# US 2018/0322403 A1 · Dynamic response prediction for improved bot task processing > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2018/0322403 A1 - Document ID: US20180322403A1 - Status: Published application - Kind: A1 - Application number: 15/971,708 - Filed: 2018-05-04 - Published: 2018-11-08 - Earliest claimed: 2017-05-05 - Assignees: Liveperson, Inc. - Applicants: Liveperson, Inc. - Inventors: Ofer Ron; Eran Vanounou; Gali Konky; Seetvun Amir - Web page: https://patentgrep.com/patent/US20180322403A1 ## CPC - G06N 5/04: Inference or reasoning models - G06N 3/044: Recurrent networks, e.g. Hopfield networks - G06N 3/0464: Convolutional networks [CNN, ConvNet] - G06N 3/08: Learning methods - G06N 3/09: Supervised learning - G06N 20/00: Machine learning - G06Q 10/10: Office automation; Time management - G06Q 10/107: Computer-aided management of electronic mailing [e-mailing] - H04L 51/02: using automatic reactions or user delegation, e.g. automatic replies or chatbot-generated messages - H04L 51/216: Handling conversation history, e.g. grouping of messages in sessions or threads - H04L 67/02: based on web technology, e.g. hypertext transfer protocol [HTTP] - H04L 67/63: Routing a service request depending on the request content or context - G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks - G06N 20/10: using kernel methods, e.g. support vector machines [SVM] - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS - G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR - H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M) ## Abstract Systems and methods can be provided for predicting responses during communication sessions with network devices. In some implementations, systems and methods can facilitate predicting responses using machine learning techniques. Messages received through a platform can be stored in a repository. A machine learning model may be trained using the stored messages. When a terminal device is communicating with a network device in a communication session, the messages exchanged in the communication session and the machine learning model can be used to predict future responses in real-time. The predicted future responses can be presented at the terminal device. A predicted response can be selected at the terminal device. Upon selection, the selected predicted response is transmitted to the network device during the communication session. ## Claims ### Claim 1 A computer-implemented method, comprising: collecting a data set for training a machine-learning model to predict response messages, wherein collecting the data set includes storing one or more previous messages included in a previous communication session between a network device and a terminal device associated with an agent; facilitating a communication session between a terminal device and a network device, wherein the communication session enables the terminal device and the network device to exchange one or more messages; receiving a new message during the communication session; evaluating the new message using the trained machine-learning model, wherein evaluating the new message includes evaluating any messages exchanged before the new message was received; predicting a response to the new message, wherein predicting the response includes using a result of the evaluation to determine which previous message to select from the data set as the predicted response to the new message; and facilitating displaying the predicted response at the terminal device, wherein when the predicted response is selected, the predicted response is automatically transmitted to the network device during the communication session. ### Claim 2 (depends on 1) The computer-implemented method of claim 1, further comprising: identifying an attribute associated with a previous message included in the data set; appending an attribute tag to the previous message, wherein the attribute tag corresponds to the identified attribute; and training the machine-learning model using the attribute tag. ### Claim 3 (depends on 1) The computer-implemented method of claim 1, wherein the communication session between the terminal device and the network device includes a bot, wherein the bot is configured to autonomously communicate with the network device. ### Claim 4 (depends on 3) The computer-implemented method of claim 3, further comprising: evaluating the new message using the bot, wherein evaluating the new message includes automatically identifying a query from the new message, and wherein the new message is evaluated in real-time as the new message is received; querying a database using the identified query; and displaying one or more results of the querying at the terminal device. ### Claim 5 (depends on 1) The computer-implemented method of claim 1, wherein the machine-learning model includes a recurrent neural network (RNN) or a convolutional neural network (CNN). ### Claim 6 (depends on 1) The computer-implemented method of claim 1, wherein determining the predicted response message further comprises: generating a vector representation of the new message; using the machine-learning model to compare the vector representation of the new message to an additional vector representation of the one or more previous messages included in the data set; and selecting a previous message from the one or more previous messages included in the data set, wherein the previous message is selected based on the comparison. ### Claim 7 (depends on 1) The computer-implemented method of claim 1, further comprising: generating a matrix for a candidate message, wherein the matrix includes a number of words included in the previous message and a vector representation of the previous message; generating an additional matrix for the new message, wherein the additional matrix includes a number of words included in the new message and a vector representation of the new message; and comparing the matrix with the additional matrix to determine whether to select the previous message as the predicted response. ### Claim 8 A system, comprising: one or more data processors; and a non-transitory computer-readable storage medium containing instructions which, when executed on the one or more data processors, cause the one or more data processors to perform operations including: collecting a data set for training a machine-learning model to predict response messages, wherein collecting the data set includes storing one or more previous messages included in a previous communication session between a network device and a terminal device associated with an agent; facilitating a communication session between a terminal device and a network device, wherein the communication session enables the terminal device and the network device to exchange one or more messages; receiving a new message during the communication session; evaluating the new message using the trained machine-learning model, wherein evaluating the new message includes evaluating any messages exchanged before the new message was received; predicting a response to the new message, wherein predicting the response includes using a result of the evaluation to determine which previous message to select from the data set as the predicted response to the new message; and facilitating displaying the predicted response at the terminal device, wherein when the predicted response is selected, the predicted response is automatically transmitted to the network device during the communication session. ### Claim 9 (depends on 8) The system of claim 8, wherein the operations further comprise: identifying an attribute associated with a previous message included in the data set; appending an attribute tag to the previous message, wherein the attribute tag corresponds to the identified attribute; and training the machine-learning model using the attribute tag. ### Claim 10 (depends on 8) The system of claim 8, wherein the communication session between the terminal device and the network device includes a bot, wherein the bot is configured to autonomously communicate with the network device. ### Claim 11 (depends on 10) The system of claim 10, further comprising: evaluating the new message using the bot, wherein evaluating the new message includes automatically identifying a query from the new message, and wherein the new message is evaluated in real-time as the new message is received; querying a database using the identified query; and displaying one or more results of the querying at the terminal device. ### Claim 12 (depends on 8) The system of claim 8, wherein the machine-learning model includes a recurrent neural network (RNN) or a convolutional neural network (CNN). ### Claim 13 (depends on 8) The system of claim 8, wherein predicting the response further comprises: generating a vector representation of the new message; using the machine-learning model to compare the vector representation of the new message to an additional vector representation of the one or more previous messages included in the data set; and selecting a previous message from the one or more previous messages included in the data set, wherein the previous message is selected based on the comparison. ### Claim 14 (depends on 8) The system of claim 8, wherein the operations further comprise: generating a matrix for a candidate message, wherein the matrix includes a number of words included in the previous message and a vector representation of the previous message; generating an additional matrix for the new message, wherein the additional matrix includes a number of words included in the new message and a vector representation of the new message; and comparing the matrix with the additional matrix to determine whether to select the previous message as the predicted response. ### Claim 15 A computer-program product tangibly embodied in a non-transitory machine-readable storage medium, including instructions configured to cause a data processing apparatus to perform operations including: collecting a data set for training a machine-learning model to predict response messages, wherein collecting the data set includes storing one or more previous messages included in a previous communication session between a network device and a terminal device associated with an agent; facilitating a communication session between a terminal device and a network device, wherein the communication session enables the terminal device and the network device to exchange one or more messages; receiving a new message during the communication session; evaluating the new message using the trained machine-learning model, wherein evaluating the new message includes evaluating any messages exchanged before the new message was received; predicting a response to the new message, wherein predicting the response includes using a result of the evaluation to determine which previous message to select from the data set as the predicted response to the new message; and facilitating displaying the predicted response at the terminal device, wherein when the predicted response is selected, the predicted response is automatically transmitted to the network device during the communication session. ### Claim 16 (depends on 15) The computer-program product of claim 15, wherein the operations further comprise: identifying an attribute associated with a previous message included in the data set; appending an attribute tag to the previous message, wherein the attribute tag corresponds to the identified attribute; and training the machine-learning model using the attribute tag. ### Claim 17 (depends on 15) The computer-program product of claim 15, wherein the communication session between the terminal device and the network device includes a bot, wherein the bot is configured to autonomously communicate with the network device. ### Claim 18 (depends on 17) The computer-program product of claim 17, further comprising: evaluating the new message using the bot, wherein evaluating the new message includes automatically identifying a query from the new message, and wherein the new message is evaluated in real-time as the new message is received; querying a database using the identified query; and displaying one or more results of the querying at the terminal device. ### Claim 19 (depends on 15) The computer-program product of claim 15, wherein the machine-learning model includes a recurrent neural network (RNN) or a convolutional neural network (CNN). ### Claim 20 (depends on 15) The computer-program product of claim 15, wherein predicting the response further comprises: generating a vector representation of the new message; using the machine-learning model to compare the vector representation of the new message to an additional vector representation of the one or more previous messages included in the data set; and selecting a previous message from the one or more previous messages included in the data set, wherein the previous message is selected based on the comparison.