A search system that receives and returns results for split stays is described. The search system receives, from a searching end-user, a listing request specifying a multiple-day length of stay parameter. The search system determines that the multiple-day length of stay parameter of the listing request transgresses a minimum length of stay threshold and, in response, generates a combined listing that includes a first listing of the plurality of listings associated with a first portion of the multiple-day length of stay parameter and a second listing of the plurality of listings associated with a second portion of the multiple-day length of stay parameter. The combined listing is presented with one or more other listings of the plurality of listings that match the listing request in a ranked order.
There is provided a method that includes receiving, from a client device, a search request for a set of listings, the search request including search parameters defining a search query. The method further includes generating a set of listings based on the search query and the search parameters and extracting price-indicative and non-price-indicative features. The method also includes computing a probability of booking and an estimate of quality, by inputting the price-indicative features and non-price-indicative features to trained machine learning models. The trained machine learning models predict (i) an affordability metric based on the price-indicative features and (ii) a quality metric based on the non-price- indicative features, separately. The affordability metric and the quality metric are representative of the probability of booking, and the quality metric is representative of the estimate of quality. The method further includes ranking the set of listings based on the booking probability and the quality estimate.
A server obtains content in a first language and receives a first request from a first client device to view the content, wherein the first client device is associated with a second language selected by a user operating the first client device. In response to receiving the first request, the server determines that the first language is different from the second language and determines that a storage of the server does not include a machine-translated version of the content in the second language. In accordance with these determinations, the server obtains a machine-translated version of the content in the second language and stores the machine-translated version of the content in the second language in the storage for subsequent requests to view the content in the second language.
Systems and methods are provided for receiving image data via a camera of a computing device, the image data comprising a plurality of image frames; displaying a 3D reconstruction of the image data on a graphical user interface (GUI) displayed on a computing device as the image data is received and the 3D reconstruction of the image data is generated; detecting at least one object corresponding to one or more of a plurality of predefined object types in the image data; determining dimensions of the at least one object in 3D space based on the 3D reconstruction of the image data; and displaying in the GUI the at least one detected object.
Systems and methods are provided for extracting a plurality of features for a listing from a datastore comprising a plurality of listings and a plurality of features for each of the plurality of listings, determining a cluster of similar listings to the listing and generating a set of cluster features for the cluster of similar listings, analyzing the set of cluster features for the cluster of similar listings based on a booking price, using a first trained machine learning model to determine a cluster-level probability of booking the listing on the given date, analyzing the plurality of features for the listing using the booking price, using a second trained machine learning model to determine a listing-level probability of booking the listing on the given date, and generating a final probability of booking by combining the cluster-level probability of booking and the listing-level probability of booking.
Systems and methods are disclosed for retrieving, from a database, over a network, historical routing data for multiple attributes and determining, for each attribute, based on its respective historical routing data, whether processing volume and processing error rates for each attribute exceed respective threshold. If both processing volume and error rate exceed their respective thresholds, the systems and methods describe herein calculate, for each qualifying attribute, a degree to which routing for each attribute can be improved. The systems and methods described herein output a ranking for each qualifying attribute based on their respective degrees to which routing can be improved for the respective attributes.
Systems and methods are provided for receiving a request for services in a given location from a client device operated by a user and generating a set of features based on information included in the request for services in the given location. The systems and methods further provide for analyzing the set of features using a machine learning model to predict whether only services that can be instantly booked should be provided in response to the request for services in the given location, analyzing a prediction output by the machine learning model to determine that only services that can be instantly booked should be provided in response to the request for services in the given location, and generating a list with only services that can be instantly booked.
Systems and methods are provided for generating a first trained machine learning model, the first machine learning model comprising a plurality of hard layers for learning correlations between listing features and a plurality of soft layers, each soft layer for learning correlations for a prespecified listing feature. The systems and methods further provide for analyzing, using the first trained machine learning model, each of a plurality of price changes and price independent listing features for the first listing to determine a predicted value for each of a prespecified price dependent listing feature for each of the plurality of price changes for the first listing and generating, using the first trained machine learning model, the predicted value for each of the prespecified price dependent listing features for each of the plurality of price changes for the first listing.
A computer implemented method for incorporating multiple objectives in a ranked list of search results includes receiving a search query from a client device, accessing a set of stored listings for goods or services and probabilities of serving the listings, defining a serving vector as a probability distribution over the set of listings, providing a serving vector as input to a multi -objective function, decomposing the multi-objective function into one or more objective functions, generating a ranked list of the listings based at least in part on the serving vector that maximizes the decomposed multi-objective function, and providing the listings to the client device according to the order of the ranked list. Each objective function addresses a different goal in an overall diversity optimization.
A lifetime value estimation of a web-based advertisement can be generated at the time of creation of a reservable listing in response to the advertisement, even when little to no information about conversion of the reservable listing is available. In a case that there is no information about a conversion time for a listing, the estimation of the lifetime value is done by calculating a real conversion rate from an advertising platform-dependent historical conversion rate, an average percentage of converted listings that converted from unreserved to reserved within a set period of days, and a scaling multiplier. Where there are relatively few conversions, the estimation of a lifetime value for the advertisement can be done using a weighted average of a campaign-level global conversion rate, which encompasses a large number of keywords that have seen conversions, and a keyword-level local conversion rate. The resulting estimated valuation, by either method, is used to submit a keyword bid to an online advertising service for display alongside a user's search results.
An online reservation system (15) is configured to receive requests from a guest for searching property listings (77) and to return property listings that satisfy the search criteria of the requests. The online reservation system uses a machine learning algorithm to rank the property listings returned by the search. The reservation system uses objective functions to determine parameters for each property listing and assign a ranking based on the parameters. A first objective function generates a parameter indicating an extent to which a property listing matches preferences of the guest, and is based on data about the guest's interactions with the reservation system. A second objective function generates another parameter indicating an extent to which the search request matches the preferences of the host associate with the property listing, and is based on data about the host's responses to reservation requests.
Systems and methods for providing contextual calendar reminders are provided. A host can receive a calendar reminder notifying the host that a calendar for a property listing maintained by the host may need to be updated. The system can provide the host with a reminder to update the calendar for the property listing based on criteria related to the expected number of times the host will view the calendar within a designated time period and the timing of the host's last interaction with the calendar. If the designated criteria are satisfied, the system can generate a reminder for the host to update the calendar for the property listing. Before sending the reminder to the host, the system confirms that the host did not recently interact with the calendar and if the host did interact with the calendar, the system cancels the sending of the reminder.
Systems and methods are provided for analyzing a cancellation policy for a trip item to determine payment parameters set for the trip item, and for each of a plurality of predetermined number of installment payments, calculating a payment period for each payment parameter set for the trip item based on a time period before payment is due to meet each payment parameter and the predetermined number of installments. The systems and methods further provide for selecting a number of predetermined installment payments of the plurality of predetermined number of installment payments that comprises an optimal payment period between a final installment payment and the start date of the trip item, generating an installment plan based on the selected number of predetermined installment payments, the installment plan comprising a date and amount for each installment payment, and causing the installment plan to be presented via the computing device.
Systems and methods are provided for receiving from a first computing device associated with a first user, a request to register a group trip comprising at least one trip item, the request including parameters for the group trip, and receiving authorization from a second computing device associated with a second user to be included in the group trip. The systems and method further providing for receiving from the first computing device, a request to book a trip item for the group trip, approving the request to book the trip item for the group trip based on determining that the trip item meets the parameters for the group trip, and automatically charging a payment device associated with the first user and a payment device associated with the second user according to the parameters related to the group trip.
Systems and methods are provided for generating a search query based on received user data to perform an internet search using the search query. The systems and methods further extract data from internet search results from the internet search using the search query, generate internet search income records from the extracted data, generate income records from one or more databases comprising income records, and combine the internet search income records and the generated income records from the one or more databases to form combined income record results. The systems and methods further identify candidate income records from the combined income record results, extract features from each candidate income record for generating an income prediction, and generate the income prediction using a machine learning model to predict an income, based on the extracted features from the candidate income records.
G06F 17/30 - Information retrieval; Database structures therefor
G06Q 90/00 - Systems or methods specially adapted for administrative, commercial, financial, managerial or supervisory purposes, not involving significant data processing
G06N 3/10 - Interfaces, programming languages or software development kits, e.g. for simulating neural networks
17.
MACHINE LEARNING MODELING FOR GENERATING CLIENT RESERVATION VALUE
Systems and methods are provided for analyzing booking session data to generate a plurality of feature vectors for each booking session of the plurality of booking sessions, and generating training data comprising the plurality of feature vectors for each booking session and at least a first constraint. The systems and methods further providing for calculating a set of weights using the training data, wherein each weight is a lowest weight satisfying the most constraints possible, wherein the set of weights comprises a weight associated with each feature in the plurality of feature vectors, and computing a reservation value for each of a plurality of clients for each of a plurality of listings and for each date of a plurality of dates, based on the set of weights and the plurality of feature vectors.
G06Q 10/02 - Reservations, e.g. for tickets, services or events
G06F 15/18 - in which a program is changed according to experience gained by the computer itself during a complete run; Learning machines (adaptive control systems G05B 13/00;artificial intelligence G06N)
This disclosure includes methods for predicting demand based on the price of a time-expiring inventory. An online system provides a connection between a manager of a time-expiring inventory and a plurality of clients. The online system provides a listing for the manager's time-expiring inventory to clients on the online system. The manager specifies the price of the time-expiring inventory in the listing and is presented with price tips generated by the online system. A demand function predicts the demand for the time-expiring inventory based on features of the listing and the time-expiring inventory. A manager option function predicts the likelihood of acceptance of a price tip by the manager. The online system uses the demand function and the manager option function to create a Monte Carlo pricing model to provide to the manager price tips for the listing.
This disclosure includes systems for regression-tree-modified feature vector machine learning models for utilization prediction in time-expiring inventory. An online computing system receives a feature vector for a listing and inputs the feature vector and modified feature vectors into a demand function to generate demand estimates. The system inputs the demand estimates into a likelihood model to generate a set of request likelihoods, each request likelihood representing a likelihood that the time-expiring inventory will receive a transaction request at each of a set of test price and test times to expiration. The system further trains a regression tree model based on a set of training data comprising each of the request likelihoods from the set and the test price and test time period to expiration used to generate the demand estimate that was used to generate the request likelihood.
A communications system provides access to services when direct Internet connectivity is not practical. The system includes a beam modem and a beam API server. The beam modem receives a web request from a client device through a short range interface, modifies the request, and transmits the modified web request to the beam API server via a cellular connection. The beam API server then extracts an endpoint address and request data from the web request and determines an external web service from the endpoint address. The server transmits the request data to the external web service and, after receiving a response to the request data, reduces the size of the response data and sends it back to the beam modem via the cellular connection. The beam modem converts the response data to client device readable form and transmits it to the client device via the short range interface.
H04L 29/08 - Transmission control procedure, e.g. data link level control procedure
H04L 29/06 - Communication control; Communication processing characterised by a protocol
H04W 4/00 - Services specially adapted for wireless communication networks; Facilities therefor
H04W 4/18 - Information format or content conversion, e.g. adaptation by the network of the transmitted or received information for the purpose of wireless delivery to users or terminals
H04W 88/04 - Terminal devices adapted for relaying to or from another terminal or user
H04W 88/06 - Terminal devices adapted for operation in multiple networks, e.g. multi-mode terminals
H04W 88/10 - Access point devices adapted for operation in multiple networks, e.g. multi-mode access points
H04W 12/00 - Security arrangements; Authentication; Protecting privacy or anonymity
21.
AUTOMATED DATABASE RECORD ACTIVATION USING PREDICTIVE MODELING OF DATABASE ACCESS
A computer implemented system and method for selecting and notifying operators of the option to enable a record activation feature for a short interval of time for the records they offer in a selected geographic area. Enabling record activation for a record indicates that the record may be booked by a user without first requesting the operator to manually approve the transaction request and waiting for the operator's approval of the request. Before selecting and notifying operators, a demand for database requests is predicted. Operators that are most likely to offer their record for record activation are identified. A quality score is determined for each identified record based on the likelihood that the record will get booked once the operator has programmatically enabled record activation. The records needed to fulfil the demand for database requests are selected based on their quality score and the operators of the selected records are notified of the option to enable record activation.
An online computer system including a database uses an encrypted table that allows for write protection its contents. Middleware logic operating on the system acts as an interface for access to the database, so that any business logic on the system accesses the database through simple procedural calls to the middleware rather than directly to the database itself. The middleware logic abstracts logic that helps implement write protection with the encrypted table. Data to be encrypted that has been traditionally written to other tables is migrated to the encrypted table, where the data encrypted using an authenticated encryption with additional data (AEAD) algorithm. To implement AEAD, the original table, column, and primary key indicating where the data would have otherwise been stored are together used as additional authenticated data (AAD). This tuple of information is also stored in the encrypted table.
Methods and systems for enforcing host standards in a reservation system. In one embodiment, a reservation system provides two methods of communication between a potential guest and a host, hard communications and soft communications. The reservation system associates both types of communication with threads, each thread associated with the host and an associated reservation. The reservation system determines suspension metrics for each host based on the communications of the threads they are associated with. When the suspension metrics associated with a host satisfy one or more suspension criteria, the reservation system issues a suspension to the host.
This disclosure includes methods for predicting demand based on the price of a time-expiring inventory. An online system provides a connection between a manager of a time-expiring inventory and a plurality of clients. The online system provides a listing for the manager's time-expiring inventory to clients on the online system. The manager specifies the price of the time-expiring inventory in the listing. A demand function predicts the demand for the time-expiring inventory based on features of the listing and the time-expiring inventory. The online system determines a likelihood of receiving a request for (he time-expiring inventory from a client on the online system based on the predicted demand. The online system may use the determined likelihoods to provide to the manager information about how changes in the price of the listing are likely to affect the demand for the time-expiring inventory.
G06Q 10/02 - Reservations, e.g. for tickets, services or events
G06Q 10/04 - Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
G06Q 10/06 - Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
25.
PASSWORD MANIPULATION FOR SECURE ACCOUNT CREATION AND VERIFICATION THROUGH THIRD-PARTY SERVERS
A method and system for deterring attacks at potential breach points between servers and an account and login server for creating and subsequent verification of accounts. Various cryptographic primitives are used to manipulate passwords to generate verifiers. The verifiers are used with external hardware security modules (HSMs) to eliminate HSMs and intermediate steps between the HSM and login servers as potential breach points.
H04L 9/32 - Arrangements for secret or secure communications; Network security protocols including means for verifying the identity or authority of a user of the system
26.
DETERMINING HOST PREFERENCES FOR ACCOMMODATION LISTINGS
Methods and systems for determining the preferences of hosts offering accommodations are disclosed. In one embodiment, an online booking system models the preferences of hosts based on statistical relationships between features of previously received accommodation reservation requests and the acceptance of those reservation requests by the hosts. In particular, the system classifies reservation requests based on several features - a reservation request either possesses a feature or does not possess a feature. The preference of a host for a particular request feature is modeled based on the relationship between the reservation requests that possess the feature and the reservation requests that are accepted by the host.
Listings and reviews of listings can be processed to identify descriptive attributes for locations associated with the listings. To do this, a corpus of words is generated for various locations based on listings in the locations and reviews of those listings. An expected frequency, and per-location frequency for each word is determined. These numbers are in turn used to determine a number of high frequency listing locations, and a number of below expected frequency listing locations for each word. Based on a comparison of the number of high frequency listing locations and the number of below expected frequency listing locations of a word with an attribute reference number, the word can be identified either as an attribute that is likely descriptive of the location, or not.
An online booking system allows users to search and book listings of goods or services. When a user searches for listings, the listings are ranked and scored based on a number of factors including the location and the price of the listing, the number and quality of reviews and the number of successful prior bookings. In some situations, the listing scores overly skew the top ranking results to a particular region. The listing scores may be modified to address this skewing of results. When diversity in search results is desired, the listing scores are modified such that the top ranking results that are located in a diverse set of regions. When granular relevance of search results is desired, the listing scores are modified such that the top ranking results are located in regions that are more relevant to the search than the region to which the results are skewed.
An online booking system allows users to creates, search, and book listings of goods or services. When a user searches for listings, the listings are ranked at least in part based on a location relevance score including at least one of a city relevance subscore, a neighborhood subscore, and a distance subscore. Generally, the city relevance subscore quantifies the probability that a searching user may have actually intended to look for listings in a city other than the city specified in a search query. Generally, the neighborhood relevance subscore quantifies the popularity of specific neighborhoods within a city as a replacement or addition to the distance subscore that determines a real world distance between a listing's real world location and a location specified in a search query.
Methods and systems for verifying the identity and trustworthiness of a user of an online system are disclosed. In one embodiment, the method comprises receiving online and offline identity information for a user and comparing them to a user profile information provided by the user. Furthermore, the user's online activity in a third party online system and the user's offline activity are received. Based on the online activity and the offline activity a trustworthiness score may be calculated.
Methods and systems for updating a calendar entry for an accommodation listing are disclosed. In one embodiment, the method comprises generating an availability model and an acceptance model for an accommodation listing in an accommodation reservation system and determining based on those models the probability that the accommodation listing would be able to be booked. Furthermore, the result of an accommodation search query can be filtered and/or sorted using the determined probability of booking.
Systems and methods for empirical expert determination and question and answer routing are provided. They system collects location tracking data about each user and analyzes the location tracking data to empirically determine the level of expertise a particular user has for a specific venue/event or a specific geographic region at a particular scale on a map. The system receives questions about a specific venue/event or about a category of venue/event in a specific geographic region at a particular scale on a map and routes those questions in real time to one or more experts for the specific venue/event or the category of venue/event in the specific geographic region. The system receives a response to the question from at least one of the one or more experts and routes the response back to the requestor, also in real time. The system also efficiently represents the location of a plurality of venues/events and/or users within a specific geographic region on a displayed map at a plurality of scales of the map.