Abstract: Examples of the disclosure relate to apparatus, methods and computer programs for audio signal enhancement using a dataset for a target use case. In examples of the disclosure an apparatus is configured to enable access to a trained computer program. The trained computer program is configured for processing one or more audio signals to enhance audibility of sounds within the one or more audio signals. The trained computer program is trained using a generic dataset. The apparatus is also configured to obtain a dataset. The dataset includes data samples with inputs and outputs for the computer program. The apparatus is configured to use the dataset to update the trained computer program.

Abstract: An apparatus is disclosed, which may include: at least one processor; and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to perform: receiving downlink channel estimates between a plurality of access point (AP) antennas of one or more APs and a plurality of user equipment (UE) antennas of one or more UEs; iteratively updating a set of power allocation weights for the AP antennas using weights extracted from a precoder; computing the precoder based on the received channel estimates and a previous set of power allocation weights; performing the iterative updates until a stopping criterion is reached; and outputting data representative of a precoder computed from the received channel estimates and the updated set of power allocation weights.

Abstract: An apparatus is disclosed, which may include: at least one processor; and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to perform: receiving downlink channel estimates between a plurality of access point (AP) antennas of one or more APs and a plurality of user equipment (UE) antennas of one or more UEs; iteratively updating a set of power allocation weights for the AP antennas using weights extracted from a precoder; computing the precoder based on the received channel estimates and a previous set of power allocation weights; performing the iterative updates until a stopping criterion is reached; and outputting data representative of a precoder computed from the received channel estimates and the updated set of power allocation weights.

Abstract: Embodiments relate to an apparatus comprising means configured for: obtaining echo response data representative of the echo response of a communication line, wherein the echo response data specifies the echo response based on two dimensions and includes first dimension data and second dimension data, determining at least one property of the communication line based on processing the echo response data with a neural network, wherein the neural network comprises at least: a first convolutional branch for processing the first dimension data, a second convolutional branch for processing the second dimension data, a dense part for processing the outputs of the first and second convolutional branches.

Abstract: Example embodiments describe means (200) for performing i) pre-compensating (210, N sets of K1 tone data values (220) for crosstalk between N communication lines; the N sets of K1 tone data values pertaining to respective N terminal nodes of a digital communication system; ii) calculating (215) from the pre-compensated N sets of K1 tone data values (221) N sets of first time domain symbols (225); iii) calculating (283) a second time domain symbol (284) from a set of K2 tones values (280); the K2 tone data values pertaining to a selected one of the N terminal nodes; and iv) adding (212) the second time domain symbol in a weighted manner to the first time domain symbols such that the second time domain symbol is added to the first time domain symbol for the selected terminal node and to at least one other of the first time domain symbols for the respective other terminal nodes.

Abstract: The present disclosure relates to an equalizer training unit for deriving equalization parameters for compensating data-dependent distortion in received samples by use of a training sequence including a sequence p>1 times and cyclically comprising N sub-sequences of respective combinations of L time-domain symbols of a modulation scheme, wherein the N sub-sequences are cyclically arranged in a selected order and such that L?1 symbols of a respective sub-sequence overlap with symbols in the preceding and following sub-sequences. The present disclosure further relates to a training sequence generator unit for generative the training sequence and an equalizer employing the equalizer training unit.

Abstract: Examples of the disclosure enable tuning of the noise suppression in audio signals such as microphone captured signals. In examples of the disclosure, a machine learning program can be used to obtain outputs for a plurality of different frequency bands. One or more tuning parameters can also be obtained. The outputs can be processed to determine at least one uncertainty value and a gain coefficient for the plurality of different frequency bands. The at least one uncertainty value provides a measure of uncertainty for the gain coefficient, and the gain coefficient is adjusted by the at least one uncertainty value and the one or more tuning parameters. The adjusted gain coefficient is configured to be applied to a signal associated with at least one microphone output signal within the plurality of different frequency bands to control noise suppression for speech audibility.

Abstract: Apparatus, methods and computer programs are provided for reducing echo in microphone signals. The apparatus can be configured to obtain an echo path delay estimate and obtain information indicative of a number of future frames used by an acoustic echo cancellation filter. The apparatus can also be configured to obtain information indicative of a time shift between consecutive frames processed by the acoustic echo cancellation filter. The apparatus can also be configured to determine at least a first delay to be added to a loudspeaker signal and determine at least a second delay to be added to a microphone signal. The determination of the first delay and the second delay is based on the echo path delay estimate, the number of future frames and the time shift between consecutive frames processed by the acoustic echo cancellation filter.

Abstract: A wired communication systems, a subscriber device and a method, and more specifically related to initialization and/or update of communication parameters used for communication through the wired communication systems. An access node comprises at least one processor and memory storing instructions that when executed by the at least one processor cause a local transceiver of the access node to transmit signaling data via a subscriber line to a remote transceiver, the signaling data identifying particular symbol positions within a sequence of symbols transmitted over the subscriber line during show-time as being suitable for characterizing a direct communication channel between the local transceiver and the remote transceiver.

Abstract: The present invention relates to a communication controller (131; 132) and method for controlling communications between an access node (101; 102) and a plurality of remote communication units (211; 212) coupled to the access node via at least one wired transmission medium (20; 40). At least one communication unit (111, 211; 112, 212) of the access node and of the plurality of remote communication units is configured to operate in full-duplex mode according to a first full-duplex communication profile (OP1) when using a first subset of transmission resources (TSSET1; TONESET1) selected from a whole set of transmission resources available for communication over the at least one transmission medium, and according to a second full-duplex communication profile (OP2) when using a second non-overlapping subset of transmission resources (TSSET2; TONESET2) selected from the whole set of transmission resources.

Abstract: An optical line terminal is disclosed. The optical line terminal comprises at least one processor; and at least one memory including machine-readable instructions. The at least one memory and the machine-readable instructions are configured to, with the at least one processor, cause the optical line terminal to determine based on one or more variables a relationship between bandwidth efficiency and latency for communication of contents of a queue buffer of an optical network unit with the optical line terminal via an optical distribution network, and determine a burst schedule for the queue buffer based on the determined relationship.

Abstract: A communication system in which DRA control is aided by RL. An example embodiment may control one or more buffer queues populated by downstream and/or upstream data streams. The egress rates of the buffer queues can be dynamically controlled using an RL technique, according to which a learning agent can adaptively change the state-to-action mapping function of the DRA controller while circumventing the RL exploration phase and relying on extrapolation of the already taken actions instead. This feature may result in at least two benefits: (i) cancellation of a performance penalty typically associated with RL exploration; and (ii) faster learning of the environment, as the learning agent can determine the performance metrics of many actions per state in a single occurrence of the state. In an example embodiment, the communication system may be a DSL system, a PON system, or a wireless communication system.

Abstract: The present disclosure relates to an equalizer training unit for deriving equalization parameters for compensating data-dependent distortion in received samples by use of a training sequence including a sequence p>1 times and cyclically comprising N sub-sequences of respective combinations of L time-domain symbols of a modulation scheme, wherein the N sub-sequences are cyclically arranged in a selected order and such that L?1 symbols of a respective sub-sequence overlap with symbols in the preceding and following sub-sequences. The present disclosure further relates to a training sequence generator unit for generative the training sequence and an equalizer employing the equalizer training unit.

Abstract: A vectoring controller for configuring a vectoring processor that jointly processes DMT communication signals to be transmitted over, or received from, a plurality of N subscriber lines according to a vectoring matrix. In accordance with an embodiment, the vectoring controller is configured, for given ones of a plurality of tones, to enable the given tone for direct data communication over a first set of N?Mk targeted lines out of the plurality of N subscriber lines, and to disable the given tone for direct data communication over a second disjoint set of Mk supporting lines out of the plurality of N subscriber lines, Mk denoting a non-null positive integer. The vectoring controller is further configured to configure the vectoring matrix to use an available transmit or receive power at the given tone over the second set of Mk supporting lines for further enhancement of data signal gains at the given tone over the first set of N?Mk targeted lines.

Abstract: A communication system in which DRA control is aided by RL. An example embodiment may control one or more buffer queues populated by downstream and/or upstream data streams. The egress rates of the buffer queues can be dynamically controlled using an RL technique, according to which a learning agent can adaptively change the state-to-action mapping function of the DRA controller while circumventing the RL exploration phase and relying on extrapolation of the already taken actions instead. This feature may result in at least two benefits: (i) cancellation of a performance penalty typically associated with RL exploration; and (ii) faster learning of the environment, as the learning agent can determine the performance metrics of many actions per state in a single occurrence of the state. In an example embodiment, the communication system may be a DSL system, a PON system, or a wireless communication system.

Abstract: A digital circuit for implementing a channel-tracking functionality, in which an adaptive (e.g., FIR) filter is updated based on reinforcement learning. In an example embodiment, the adaptive filter may be updated using an LMS-type algorithm. The digital circuit may also include an electronic controller configured to change the convergence coefficient of the LMS algorithm using a selection policy learned by applying a reinforcement-learning technique and based on residual errors and channel estimates received over a sequence of iterations. In some embodiments, the electronic controller may include an artificial neural network. An example embodiment of the digital circuit is advantageously capable of providing improved performance after the learning phase, e.g., for communication channels exhibiting variable dynamicity patterns, such as those associated with aerial copper cables or some wireless channels.

Abstract: A vectoring controller is configured to determine first coefficient values for a vectoring matrix at a first tone based on a first number of iterations through an iterative update algorithm and a first channel matrix estimate at the first tone, and to determine second coefficient values for the vectoring matrix at a second neighboring tone based on a second number of iterations through the iterative update algorithm and a second channel matrix estimate at the second tone. The vectoring controller is configured to start with the first coefficient values as initial values for the respective second coefficient values in the iterative update algorithm. The second number of iterations is lower than or equal to the first number of iterations.

Abstract: A digital circuit for implementing a channel-tracking functionality, in which an adaptive (e.g., FIR) filter is updated based on reinforcement learning. In an example embodiment, the adaptive filter may be updated using an LMS-type algorithm. The digital circuit may also include an electronic controller configured to change the convergence coefficient of the LMS algorithm using a selection policy learned by applying a reinforcement-learning technique and based on residual errors and channel estimates received over a sequence of iterations. In some embodiments, the electronic controller may include an artificial neural network. An example embodiment of the digital circuit is advantageously capable of providing improved performance after the learning phase, e.g., for communication channels exhibiting variable dynamicity patterns, such as those associated with aerial copper cables or some wireless channels.

Abstract: A digital circuit for implementing a channel-tracking functionality, in which an adaptive (e.g., FIR) filter is updated based on reinforcement learning. In an example embodiment, the adaptive filter may be updated using an LMS-type algorithm. The digital circuit may also include an electronic controller configured to change the convergence coefficient of the LMS algorithm using a selection policy learned by applying a reinforcement-learning technique and based on residual errors and channel estimates received over a sequence of iterations. In some embodiments, the electronic controller may include an artificial neural network. An example embodiment of the digital circuit is advantageously capable of providing improved performance after the learning phase, e.g., for communication channels exhibiting variable dynamicity patterns, such as those associated with aerial copper cables or some wireless channels.

Abstract: Embodiments relate to an apparatus comprising means configured for: obtaining echo response data representative of the echo response of a communication line, wherein the echo response data specifies the echo response based on two dimensions and includes first dimension data and second dimension data, determining at least one property of the communication line based on processing the echo response data with a neural network, wherein the neural network comprises at least: a first convolutional branch for processing the first dimension data, a second convolutional branch for processing the second dimension data, a dense part for processing the outputs of the first and second convolutional branches.