Systems and methods for hyper-personalized, evidence-based sleep product, service, and intervention recommendation

The system addresses the lack of personalization in sleep recommendations by using AI to analyze user-specific sleep data and contextual information, offering adaptive, evidence-based interventions that enhance sleep quality.

WO2026236760A1PCT designated stage Publication Date: 2026-11-12SLEEPSCORE INC
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
PCT/US2026/027676
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2025-05-09
Filing Date
2026-05-11
Publication Date
2026-11-12

AI Technical Summary

Technical Problem

Existing sleep recommendation systems lack personalization and scientific validation, failing to account for individual sleep patterns and dynamics, leading to suboptimal sleep quality.

Method used

A hyper-personalized sleep product recommendation system that integrates longitudinal sleep data and contextual information, using AI to generate tailored recommendations based on user phenotypes, and incorporates feedback loops to refine models over time.

Benefits of technology

Provides precise, scientifically grounded recommendations that adapt to individual user needs, improving sleep quality and satisfaction through continuous learning and evidence-based interventions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure US2026027676_12112026_PF_FP_ABST
    Figure US2026027676_12112026_PF_FP_ABST
Patent Text Reader

Abstract

A sleep product recommendation system and method provides hyper-personalized and evidence-based recommendations for sleep improvement products, services, and interventions. A recommendation engine collects and analyzes longitudinal data from one or more personal devices, the longitudinal data including objective sleep architectural metrics such as total sleep time, sleep onset latency, wake after sleep onset, and sleep stage distribution, as well as lifestyle, behavioral, chronotypic and circadian, self-reported sleep, geolocation, and demographic data. Using artificial intelligence and algorithmic approaches, the engine generates a user phenotype, queries a scientific research database and an intervention database, and produces a curated set of evidence-based, ranked recommendations that are transmitted to the user. Post-intervention objective sleep data and subjective user feedback are returned to the engine to dynamically refine its predictive models, thereby closing the loop and improving the accuracy of subsequent recommendations.
Need to check novelty before this filing date? Find Prior Art

Description

Attorney Docket 022546.005WO1SYSTEMS AND METHODS FOR HYPER-PERSONALIZED, EVIDENCE-BASED SLEEP PRODUCT, SERVICE, AND INTERVENTION RECOMMENDATIONCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of and priority to U.S. Provisional Patent Application No. 63 / 802,997, entitled “Sleep Product Recommendation,” filed on May 9, 2025, the entire disclosure of which is hereby incorporated herein by reference for all purposes.TECHNICAL FIELD

[0002] The subject matter described herein relates generally to systems and methods for improving sleep quality, and more particularly to a hyper-personalized, evidence-based recommendation engine that leverages longitudinal sleep architectural data and contextual data regarding a user to provide tailored sleep product, service, and intervention recommendations.BACKGROUND

[0003] Sleep is a critical component of human health. Sleep is a cyclical process that includes a number of distinct stages, each being characterized by particular physiological and biological states of a person.

[0004] Light Sleep (N1 and N2) refers to the initial stages of sleep in which the body begins to relax, heart rate slows, and brain activity decreases. N1 is a very light, transitional sleep, while N2 involves deeper relaxation with sleep spindles and K-complexes, preparing the body for deeper sleep.Attorney Docket 022546.005WO1

[0005] Deep Sleep (N3, slow-wave sleep) refers to the restorative stage during which the body repairs tissues, strengthens the immune system, and consolidates memories. During deep sleep, brain waves slow significantly (e.g., delta waves).

[0006] REM Sleep (Rapid Eye Movement) is characterized by vivid dreams, rapid eye movements, and high brain activity similar to that of wakefulness. REM sleep is crucial for cognitive functions such as memory processing and emotional regulation.

[0007] A typical sleep period consists of multiple sleep cycles (usually four to six per night), each lasting approximately 90 minutes. Each cycle progresses through light sleep, deep sleep, and REM sleep, with deep sleep dominating earlier cycles and REM sleep becoming more prominent in later cycles.

[0008] Although sleep has been studied and measured extensively, many individuals struggle with achieving optimal sleep quality due to a lack of personalized, evidence-based solutions. Existing recommendation systems for sleep products often rely on generic or subjective data, and fail to account for individual sleep patterns, scientific validation, or dynamic feedback from the user. There is therefore a need for an innovative system that integrates comprehensive sleep data with advanced algorithmic approaches to deliver precise, scientifically grounded recommendations tailored to individual users.SUMMARY

[0009] In one aspect, a sleep product recommendation system is provided. The system includes at least one personal device configured to capture longitudinal sleep data and contextual data associated with a user; a recommendation engine in communication with the at least one personal device over a communications network, the recommendation engine including at least one processor and a non-transitory machine-readable medium storing instructions that, whenAttorney Docket 022546.005WO1 executed by the at least one processor, cause the recommendation engine to perform operations comprising: (i) receiving the longitudinal sleep data and contextual data; (ii) generating a user phenotype based on the longitudinal sleep data and contextual data; (iii) querying a scientific research database based at least in part on the user phenotype; (iv) querying an intervention database based at least in part on the user phenotype and on results returned by the scientific research database; (v) applying an artificial intelligence model to rank candidate interventions; and (vi) transmitting at least one personalized recommendation to the user; and a feedback module configured to receive post-intervention objective sleep data and self-reported feedback from the user and to update the artificial intelligence model based at least in part on the post-intervention objective sleep data and the self-reported feedback.

[0010] In another aspect, a computer-implemented method is provided. The method includes collecting longitudinal objective sleep architectural data and additional contextual data from one or more personal devices; generating a user phenotype based on the collected data; querying a scientific research database to identify evidence-based insights relevant to the user phenotype; querying an intervention database to identify candidate interventions relevant to the user phenotype and to the evidence-based insights; applying one or more artificial intelligence or machine-learning algorithms to rank the candidate interventions; transmitting at least one personalized recommendation derived from the ranked candidate interventions to the user; collecting post-intervention objective sleep data and user-reported feedback indicative of hedonic value and self-reported efficacy of the at least one personalized recommendation; and updating the one or more artificial intelligence or machine-learning algorithms based at least in part on the postintervention objective sleep data and the user-reported feedback.Attorney Docket 022546.005WO1

[0011] Tn another aspect, a non-transitory machine-readable medium is provided, the medium storing instructions which, when executed by at least one programmable processor, cause the at least one programmable processor to perform operations consistent with the methods described herein.

[0012] Implementations of the current subject matter can include, but are not limited to, methods consistent with the descriptions provided herein as well as articles that comprise a tangibly embodied machine-readable medium operable to cause one or more machines (for example, computers, etc.) to result in operations implementing one or more of the described features. Similarly, computer systems are also described that may include one or more processors and one or more memories coupled to the one or more processors. A memory, which can include a non-transitory computer-readable or machine-readable storage medium, may include, encode, store, or the like, one or more programs that cause one or more processors to perform one or more of the operations described herein. Computer-implemented methods consistent with one or more implementations of the current subject matter can be implemented by one or more data processors residing in a single computing system or multiple computing systems. Such multiple computing systems can be connected and can exchange data and / or commands or other instructions or the like via one or more connections, including but not limited to a connection over a network (e.g., the Internet, a wireless wide area network, a local area network, a wide area network, a wired network, or the like), via a direct connection between one or more of the multiple computing systems, etc.

[0013] The details of one or more variations of the subject matter described herein are set forth in the accompanying drawings and the description below. Other features and advantages of the subject matter described herein will be apparent from the description and drawings, and from the claims. While certain features of the currently disclosed subject matter are described forAttorney Docket 022546.005WO1 illustrative purposes in relation to a sleep product recommendation engine and system, it should be readily understood that such features are not intended to be limiting. The claims that follow this disclosure are intended to define the scope of the protected subject matter.BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The accompanying drawings, which are incorporated in and constitute a part of this specification, show certain aspects of the subject matter disclosed herein and, together with the description, help explain some of the principles associated with the disclosed implementations. In the drawings:

[0015] FIG. 1 is a sequence diagram illustrating the data and control flow between a user, a sleep product recommendation engine, one or more personal devices, a scientific research database, an artificial intelligence and algorithms module, and an intervention database, in accordance with one or more implementations of the present subject matter.

[0016] FIG. 2 is a schematic block diagram of a sleep product recommendation system, in accordance with one or more implementations of the present subject matter.

[0017] FIG. 3 is a schematic block diagram of internal components of a sleep product recommendation engine, in accordance with one or more implementations of the present subject matter.

[0018] FIG. 4 is a process flow diagram illustrating a method for generating a hyperpersonalized, evidence-based sleep product, service, or intervention recommendation, in accordance with one or more implementations of the present subject matter.

[0019] FIG. 5 is a process flow diagram illustrating a method for incorporating user feedback and dynamically refining recommendation logic, in accordance with one or more implementations of the present subject matter.Attorney Docket 022546.005WO1

[0020] FIG. 6 is a schematic block diagram of an example computing system that may be used to implement aspects of the subject matter described herein, in accordance with one or more implementations of the present subject matter.

[0021] When practical, similar reference numbers denote similar structures, features, or elements.DETAILED DESCRIPTION

[0022] This document describes a sleep product recommendation system (hereinafter the “system”) and a sleep product recommendation engine (hereinafter the “Engine”) designed to provide hyper-personalized recommendations for sleep architecture, including, without limitation, sleep improvement products, services, and interventions. The recommendations are based on artificial intelligence (Al) and algorithmic approaches applied to an evidence-based dataset (with proprietary and public research and monitoring data) that has been built over time. As used herein, sleep architecture refers to the structure and organization of sleep, encompassing the cyclical patterns and stages that occur during a sleep period, as well as a sleep environment (e.g., bed, room temperature, lighting level, ambient sound, and the like).

[0023] Referring now to FIG. 2, in some implementations a sleep product recommendation system 100 includes one or more personal devices 104 associated with a user 102, a communications network 106, a sleep product recommendation engine 108, a scientific research database 110, an artificial intelligence and algorithms module 112, and an intervention database 114. The one or more personal devices 104 may include, without limitation, a smartphone 104a, a wearable monitoring device 104b (such as a smart watch, ring, or band), and one or more proximal or external sensors 104c (such as a bedside monitor, mattress sensor, environmental sensor, or polysomnography device). A client application 116 executing on at least one of the personalAttorney Docket 022546.005WO1 devices 104 provides a user interface through which the user 102 logs sleep data and personal information and through which personalized recommendations are surfaced.

[0024] The personal devices 104 communicate with the Engine 108 via the communications network 106, which may include, without limitation, the Internet, a wireless wide area network, a local area network, a wide area network, a wired network, or any combination thereof. The Engine 108 is in further communication with the scientific research database 110, which stores historical and current information from previous studies, clinical research, and other scientific literature relating to sleep health, sleep interventions, and outcomes; and with the intervention database 114, which stores a catalog of candidate interventions, including, without limitation, behavioral changes (e.g., sleep hygiene practices), technological solutions (e.g., white noise machines), product offerings (e.g., mattresses, pillows, bedding, sleepwear, supplements), and clinical recommendations (e.g., cognitive behavioral therapy for insomnia). The Al and algorithms module 112 is integrated with, or in communication with, the Engine 108 and applies one or more artificial intelligence, machine-learning, deep-learning, or statistical models to data accessible to the Engine 108. A feedback channel 118 allows post-intervention feedback (both objective and subjective) from the user 102 to be returned to the Engine 108.

[0025] Referring now to FIG. 3, in some implementations the Engine 108 includes a data ingestion module 120, a user phenotyping module 122, a recommendation generator module 124, a feedback module 126, a model refinement module 128, an output or delivery module 130, and a local datastore or cache 132. The data ingestion module 120 is configured to receive longitudinal data from the personal devices 104 and to normalize and validate the received data. The user phenotyping module 122 is configured to generate a personalized profile, referred to herein as a user phenotype, based on the ingested data. The recommendation generator module 124 isAttorney Docket 022546.005WO1 configured to query the scientific research database 110 and the intervention database 114, and, in cooperation with the Al and algorithms module 112, to rank and select candidate interventions for the user 102. The output or delivery module 130 is configured to transmit one or more selected recommendations back to the user 102, preferably via the client application 116. The feedback module 126 receives post-intervention feedback, and the model refinement module 128 uses such feedback together with newly collected sleep data to update predictive models within the Al and algorithms module 112 and to update recommendation logic within the recommendation generator module 124. The local datastore or cache 132 may store user-specific data, intermediate computations, and a local copy of evidence-based insights returned by the scientific research database 110.

[0026] The Engine 108 charts the progression of sleep stages over time. Sleep architecture is derived from objective data, often collected via devices such as polysomnography devices or wearable sleep trackers. The Engine 108 is configured to process this data to generate a curated set of recommendations with high confidence in improving sleep quality, enhancing post-purchase satisfaction, and reducing product returns, among other benefits.

[0027] The Engine 108 is further configured to collect and analyze longitudinal data from one or more personal devices 104, such as smart phones, wearable monitoring devices, and external monitoring devices, where the longitudinal data includes objective sleep architectural metrics (for example, total sleep time, sleep onset latency, wake after sleep onset, and sleep stages such as light, deep, and REM sleep), as well as lifestyle, behavioral, chronotypic and circadian, self-reported sleep, geolocation, and demographic data (for example, age and sex).

[0028] Input data to the Engine 108 includes, without limitation: Objective Sleep Architectural Data (OSAD) - metrics such as total sleep time, sleep onset latency, wake after sleepAttorney Docket 022546.005WO1 onset, and sleep stage distribution (light, deep, REM); and Additional Contextual Data Lifestyle (ACDL) factors, such as self-reported sleep quality, behavior logging or tagging, chronotypic and circadian rhythms, self-reported sleep disorders, geolocation, and demographic characteristics (eg., age, sex). By integrating these diverse data points, the Engine 108 customizes recommendations to address specific sleep needs and patterns unique to each user 102.

[0029] Recommendations generated by the Engine 108 are derived from empirical data collected through sleep-improvement research studies stored in the scientific research database 110. This ensures that each suggested product, service, or intervention is supported by validated scientific evidence, distinguishing the Engine 108 from generic recommendation systems.

[0030] The Engine 108 analyzes a broad spectrum of sleep metrics to identify nuanced sleep issues that the user 102 may not explicitly recognize. For example, the Engine 108 may detect irregularities in REM sleep duration or excessive wake after sleep onset and recommend interventions to address these specific issues.

[0031] User phenotyping is the process of analyzing a wide range of individual data to create a detailed, personalized profile of the user’s sleep characteristics, behaviors, and needs. User phenotyping may involve examining metrics including, without limitation, objective sleep data (e g., total sleep time, sleep stages, sleep onset latency), lifestyle factors, self-reported sleep habits, chronotype (circadian rhythm preferences), demographics (e.g., age, sex), and other contextual data such as geolocation or reported sleep disorders. By integrating these data points, the user phenotyping module 122 identifies unique sleep patterns and potential issues — often issues of which the user may not be aware, such as subtle inefficiencies in sleep architecture. This enables the Engine 108 to tailor hyper-personalized recommendations for products, services, orAttorney Docket 022546.005WO1 interventions that address specific sleep challenges, grounded in scientific evidence and based on a database of historical information, to optimize sleep quality and overall health.

[0032] The user 102 provides feedback on the hedonic value (enjoyment), self-reported efficacy, and accuracy of sleep predictions and product recommendations. The feedback module 126 of the Engine 108 incorporates objectively measured sleep data post-intervention, and the model refinement module 128 continuously updates and refines the predictive models used by the Al and algorithms module 112. This feedback loop ensures that recommendations evolve with user experiences and with advances in sleep science.

[0033] Referring back to FIG. 1, shown is a sequence diagram representing one example interaction between the user 102, the Engine 108, the personal devices 104, the scientific research database 110, the Al and algorithms module 112, and the intervention database 114.

[0034] The system operates as a closed-loop process, starting with the user’s input (sleep data and feedback) and cycling through data analysis, scientific research, intervention selection, Al-driven refinement, and personalized recommendations. The integration of Al ensures that the system learns and improves over time, while the use of scientific research and a dedicated intervention database ensures that recommendations are evidence-based and effective. The user 102 remains at the center of the process, with the user’s data and feedback driving the system’s functionality. The following is a description of the example steps in the process.

[0035] User Logs Sleep Data and Personal Information. The process begins with the user 102, who actively logs sleep-related data (e.g., sleep duration, sleep quality, sleep patterns) and personal information (e.g., age, health conditions, lifestyle factors) through one or more personal devices 104. This step ensures that the system has a foundation of user-specific data on which to operate, which is significant for generating personalized recommendations.Attorney Docket 022546.005WO1

[0036] Personal Devices Transmit Data and Empirical Studies. The personal devices 104 collect and transmit the logged sleep data and personal information to the Engine 108 over the communications network 106. This transmission may further include aggregated data or patterns derived from the user’s device usage (e.g., trends in sleep behavior over time). This data provides the Engine 108 with raw material to analyze and process.

[0037] Engine Analyzes User Data and Research. The Engine 108 receives the transmitted data and processes it. This includes analyzing the user’s sleep data and personal information in conjunction with broader research. The Engine 108 may employ data processing techniques to identify patterns, anomalies, or specific sleep issues (e.g., insomnia, irregular sleep cycles) that may benefit from intervention. Simultaneously or sequentially, the Engine 108 prepares to query external databases to enrich its analysis with scientific insights.

[0038] Queries Scientific Research Database. The Engine 108 sends a query to the scientific research database 110, requesting relevant studies that align with the user’s data and identified sleep patterns. The scientific research database 110 contains a repository of peer-reviewed studies, clinical research, and other scientific literature related to sleep health, interventions, and outcomes.

[0039] Scientific Research Database Identifies User Phenotypes and Issues. Upon receiving the query, the scientific research database 110 processes the user’s data and matches it with relevant studies to identify specific user phenotypes (e.g., biological or behavioral traits affecting sleep, such as being a night owl or having a predisposition to sleep apnea) and sleep-related issues (e.g., difficulty falling asleep, frequent awakenings). This step leverages scientific knowledge to contextualize the user’s data, ensuring that the recommendations are grounded in evidence-based research.Attorney Docket 022546.005WO1

[0040] Engine Queries Intervention Database. With the identified phenotypes and issues, the Engine 108 queries the intervention database 114 to identify appropriate sleep interventions. The intervention database 114 contains a catalog of candidate interventions, including without limitation behavioral changes (e.g., sleep hygiene practices), technological solutions (e.g., white noise machines, smart lighting), product recommendations (e.g., mattresses, pillows, bedding), or clinical recommendations (e.g., cognitive behavioral therapy for insomnia).

[0041] Intervention Database Selects and Uses Intervention. The intervention database 114 selects interventions that best match the user’s identified phenotypes and issues, based on the scientific data provided, and then prepares the selected intervention for recommendation by formatting the intervention in a form that the Engine 108 can deliver to the user 102 (e.g., as a step-by-step plan or a device-setting adjustment).

[0042] Al and Algorithms Collect Usage Data. In parallel, the Al and algorithms module 112 collects usage data from the entire system, including how the user 102 interacts with the personal devices 104, the Engine 108, and the interventions. This data may include metrics such as how often the user 102 follows recommendations, the effectiveness of past interventions, and changes in sleep patterns over time.

[0043] Al and Algorithms Refine Predictive Models. Using the collected usage data, the Al and algorithms module 112 refines its predictive models. This may involve machine-learning techniques to improve the accuracy of future recommendations. For example, where the user 102 consistently fails to follow a certain type of intervention (e.g., avoiding screens before bed), the Al and algorithms module 112 may adjust its model to prioritize alternative interventions that the user 102 is more likely to adopt.Attorney Docket 022546.005WO1

[0044] AT and Algorithms Update Recommendation Logic. The refined predictive models are used to update the recommendation logic within the Engine 108. This ensures that the Engine’s decision-making process evolves over time, becoming more tailored to the user’s behavior and needs as more data is gathered.

[0045] Engine Suggests Personalized Interventions. The Engine 108 compiles the processed data — the user data, scientific research, selected interventions, and updated recommendation logic — to suggest personalized sleep interventions to the user 102. These suggestions are delivered back through the user’s personal devices 104, preferably presented in a user-friendly format (e.g., a notification with a recommendation to adjust bedtime routines or try a relaxation technique).

[0046] User Provides Feedback. After receiving and implementing the suggested interventions, the user 102 provides feedback to the system. The feedback may include whether the intervention was effective (e.g., improved sleep quality), any challenges faced (e.g., difficulty following the recommendation), or other observations (e.g., side effects such as increased daytime alertness). The feedback is sent back to the Engine 108, typically through the personal devices 104.

[0047] Feedback Loops Back to Update Recommendations. The feedback is used by the Engine 108 to update its recommendations. For example, where an intervention was ineffective, the Engine 108 may select a different intervention from the intervention database 114 or adjust its approach based on the user’s input. This creates a continuous feedback loop, ensuring that the system adapts to the user’s evolving needs and preferences.

[0048] FIG. 4 illustrates a method for generating a hyper-personalized, evidence-based sleep product, service, or intervention recommendation, in accordance with one or moreAttorney Docket 022546.005WO1 implementations of the present subject matter. At 402, the Engine 108 collects longitudinal objective sleep architectural data (OSAD) from the one or more personal devices 104. At 404, the Engine 108 collects additional contextual and lifestyle data (ACDL), such as self-reported sleep quality, behavior logs, chronotype, self-reported sleep disorders, geolocation, and demographic characteristics.

[0049] At 406, the user phenotyping module 122 generates a user phenotype based on the collected OSAD and ACDL data. At 408, the Engine 108 queries the scientific research database 110 based at least in part on the user phenotype. At 410, the Engine 108 receives evidence-based insights, such as identification of user phenotypes and sleep-related issues, from the scientific research database 110.

[0050] At 412, the Engine 108 queries the intervention database 114 to identify candidate interventions that match the user phenotype and evidence-based insights. At 414, the Al and algorithms module 112 applies one or more artificial intelligence or algorithmic ranking techniques to the candidate interventions to produce a ranked set of candidate interventions. The ranking may take into account the user phenotype, the evidence-based insights, prior feedback from the user 102, and crowdsourced feedback from one or more other users. At 416, the Engine 108 outputs one or more personalized recommendations to the user 102, preferably via the client application 116 executing on at least one of the personal devices 104.

[0051] FIG. 5 illustrates a method for incorporating user feedback and dynamically refining the recommendation logic of the Engine 108. At 502, the Engine 108 delivers a recommendation to a user device. At 504, the user 102 implements the recommended intervention. At 506, the Engine 108 collects post-intervention objective sleep data from the one or moreAttorney Docket 022546.005WO1 personal devices 104. At 508, the Engine 108 collects user-reported feedback indicative of hedonic value and self-reported efficacy of the recommended intervention.

[0052] At 510, the model refinement module 128 updates one or more predictive models within the Al and algorithms module 112 based at least in part on the post-intervention objective sleep data and the user-reported feedback. At 512, the Engine 108 updates its recommendation logic based on the updated predictive models. At 514, the Engine 108 generates one or more refined recommendations. The refined recommendations may be delivered to the user 102, thereby completing a feedback loop that allows the system to adapt to the user’s evolving needs and preferences. Optionally, the post-intervention data and feedback collected at 506 and 508 are anonymized and aggregated as crowdsourced data that further enhances the Engine’s predictive capabilities across a plurality of users.

[0053] FIG. 6 illustrates an example computing system 600 that may be used to implement one or more aspects of the subject matter described herein. The computing system 600 may include a processor 610, a non-transitory machine-readable medium 612 (e g., memory), a storage 614, an input / output (I / O) interface 616, a display 618, an input device 620, and a network interface 622, all coupled together by a bus or interconnect 624. The processor 610 may include one or more central processing units (CPUs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), graphics processing units (GPUs), tensor processing units (TPUs), or other programmable processors. The non-transitory machine-readable medium 612 stores instructions executable by the processor 610 to perform operations described herein. The display 618 may include a cathode ray tube (CRT) display, a liquid crystal display (LCD), or a light-emitting diode (LED) display. The input device 620 may include a keyboard, a pointing device (e g., a mouse or trackball), a touchscreen, a single- or multi-point resistive or capacitiveAttorney Docket 022546.005WO1 trackpad, voice recognition hardware and software, an optical scanner, an optical pointer, a digital image capture device, or any combination thereof.

[0054] The network interface 622 enables the computing system 600 to communicate with a communications network 626, which may include, without limitation, the Internet, a wireless wide area network, a local area network, a wide area network, a wired network, or any combination thereof. The communications network 626 may further connect the computing system 600 to one or more servers 628 and one or more clients 630, which may be remote from each other and typically interact through the communications network 626. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a clientserver relationship to each other.

[0055] One or more aspects or features of the subject matter described herein can be realized in digital electronic circuitry, integrated circuitry, specially designed application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), computer hardware, firmware, software, and / or combinations thereof. These various aspects or features can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device. The programmable system or computing system may include clients and servers. A client and server are generally remote from each other and typically interact through a communications network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.Attorney Docket 022546.005WO1

[0056] These computer programs, which can also be referred to as programs, software, software applications, applications, components, or code, include machine instructions for a programmable processor, and can be implemented in a high-level procedural language, an object-oriented programming language, a functional programming language, a logical programming language, and / or in assembly or machine language. As used herein, the term “machine-readable medium” refers to any computer program product, apparatus, and / or device, such as for example magnetic discs, optical disks, memory, and Programmable Logic Devices (PLDs), used to provide machine instructions and / or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term “machine-readable signal” refers to any signal used to provide machine instructions and / or data to a programmable processor. The machine-readable medium can store such machine instructions non-transitorily, such as for example as would a non-transient solid-state memory or a magnetic hard drive or any equivalent storage medium. The machine-readable medium can alternatively or additionally store such machine instructions in a transient manner, such as for example as would a processor cache or other random access memory associated with one or more physical processor cores.

[0057] To provide for interaction with a user, one or more aspects or features of the subj ect matter described herein can be implemented on a computer having a display device, such as for example a cathode ray tube (CRT), a liquid crystal display (LCD), or a light-emitting diode (LED) monitor for displaying information to the user and a keyboard and a pointing device, such as for example a mouse or a trackball, by which the user may provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well. For example, feedback provided to the user can be any form of sensory feedback, such as for example visual feedback,Attorney Docket 022546.005WO1 auditory feedback, or tactile feedback; and input from the user may be received in any form, including but not limited to acoustic, speech, or tactile input. Other possible input devices include, but are not limited to, touch screens or other touch-sensitive devices such as single- or multi-point resistive or capacitive trackpads, voice recognition hardware and software, optical scanners, optical pointers, digital image capture devices and associated interpretation software, and the like.

[0058] In the descriptions above and in the claims, phrases such as “at least one of’ or “one or more of’ may occur followed by a conjunctive list of elements or features. The term “and / or” may also occur in a list of two or more elements or features. Unless otherwise implicitly or explicitly contradicted by the context in which it is used, such a phrase is intended to mean any of the listed elements or features individually or any of the recited elements or features in combination with any of the other recited elements or features. For example, the phrases “at least one of A and B,” “one or more of A and B,” and “A and / or B” are each intended to mean “A alone, B alone, or A and B together.” A similar interpretation is also intended for lists including three or more items. For example, the phrases “at least one of A, B, and C,” “one or more of A, B, and C,” and “A, B, and / or C” are each intended to mean “A alone, B alone, C alone, A and B together, A and C together, B and C together, or A and B and C together.” Use of the term “based on,” above and in the claims, is intended to mean “based at least in part on,” such that an unrecited feature or element is also permissible.

[0059] The subject matter described herein can be embodied in systems, apparatus, methods, and / or articles depending on the desired configuration. The implementations set forth in the foregoing description do not represent all implementations consistent with the subject matter described herein. Instead, they are merely some examples consistent with aspects related to the described subject matter. Although a few variations have been described in detail above, otherAttorney Docket 022546.005WO1 modifications or additions are possible. In particular, further features and / or variations can be provided in addition to those set forth herein. For example, the implementations described above can be directed to various combinations and subcombinations of the disclosed features and / or combinations and subcombinations of several further features disclosed above. In addition, the logic flows depicted in the accompanying figures and / or described herein do not necessarily require the particular order shown, or sequential order, to achieve desirable results. Other implementations may be within the scope of the following claims.

Claims

Attorney Docket 022546.005WO1CLAIMSWhat is claimed is:

1. A sleep product recommendation system, comprising:at least one personal device configured to capture longitudinal sleep data and contextual data associated with a user; anda recommendation engine in communication with the at least one personal device over a communications network, the recommendation engine comprising at least one processor and a non-transitory machine-readable medium storing instructions that, when executed by the at least one processor, cause the recommendation engine to perform operations comprising:receiving the longitudinal sleep data and the contextual data; generating a user phenotype based at least in part on the longitudinal sleep data and the contextual data; querying a scientific research database based at least in part on the user phenotype, and receiving from the scientific research database one or more evidence-based insights relevant to the user phenotype;querying an intervention database based at least in part on the user phenotype and the one or more evidence-based insights, and receiving from the intervention database a plurality of candidate interventions; applying an artificial intelligence model to rank the plurality of candidate interventions to produce a ranked set of candidate interventions; andtransmitting at least one personalized recommendation derived from the ranked set of candidate interventions to the user; and a feedback module configured to receive, from at least one of the at least one personal device, post-intervention objective sleep data and self-reported feedback associated with the at least one personalizedAttorney Docket 022546.005WO1 recommendation, and to update the artificial intelligence model based at least in part on the post-intervention objective sleep data and the self-reported feedback.

2. The system of claim 1, wherein the longitudinal sleep data comprises one or more of total sleep time, sleep onset latency, wake after sleep onset, and sleep stage distribution, the sleep stage distribution including distribution among at least light sleep, deep sleep, and rapid eye movement (REM) sleep.

3. The system of claim 1, wherein the contextual data comprises one or more of self-reported sleep quality, behavior logging or tagging data, chronotypic or circadian data, selfreported sleep disorders, geolocation data, and demographic data including age and sex.

4. The system of claim 1, wherein the at least one personal device comprises one or more of a smartphone, a wearable monitoring device, and a proximal or external sensor.

5. The system of claim 4, wherein the proximal or external sensor comprises at least one of a mattress sensor, a bedside monitor, an environmental sensor, and a polysomnography device.

6. The system of claim 1, wherein the recommendation engine further comprises a data ingestion module, a user phenotyping module, a recommendation generator module, a feedback module, a model refinement module, and an output module.Attorney Docket 022546.005WO1 7. The system of claim 1, wherein the artificial intelligence model comprises one or more of a machine-learning model, a deep-learning model, an ensemble model, and a statistical model.

8. The system of claim 1, wherein the scientific research database stores at least one of peer-reviewed studies, clinical research data, and other scientific literature relating to sleep health, sleep interventions, and outcomes.

9. The system of claim 1, wherein the intervention database stores a catalog of candidate interventions comprising one or more of a behavioral intervention, a sleep hygiene practice, a technological solution, a sleep product offering, and a clinical intervention.

10. The system of claim 9, wherein the sleep product offering comprises at least one of a mattress, a pillow, bedding, sleepwear, a sleep supplement, a sleep tracker, a smartlighting device, and a sound machine.

11. The system of claim 1, wherein the feedback module is further configured to anonymize and aggregate the post-intervention objective sleep data and the self-reported feedback as crowdsourced data, and to incorporate the crowdsourced data into one or more predictive models used by the artificial intelligence model.

12. The system of claim 1, wherein the operations further comprise presenting, via a client application executing on the at least one personal device, a personalized experimentAttorney Docket 022546.005WO1 in which the user is invited to evaluate one or more of the candidate interventions and to return objective post-intervention sleep data and subjective feedback.

13. The system of claim 1, wherein generating the user phenotype comprises identifying one or more sleep-architecture irregularities in the longitudinal sleep data that are not explicitly reported by the user, the one or more sleep-architecture irregularities including at least one of an REM sleep duration irregularity and an excessive wake-after-sleep-onset value.

14. The system of claim 1, wherein the recommendation engine is configured to dynamically update recommendation logic, in substantially continuous fashion, in response to the post-intervention objective sleep data and the self-reported feedback.

15. A computer-implemented method for generating a hyper-personalized, evidence-based sleep recommendation, the method comprising:collecting, by a recommendation engine and from one or more personal devices, longitudinal objective sleep architectural data and additional contextual data associated with a user;generating, by the recommendation engine, a user phenotype based at least in part on the longitudinal objective sleep architectural data and the additional contextual data;querying, by the recommendation engine, a scientific research database based at least in part on the user phenotype to identify one or more evidence-based insights relevant to the user phenotype;Attorney Docket 022546.005WO1 querying, by the recommendation engine, an intervention database based at least in part on the user phenotype and the one or more evidence-based insights to identify a plurality of candidate interventions;applying, by the recommendation engine, at least one artificial intelligence or machine-learning algorithm to rank the plurality of candidate interventions and to produce a ranked set of candidate interventions;transmitting, by the recommendation engine, at least one personalized recommendation derived from the ranked set of candidate interventions to a personal device associated with the user;collecting, by the recommendation engine, post-intervention objective sleep data and user-reported feedback indicative of hedonic value and self-reported efficacy of the at least one personalized recommendation; andupdating, by the recommendation engine, the at least one artificial intelligence or machine-learning algorithm based at least in part on the post-intervention objective sleep data and the user-reported feedback.

16. The method of claim 15, wherein collecting the longitudinal objective sleep architectural data comprises receiving, from at least one of the one or more personal devices, one or more of total sleep time, sleep onset latency, wake after sleep onset, and sleep stage distribution among light sleep, deep sleep, and rapid eye movement (REM) sleep.Attorney Docket 022546.005WO1 17. The method of claim 15, wherein generating the user phenotype comprises identifying one or more sleep-architecture irregularities that are not explicitly reported by the user.

18. The method of claim 15, further comprising anonymizing and aggregating the post-intervention objective sleep data and the user-reported feedback as crowdsourced data, and using the crowdsourced data to refine the at least one artificial intelligence or machine-learning algorithm across a plurality of users.

19. The method of claim 15, further comprising presenting, via a client application executing on the personal device, a personalized experiment in which the user is invited to evaluate one or more of the candidate interventions and to return objective postintervention sleep data and subjective feedback to the recommendation engine.

20. The method of claim 15, wherein applying the at least one artificial intelligence or machine-learning algorithm comprises applying one or more of a machinelearning model, a deep-learning model, an ensemble model, and a statistical model.

21. A non-transitory machine-readable medium storing instructions which, when executed by at least one programmable processor, cause the at least one programmable processor to perform operations comprising:receiving, from one or more personal devices, longitudinal objective sleep architectural data and additional contextual data associated with a user;Attorney Docket 022546.005WO1 generating a user phenotype based at least in part on the longitudinal objective sleep architectural data and the additional contextual data;querying a scientific research database based at least in part on the user phenotype to identify one or more evidence-based insights;querying an intervention database based at least in part on the user phenotype and the one or more evidence-based insights to identify a plurality of candidate interventions;applying an artificial intelligence model to rank the plurality of candidate interventions;transmitting at least one personalized recommendation derived from the ranked plurality of candidate interventions to a personal device associated with the user;receiving post-intervention objective sleep data and user-reported feedback indicative of hedonic value and self-reported efficacy of the at least one personalized recommendation; andupdating the artificial intelligence model based at least in part on the postintervention objective sleep data and the user-reported feedback.

22. The non-transitory machine-readable medium of claim 21, wherein the operations further comprise dynamically updating recommendation logic in response to the postintervention objective sleep data and the user-reported feedback.

23. The non-transitory machine-readable medium of claim 21, wherein the operations further comprise anonymizing and aggregating the post-intervention objective sleepAttorney Docket 022546.005WO1 data and the user-reported feedback as crowdsourced data and using the crowdsourced data to refine the artificial intelligence model across a plurality of users.