Abstract: A root behavior tree is stored with a flow control node and a plurality of child nodes. The flow control node may determine a sequence of execution of child nodes. A first-level behavior tree is stored and executed to accomplish a task with associated intent. The first-level tree comprises a child node of the root tree and nodes defining actions. A dialog between contact center and user is hosted over a communication channel and a processor receives inputs from the dialog and executes the root tree and the first-level tree as a child node of the root tree in response. Upon detecting a match between an input and a data pattern, the execution sequence within the root tree or a further tree is altered. Responsive to the altered sequence, an output is provided to cause an action.

Abstract: A computer-implemented method for managing a dialog between a contact center system and a user thereof, comprising the steps of: hosting a dialog over a communication channel between an automated dialog engine of said contact center and said user thereof, said dialog comprising messages sent between said automated dialog engine and said user in both directions; said automated dialog engine receiving input messages from said user, and determining response messages in response to said inputs; detecting from said messages of said dialog a trigger event matching a rule; in response to detection of said trigger event: providing said agent station with a summary of said dialog; and providing control of said automated dialog engine to said agent station.

Abstract: A processor receives inputs from a dialog between an agent and a user performed over a communication channel A knowledge base is stored, comprising entries with a subset of said entries defined as higher priority. A match is detected between an input from said dialog and a plurality of said knowledge base entries. At least one of said plurality of entries is retrieved corresponding to said match. The at least one of said plurality of entries is pushed as an output to one or more of a device operated by said user and at least one device operated by one of said user and said agent. A further match is detected between an input from said dialog and a higher priority entry in said knowledge base and pushed to said at least one device while the dialog is ongoing.

Abstract: A system and method are presented for the correction of packet loss in audio in automatic speech recognition (ASR) systems. Packet loss correction, as presented herein, occurs at the recognition stage without modifying any of the acoustic models generated during training. The behavior of the ASR engine in the absence of packet loss is thus not altered. To accomplish this, the actual input signal may be rectified, the recognition scores may be normalized to account for signal errors, and a best-estimate method using information from previous frames and acoustic models may be used to replace the noisy signal.

Abstract: A system and method are presented for customer journey event representation learning and outcome prediction using neural sequence models. A plurality of events are input into a module where each event has a schema comprising characteristics of the events and their modalities (web clicks, calls, emails, chats, etc.). The events of different modalities can be captured using different schemas and therefore embodiments described herein are schema-agnostic. Each event is represented as a vector of some number of numbers by the module with a plurality of vectors being generated in total for each customer visit. The vectors are then used in sequence learning to predict real-time next best actions or outcome probabilities in a customer journey using machine learning algorithms such as recurrent neural networks.

Abstract: A system is presented for load balancing across media server instances. In an embodiment, media is broken out into a multi-tenanted service allowing the media to be scaled independently of the number of organizations supported on a cloud-based collaboration platform. Scaling may occur in a scaling-out or a scaling-in operation. States for a media service may comprise in-service, quiescing, quiesced, failed, etc. The states may be used to monitor sessions associated with an instance and determine which media instances to terminate during a scaling-in operation. In an embodiment, new instances may be added to a collection of media instances in response to an increased workload in a scaling-out operation.

Abstract: A system and method are presented for the correction of packet loss in audio in automatic speech recognition (ASR) systems. Packet loss correction, as presented herein, occurs at the recognition stage without modifying any of the acoustic models generated during training. The behavior of the ASR engine in the absence of packet loss is thus not altered. To accomplish this, the actual input signal may be rectified, the recognition scores may be normalized to account for signal errors, and a best-estimate method using information from previous frames and acoustic models may be used to replace the noisy signal.

Abstract: A system and method are presented for the synthesis of speech from provided text. Particularly, the generation of parameters within the system is performed as a continuous approximation in order to mimic the natural flow of speech as opposed to a step-wise approximation of the feature stream. Provided text may be partitioned and parameters generated using a speech model. The generated parameters from the speech model may then be used in a post-processing step to obtain a new set of parameters for application in speech synthesis.

Abstract: A system and method are presented for load balancing across media server instances. In an embodiment, media is broken out into a multi-tenanted service allowing the media to be scaled independently of the number of organizations supported on a cloud-based collaboration platform. Scaling may occur in a scaling-out or a scaling-in operation. States for a media service may comprise in-service, quiescing, quiesced, failed, etc. The states may be used to monitor sessions associated with an instance and determine which media instances to terminate during a scaling-in operation. In an embodiment, new instances may be added to a collection of media instances in response to an increased workload in a scaling-out operation.

Abstract: A system and method are presented for model derivation for entity prediction. An LSTM with 100 memory cells is used in the system architecture. Sentences are truncated and provided with feature information to a named-entity recognition model. A forward and a backward pass of the LSTM are performed, and each pass is concatenated. The concatenated bi-directional LSTM encodings are obtained for the various features for each word. A fully connected set of neurons shared across all encoded words is obtained and the final encoded outputs with dimensions equal to the number of entities is determined.

Abstract: Technologies for authenticating a speaker in a voice authentication system using voice biometrics include a speech collection computing device and a speech authentication computing device. The speech collection computing device is configured to collect a speech signal from a speaker and transmit the speech signal to the speech authentication computing device. The speech authentication computing device is configured to compute a speech signal feature vector for the received speech signal, retrieve a speech signal classifier associated with the speaker, and feed the speech signal feature vector to the retrieved speech signal classifier. Additionally, the speech authentication computing device is configured to determine whether the speaker is an authorized speaker based on an output of the retrieved speech signal classifier. Additional embodiments are described herein.

Abstract: Systems and methods for the matching of datasets, such as input audio segments, with known datasets in a database are disclosed. In an illustrative embodiment, the use of the presently disclosed systems and methods is described in conjunction with recognizing known network message recordings encountered during an outbound telephone call. The methodologies include creation of a ternary fingerprint bitmap to make the comparison process more efficient. Also disclosed are automated methodologies for creating the database of known datasets from a larger collection of datasets.

Abstract: A method includes: loading, by a processor, a grammar specification defining at least one parameterizable grammar including a plurality of rules; setting, by the processor, an initial state of a grammar processor as a current state, the current state including parameters supplied to the rules; selecting, by the processor, a rule of the plurality of rules matching the parameters of the current state of the grammar processor; applying, by the processor, the selected rule to the audio and updating the current state; determining, by the processor, whether termination conditions have been met; in response to determining the termination conditions are not met, selecting, by the processor, from the plurality of rules in accordance with parameters of the updated state; and in response to determining the termination conditions are met, outputting, by the processor, a recognizer result of the current state.

Abstract: A communication system including a media server through which communication packets are exchanged for recording and monitoring purposes is disclosed. A tap is associated with each communication endpoint allowing for cradle to grave recording of communications despite their subsequent routing or branching. An incoming communication is routed to a first tap and upon selection of a receiving party; the first tap is routed to a second tap which forwards communication packets on to the receiving party. The taps may be used to forward communication packets to any number of other taps or destinations, such as a recording device, monitoring user, or other user in the form of a conference.

Abstract: A system and method are presented for on premises and offline survivability of an interactive voice response system in a cloud telephony system. Voice interaction control may be divided from the media resources. Survivability is invoked when the communication technology between the Cloud and the voice interaction's resource provider is degraded or disrupted. The system is capable of recovering after a disruption event such that a seamless transition between failure and non-failure states is provided for a limited impact to a user's experience. When communication paths or Cloud control is re-established, the user resumes normal processing and full functionality as if the failure had not occurred.

Abstract: A system and method are presented for optimization of audio fingerprint search. In an embodiment, the audio fingerprints are organized into a recursive tree with different branches containing fingerprint sets that are dissimilar to each other. The tree is constructed using a clustering algorithm based on a similarity measure. The similarity measure may comprise a Hamming distance for a binary fingerprint or a Euclidean distance for continuous valued fingerprints. In another embodiment, each fingerprint is stored at a plurality of resolutions and clustering is performed hierarchically. The recognition of an incoming fingerprint begins from the root of the tree and proceeds down its branches until a match or mismatch is declared. In yet another embodiment, a fingerprint definition is generalized to include more detailed audio information than in the previous definition.

Abstract: A system and method are presented for predicting speech recognition performance using accuracy scores in speech recognition systems within the speech analytics field. A keyword set is selected. Figure of Merit (FOM) is computed for the keyword set. Relevant features that describe the word individually and in relation to other words in the language are computed. A mapping from these features to FOM is learned. This mapping can be generalized via a suitable machine learning algorithm and be used to predict FOM for a new keyword. In at least embodiment, the predicted FOM may be used to adjust internals of speech recognition engine to achieve a consistent behavior for all inputs for various settings of confidence values.

Abstract: A system and method are presented for optimization of audio fingerprint search. In an embodiment, the audio fingerprints are organized into a recursive tree with different branches containing fingerprint sets that are dissimilar to each other. The tree is constructed using a clustering algorithm based on a similarity measure. The similarity measure may comprise a Hamming distance for a binary fingerprint or a Euclidean distance for continuous valued fingerprints. In another embodiment, each fingerprint is stored at a plurality of resolutions and clustering is performed hierarchically. The recognition of an incoming fingerprint begins from the root of the tree and proceeds down its branches until a match or mismatch is declared. In yet another embodiment, a fingerprint definition is generalized to include more detailed audio information than in the previous definition.

Abstract: A system and method are presented for the encoding of participants in a conference setting. In an embodiment, audio from conference participants in a voice-over-IP setting may be received and processed by the system. In an embodiment, audio may be received in a compressed form and de-compressed for processing. For each participant, return audio is generated, compressed (if applicable) and transmitted to the participant. The system may recognize when participants are using the same audio encoding format and are thus receiving audio that may be similar or identical. The audio may only be encoded once instead of for each participant. Thus, redundant encodings are recognized and eliminated resulting in less CPU usage.

Abstract: A system and method are presented for the correction of packet loss in audio in automatic speech recognition (ASR) systems. Packet loss correction, as presented herein, occurs at the recognition stage without modifying any of the acoustic models generated during training. The behavior of the ASR engine in the absence of packet loss is thus not altered. To accomplish this, the actual input signal may be rectified, the recognition scores may be normalized to account for signal errors, and a best-estimate method using information from previous frames and acoustic models may be used to replace the noisy signal.