# US 12,112,841 B2 · Systems and methods for autoregressive recurrent neural networks for identifying actionable vital alerts > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,112,841 B2 - Document ID: US12112841B2 - Status: Granted - Kind: B2 - Application number: 18/375,006 - Filed: 2023-09-29 - Published: 2024-08-22 - Granted: 2024-10-08 - Earliest claimed: 2023-02-17 - Assignees: Cadence Solutions, Inc. - Applicants: Cadence Solutions, Inc. - Inventors: Ashwyn Sharma - Web page: https://patentgrep.com/patent/US12112841B2 ## CPC - G16H 10/60: for patient-specific data, e.g. for electronic patient records - G06N 3/0442: characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] - G06N 3/0455: Auto-encoder networks; Encoder-decoder networks - G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems - G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract An emerging service in the medical industry is to provide high quality remote care to patients by remotely monitoring patient vital information. In some instances, patient vitals information is collected at a much higher frequency in comparison to a traditional clinical environment. This frequent influx of patient health information can result in a considerable amount of health-related noise that health computing systems and clinicians must evaluate. The instant systems and methods leverage autoregressive recurrent neural networks and patient embeddings to predict the likelihood of needing to address certain patient information, thereby reducing the amount of health-related noise and to enable health computing systems and clinicians to use their resources more effectively. ## Claims ### Claim 1 A system comprising: a server comprising one or more processors; and a non-transitory memory, in communication with the server, storing instructions that when executed by the one or more processors, cause the one or more processors to implement a method comprising: receiving one or more features as input at one or more encoders, wherein the one or more features are associated with patient data; converting, by the one or more encoders, the one or more features into one or more latent embeddings, wherein each of the one or more encoders converts a corresponding feature of the one or more features to generate an embedding for the corresponding feature; concatenating the one or more latent embeddings into a singular patient embedding; receiving the singular patient embedding at at least one decoder; and reconstructing, by the at least one decoder, the one or more features based on the singular patient embedding; wherein the one or more reconstructed features are used by a neural network for generating a prediction of triggering an actionable alert corresponding to the one or more features associated with the patient data. ### Claim 2 (depends on 1) The system of claim 1, wherein the one or more features includes information related to one or more of: patient vitals information, patient medicine information, patient symptoms information, notes, patient biographical information, or patient demographic information. ### Claim 3 (depends on 1) The system of claim 1, wherein each of the one or more encoders are configured to respectively encode a data type associated with each of the one or more features. ### Claim 4 (depends on 1) The system of claim 1, wherein the singular patient embedding is representative of a holistic representation of a patient's state at a specific point in time. ### Claim 5 (depends on 1) The system of claim 1, further comprising reconstructing, by the at least one decoder, patient vitals information based on the singular patient embedding. ### Claim 6 (depends on 1) The system of claim 1, wherein reconstructing the one or more features further comprises reconstructing patient vitals information. ### Claim 7 (depends on 6) The system of claim 6, further comprising generating a reconstruction error, wherein a value associated with the reconstruction error is a difference between a first vector value associated with the one or more latent embeddings and a second vector value associated with the reconstructed vitals information. ### Claim 8 A computer-implemented method comprising: receiving one or more features as input at one or more encoders, wherein the one or more features are associated with patient data; converting, by the one or more encoders, the one or more features into one or more latent embeddings, wherein each of the one or more encoders converts a corresponding feature of the one or more features to generate an embedding for the corresponding feature; concatenating the one or more latent embeddings into a singular patient embedding; receiving the singular patient embedding at at least one decoder; and reconstructing, by the at least one decoder, the one or more features based on the singular patient embedding; wherein the one or more reconstructed features are used by a neural network for generating a prediction of triggering an actionable alert corresponding to the one or more features associated with the patient data. ### Claim 9 (depends on 8) The computer-implemented method of claim 8, wherein the one or more features includes information related to one or more of: patient vitals information, patient medicine information, patient symptoms information, notes, patient biographical information, or patient demographic information. ### Claim 10 (depends on 8) The computer-implemented method of claim 8, wherein each of the one or more encoders are configured to respectively encode a data type associated with each of the one or more features. ### Claim 11 (depends on 8) The computer-implemented method of claim 8, wherein the singular patient embedding is representative of a holistic representation of a patient's state at a specific point in time. ### Claim 12 (depends on 8) The computer-implemented method of claim 8, further comprising reconstructing, by the at least one decoder, patient vitals information based on the singular patient embedding. ### Claim 13 (depends on 8) The computer-implemented method of claim 8, wherein reconstructing the one or more features further comprises reconstructing patient vitals information. ### Claim 14 (depends on 13) The computer-implemented method of claim 13, further comprising generating a reconstruction error, wherein a value associated with the reconstruction error is a difference between a first vector value associated with the one or more latent embeddings and a second vector value associated with the reconstructed vitals information. ### Claim 15 A non-transitory computer-readable medium storing instructions, that when executed by one or more processors, cause the one or more processors to implement the instructions for: receiving one or more features as input at one or more encoders, wherein the one or more features are associated with patient data; converting, by the one or more encoders, the one or more features into one or more latent embeddings, wherein each of the one or more encoders converts a corresponding feature of the one or more features to generate an embedding for the corresponding feature; concatenating the one or more latent embeddings into a singular patient embedding; receiving the singular patient embedding at at least one decoder; and reconstructing, by the at least one decoder, the one or more features based on the singular patient embedding; wherein the one or more reconstructed features are used by a neural network for generating a prediction of triggering an actionable alert corresponding to the one or more features associated with the patient data. ### Claim 16 (depends on 15) The non-transitory computer-readable medium of claim 15, wherein the one or more features includes information related to one or more of: patient vitals information, patient medicine information, patient symptoms information, notes, patient biographical information, or patient demographic information. ### Claim 17 (depends on 15) The non-transitory computer-readable medium of claim 15, wherein each of the one or more encoders are configured to respectively encode a data type associated with each of the one or more features. ### Claim 18 (depends on 15) The non-transitory computer-readable medium of claim 15, wherein the singular patient embedding is representative of a holistic representation of a patient's state at a specific point in time. ### Claim 19 (depends on 15) The non-transitory computer-readable medium of claim 15, further comprising reconstructing, by the at least one decoder, patient vitals information based on the singular patient embedding. ### Claim 20 (depends on 15) The non-transitory computer-readable medium of claim 15, wherein reconstructing the one or more features further comprises reconstructing patient vitals information; and generating a reconstruction error, wherein a value associated with the reconstruction error is a difference between a first vector value associated with the one or more latent embeddings and a second vector value associated with the reconstructed vitals information.