machine learning model, where the second machine learning models produce respective outputs. Scores associated with the second machine learning models can be determined by comparing the respective outputs with ground truth associated with the first data set. Based on the scores associated with the second machine learning models, whether the first data set is to be discarded or kept can be determined for training the first machine learning model.

Abstract: Embodiments relate to an input-encoding technique in conjunction with federation. Participating entities are arranged in a collaborative relationship. Each participating entity trains a machine learning model with an encoder on a training data set. The performance of each of the models is measured and at least one of the models is selectively identified based on the measured performance. An encoder of the selectively identified machine learning model is shared with each of the participating entities. The shared encoder is configured to be applied by the participating entities to train the first and second machine learning models, which are configured to be merged and shared in the federated learning environment.

Abstract: Techniques for classifier generalization in a supervised learning process using input encoding are provided. In one aspect, a method for classification generalization includes: encoding original input features from at least one input sample {right arrow over (x)}S with a uniquely decodable code using an encoder E(?) to produce encoded input features E({right arrow over (x)}S), wherein the at least one input sample {right arrow over (x)}S comprises uncoded input features; feeding the uncoded input features and the encoded input features E({right arrow over (x)}S) to a base model to build an encoded model; and learning a classification function {tilde over (C)}E(?) using the encoded model, wherein the classification function {tilde over (C)}E(?) learned using the encoded model is more general than that learned using the uncoded input features alone.

Abstract: Methods and apparatus are provided for calculating branch metrics, associated with possible transitions between states of a trellis, in a sequence detector for detecting symbol values corresponding to samples of an analog signal transmitted over a channel. For each sample and each transition, the method calculates a plurality of distance values indicative of distance between that sample and respective hypothesized sample values for that transition. In parallel with calculation of the distance values, the sample is compared with a set of thresholds, each defined between a pair of successive hypothesized symbol values arranged in value order, to produce a comparison result. An optimum distance value is selected as a branch metric for the transition in dependence on the comparison result.

Abstract: Sequence detectors and detection methods are provided for detecting symbol values corresponding to a sequence of input samples obtained from an ISI channel. The sequence detector comprises a branch metric unit (BMU) and a path metric unit (PMU). The BMU, which comprises an initial set of pipeline stages, is adapted to calculate, for each input sample, branch metrics for respective possible transitions between states of a trellis. To calculate these branch metrics, the BMU selects hypothesized input values, each dependent on a possible symbol value for the input sample and L>0 previous symbol values corresponding to possible transitions between states of the trellis. The BMU then calculates differences between the input sample and each hypothesized input value. The BMU compares these differences and selects, as the branch metric for each possible transition, an optimum difference in dependence on a predetermined state in a survivor path through the trellis.

Abstract: Methods and apparatus are provided for calculating branch metrics, associated with possible transitions between states of a trellis, in a sequence detector for detecting symbol values corresponding to samples of an analog signal transmitted over a channel. For each sample and each transition, the method calculates a plurality of distance values indicative of distance between that sample and respective hypothesized sample values for that transition. In parallel with calculation of the distance values, the sample is compared with a set of thresholds, each defined between a pair of successive hypothesized symbol values arranged in value order, to produce a comparison result. An optimum distance value is selected as a branch metric for the transition in dependence on the comparison result.

Abstract: A method for timing recovery for a high-speed data transmission system may be provided. The method comprises receiving an analog input signal at an ADC and passing processed digital signal samples to a Viterbi detector. The method also comprises receiving at least one processed signal sample and at least two sets of at least one candidate symbol each from the Viterbi detector and/or the processed signal samples by timing error detectors and forwarding output digital signals of the timing error detectors via loop filters to related multiplexers. Furthermore, the method comprises selecting one digital signal from each of the multiplexers using a select signal generated by the Viterbi detector, and deriving a control signal controlling a sampling clock of the analog-to-digital converter by at least one of the selected digital signals from the multiplexers.

Abstract: Sequence detectors and detection methods are provided for detecting symbol values corresponding to a sequence of input samples obtained from an ISI channel. The sequence detector comprises a branch metric unit (BMU) and a path metric unit (PMU). The BMU, which comprises an initial set of pipeline stages, is adapted to calculate, for each input sample, branch metrics for respective possible transitions between states of a trellis. To calculate these branch metrics, the BMU selects hypothesized input values, each dependent on a possible symbol value for the input sample and L>0 previous symbol values corresponding to possible transitions between states of the trellis. The BMU then calculates differences between the input sample and each hypothesized input value. The BMU compares these differences and selects, as the branch metric for each possible transition, an optimum difference in dependence on a predetermined state in a survivor path through the trellis.

Abstract: Calculating path metrics, associated with respective states of an n-state trellis, by accumulating branch metrics in a sequence detector. Each path metric is represented by N bits plus a wrap-around bit for indicating wrap-around of the N-bit value of that path metric.

Abstract: A decision-feedback equalizer for use in a receiving unit of an incoming data stream and for providing a stream of bit data outputs includes a number of comparators configured to perform a comparison related to a number of threshold values and related to a digitalized data sample and to obtain a comparison result; at least one correction block configured to receive the comparison result of a respective one of the comparators and to generate a plurality of intermediate results; and a multiplexer configured to select from the set of intermediate results depending on the output data history to provide the stream of bit data outputs.

Abstract: A decision-feedback equalizer for use in a receiving unit of an incoming data stream and for providing a stream of bit data outputs includes a number of comparators configured to perform a comparison related to a number of threshold values and related to a digitalized data sample and to obtain a comparison result; at least one correction block configured to receive the comparison result of a respective one of the comparators and to generate a plurality of intermediate results; and a multiplexer configured to select from the set of intermediate results depending on the output data history to provide the stream of bit data outputs.

Abstract: A decision-feedback equalizer for use in a receiving unit of an incoming data stream and for providing a stream of bit data outputs includes a number of comparators configured to perform a comparison related to a number of threshold values and related to a digitalized data sample and to obtain a comparison result; at least one correction block configured to receive the comparison result of a respective one of the comparators and to generate a plurality of intermediate results; and a multiplexer configured to select from the set of intermediate results depending on the output data history to provide the stream of bit data outputs.

Abstract: A decision-feedback equalizer for use in a receiving unit of an incoming data stream and for providing a stream of bit data outputs includes a number of comparators configured to perform a comparison related to a number of threshold values and related to a digitalized data sample and to obtain a comparison result; at least one correction block configured to receive the comparison result of a respective one of the comparators and to generate a plurality of intermediate results; and a multiplexer configured to select from the set of intermediate results depending on the output data history to provide the stream of bit data outputs.