Abstract: An image processing system includes a memory to store data indicative of benchmark feature vectors of benchmark images, an input interface to accept data indicative of a first feature vector of a first image and a second feature vector of a second image, and an output interface to render a similarity value between the first and the second images. The system includes a processor to determine the similarity using a first hyperplane separating the benchmark feature vectors from the first feature vector and a second hyperplane separating the benchmark feature vectors from the second feature vector. The processor determines a first normal vector to the first hyperplane as the difference between the first feature vector and the mean of the benchmark feature vectors. The processor determines an offset for the first hyperplane as the average of the maximum inner product of the benchmark feature vectors with the first normal vector and the inner product of the first feature vector with the first normal vector.

Abstract: In certain embodiments, training of a prediction model (e.g., recognition or other prediction model) may be facilitated via a training set generated based on one or more logos or other graphics. In some embodiments, graphics information associated with a logo or graphic (e.g., to be recognized via a recognition model) may be obtained. Training media items (e.g., images, videos, etc.) may be generated based on the graphics information, where each of the training media items includes (i) content other than the logo and (ii) a given representation of the logo integrated with the other content. The training media items may be processed via the recognition model to generate predictions (related to recognition of the logo or graphic for the training media items). The recognition model may be updated based on (i) the generated predictions and (ii) corresponding reference indications (related to recognition of the logo for the training media items).

Abstract: A system and method multifunction peripheral policy-based configuration includes a processor and a memory storing a policy data for each of a plurality of multifunction peripheral device configuration policies. A selectable list of the device policies is displayed. A user selects one or more device configuration policies from the list, and the processor tests selected device configuration policies. Policies are identified as single or additive policy types. Selected policies are added when not identified as a single policy type.

Abstract: An apparatus that performs image processing on input image data in parallel in units of block according to an aspect of the invention includes a derivation unit configured to derive a memory size used in conversion processing of converting data alignment sequence in the image data from a line sequence into a block sequence which is executed before rendering of the image data and a memory size used in the conversion processing which is executed after the rendering of the image data, a determination unit configured to determine a timing for executing the conversion processing based on the derivation results, and a conversion unit configured to execute the conversion processing at the determined timing.

Abstract: Implantable transponders comprising no ferromagnetic parts for use in medical implants are disclosed herein. Such transponders may assist in preventing interference of transponders with medical imaging technologies. Such transponders may optionally be of a small size, and may assist in collecting and transmitting data and information regarding implanted medical devices. Methods of using such transponders, readers for detecting such transponders, and methods for using such readers are also described.

Abstract: Methods pertaining to encoding and decoding binary identifiers within a cell array are described. A binary identifier received by computing device can be encoded according to an encoding scheme. The cell array can include multiple encoded cells, each of which indicates a predetermined sequence of two or more bits, and which includes a perimeter, and both an alignment mark and a line pattern within the perimeter. The line pattern can be one of an empty-cell line pattern, a pattern including one or more asymmetrical radial vectors, one or more diametrical vectors, a symmetric cross, or a symmetrical star, or some other line pattern. The encoding scheme can define a plurality of cell colors that correspond to a predetermined sequence of two or more bits. The bits corresponding to a cell color can be redundant to bits corresponding to a line pattern for confirming accuracy of decoding a cell.

Abstract: A two-dimensional code that is capable of suppressing unnecessary code detection processes related to detection of a code area is provided. Among outer edges of a rectangular code area, a first edge that is composed of a plurality of types of cells arrayed in a single row is configured such that a specific pattern that is composed of an array of two or more differing cells is repeated in at least an intermediate portion. A second edge composed of the plurality of types of cells arrayed in a row such as to oppose the first edge is configured such that the specific pattern is repeated in at least an intermediate portion.

Abstract: A quick response code system and methods of associating a quick response code with an item of interest are the subject of the present disclosure. In one embodiment, the system includes a code generation tool and an information modification tool. The code generation tool is configured to generate a quick response code that associates the item of interest with a record in a data repository, while the information modification tool that allows a user to access, view and modify information associated with the record. According to varying embodiments, the system prompts a first user to share the quick response code and the associated record with a second user.

Abstract: Provided is an electronic card (1) including: a card casing (2) which is formed as a plate-like body, and in which a tapered surface (2At) having an inclination in which a thickness of a cross-section decreases toward an end surface of the plate-like body on one surface side of the plate-like body, and a step (2b) protruding from the tapered surface in a direction intersecting the tapered surface is formed; an electronic device incorporated in the card casing; and a sheet shaped part (3) in which a thickness of a cross section is formed to be thinner than a height of the step, and end surface abuts against a wall surface formed on the step, and which is fixed to the one surface side of the card casing including the tapered surface.

Abstract: The technology disclosed determines which field values in a set of unique field values for a particular field in a fielded dataset are anomalous using six similarity measures. A factor vector is generated per similarity measure and combined to form an input matrix. A convolutional neural network processes the input matrix to generate evaluation vectors. A fully-connected network evaluates the evaluation vectors to generate an anomaly scalar for a particular unique field value. Thresholding is applied to anomaly scalar to determine whether the particular unique field value is anomalous.

Abstract: Configurable three-dimensional neural network array. In an exemplary embodiment, a three-dimensional (3D) neural network array includes a plurality of stacked synapse layers having a first orientation, and a plurality of synapse lines having a second orientation and passing through the synapse layers. The neural network array also includes synapse elements connected between the synapse layers and synapse lines. Each synapse element includes a programmable resistive element. The neural network array also includes a plurality of output neurons, and a plurality of select transistors connected between the synapse lines and the output neurons. The gate terminals of the select transistors receive input signals.

Abstract: A neuromorphic architecture for a spiking neural network comprising a plurality of spiking neurons, each with a plurality of synapses and corresponding synaptic weights, the architecture further comprising a synaptic competition mechanism in connection with a spike-based learning mechanism based on spikes perceived behind a synapse, in which architecture synapses of different neurons connected to the same input compete for that input and based on the result of that competition, each neuron of the neural network develops an individual perception of the presented input spikes, the perception used by the learning mechanism to adjust the synaptic weights.

Abstract: The present invention describes a method for classifying an object having the following steps: receiving at least one item of distance information of an object based on a first electromagnetic signal transmitted by a transmitter device and a first electromagnetic signal received by a receiver device; receiving at least one item of oscillation information of the object based on a second electromagnetic signal transmitted by a transmitter device and a second electromagnetic signal received by a receiver device, which represents a solid oscillation of at least one subsection of the object; classifying the object based on the received information.

Abstract: A device, system, and method is provided for storing a sparse neural network. A plurality of weights of the sparse neural network may be obtained. Each weight may represent a unique connection between a pair of a plurality of artificial neurons in different layers of a plurality of neuron layers. A minority of pairs of neurons in adjacent neuron layers are connected in the sparse neural network. Each of the plurality of weights of the sparse neural network may be stored with an association to a unique index. The unique index may uniquely identify a pair of artificial neurons that have a connection represented by the weight. Only non-zero weights may be stored that represent connections between pairs of neurons (and zero weights may not be stored that represent no connections between pairs of neurons).

Abstract: Two and three-dimensional neural network arrays. In an exemplary embodiment, a two-dimensional (2D) neural network array includes a plurality of input neurons connected to a plurality of input lines, and a plurality of output neurons connected to a plurality of output lines. The 2D neural network array also includes synapse elements connected between the input lines and the output lines. Each synapse element includes a programmable resistive element. A three-dimensional (3D) neural network array includes a plurality of stacked two-dimensional (2D) neural network arrays each having a plurality of input neurons connected to a plurality of input layers and a plurality of output neurons connected to a plurality of output layers. The output layers intersect with the input layers and include synapse elements formed between intersecting regions of the input layers and the output layers. Each synapse element includes a programmable resistive element.

Abstract: A system and method determines engagement for a future presentation for one or more future attendees/registrants for the future presentation. The system and method may extract features from the registrant data for each registrant of the future presentation, generate a model based on the extracted features and use a classification process to generate the engagement determination for each attendee/registrant of the future presentation.

Abstract: The technology disclosed determines one or more field values in a set of field values for a particular field in a fielded dataset that are similar to an input value using six similarity measures. A factor vector is generated per similarity measure and combined to form an input matrix. A convolutional neural network processes the input matrix to generate evaluation vectors. A fully-connected network evaluates the evaluation vectors to generate suggestion scalars for similarity to a particular input value. Thresholding is applied to suggestions scalars to determine one or more suggestion candidates for the particular input value.

Abstract: Methods, systems and computer program products implementing data enrichment using global structure learning are disclosed. An information enrichment system predicts a likely canonical name from a transaction record in which names may be shortened, or extra token(s) inserted. In a training phase, the information enrichment system determines tag patterns based on labeled and unlabeled training transaction records. The tag patterns include co-occurrence probability and sequential order of co-occurrence of tags. In a testing phase, the information enrichment system receives a test transaction record. The information enrichment system predicts a likely tag sequence from the test transaction record based on the tag patterns. The information enrichment system predicts a canonical name based on likely tag values and token composition. The information enrichment system can then enrich the test transaction record with the predicted canonical name.

Abstract: The present approach relates to the training of a machine learning algorithm for image generation and use of such a trained algorithm for image generation. Training the machine learning algorithm may involve using multiple images produced from a single set of tomographic projection or image data (such as a simple reconstruction and a computationally intensive reconstruction), where one image is the target image that exhibits the desired characteristics for the final result. The trained machine learning algorithm may be used to generate a final image corresponding to a computationally intensive algorithm from an input image generated using a less computationally intensive algorithm.

Abstract: A machine learning system includes a memory and a processor. The processor is configured to access and execute at least one instruction from the memory to perform inputting raw data to a first partition of a neural network, in which the first partition at least comprises an activation function of the neural network. The activation function is applied to convert the raw data into irreversible metadata. The metadata is transmitted to a second partition of the neural network as inputs to generate a learning result corresponding to the raw data.

Abstract: A device may receive, from a user device, a request to verify a machine learning (ML) application using a metamorphic testing procedure. The device may determine a type of ML process used by the ML application, and may select one or more metamorphic relations (MRs), to be used for performing the metamorphic testing procedure, based on the type of ML process. The device may receive test data to be used to test the ML application, wherein the test data is based on the one or more MRs, and may perform, by using the one or more MRs and the test data, the metamorphic testing procedure to verify one or more aspects of the ML application.

Abstract: A method including receiving an input data set. The input data set can include one of a feature domain set or a kernel matrix. The method also can include constructing dense embeddings using: (i) Nyström approximations on the input data set when the input data set comprises the kernel matrix, and (ii) clustered Nyström approximations on the input data set when the input data set comprises the feature domain set. The method additionally can include performing representation learning on each of the dense embeddings using a multi-layer fully-connected network for each of the dense embeddings to generate latent representations corresponding to each of the dense embeddings. The method further can include applying a fusion layer to the latent representations corresponding to the dense embeddings to generate a combined representation. The method additionally can include performing classification on the combined representation. Other embodiments of related systems and methods are also disclosed.

Abstract: Systems and methods for predicting performance of a modulation system are provided. A neural network model is trained using performance information of a source system. The neural network model is modified with transferable knowledge about a target system to be evaluated. The neural network model is tuned using specific characteristics of the target system to create a source-based target model. The target system performance is evaluated using the source-based target model to predict system performance of the target system.

Abstract: An image processing engine component generation method, and search method, terminal, and system are provided. The image processing engine component includes: an online feature processing component operative to receive an incremental image, and determine image content feature information of the incremental image; an offline feature processing component operative to receive a stock image, and determine image content feature information of the stock image; a search platform component operative to receive a search request, the search request including search text information; and a processing engine component operative to receive image content feature information of the incremental image and the stock image, add the received image content feature information to index tables, and determine search results corresponding to the search text information.

Abstract: A mechanism is described for facilitating multifunction perceptron-based machine learning in computing environments, according to one embodiment. A method of embodiments, as described herein, includes generating a multifunction perceptron architecture having a plurality of neurons to perform one or more neuron functions in a machine learning environment, wherein the plurality of neurons includes one or more of splitter neurons, mixer neurons, and counter neurons, wherein the plurality of neurons include heterogenous neurons.

Abstract: Systems and methods are provided for teaching and shaping the behavior of artificial intelligence agents using human input. A system may be provided that comprises a natural language processing neural network (NLP NN) and a relational reasoning neural network (RR NN). The NLP NN may receive a human observation and environment observations, and output encodings of labels for entities, actions, and/or policies, wherein the environment observations are indicative of states of the environment, and the human observation represents an observation of the environment made by a human. The RR NN may generate cross-modal embeddings from the environment observations and the encodings of labels generated by the NLP NN. The agent neural network may generate actions and/or policies from the environment observation and the cross-modal embeddings.

Abstract: Aspects of the present disclosure involve systems, methods, devices, and the like for generating compact tree representations applicable to machine learning. In one embodiment, a system is introduced that can retrieve a decision tree structure to generate a compact tree representation model. The compact tree representation model may come in the form of a matrix design to maintain the relationships expressed by the decision tree structure.

Abstract: A computer-implemented method for building a semantic analysis model. In one embodiment, the computer-implemented method includes creating proxy tags comprising a set of surface form variants. The computer-implemented method creates training examples comprising a combination of terminal tokens and at least one of the proxy tags. The computer-implemented method builds the semantic analysis model using the training examples.

Abstract: Embodiments include methods, systems, and computer program products for determining health care vital data. Aspects include receiving a health care vital measurement for a patient and health care vital data for a population. Aspects also include determining a baseline for the patient based at least in part upon the health care vital data for the population. Aspects include determining whether the patient health care vital data deviates from the baseline by more than a threshold and, applying a cognitive learning model to the patient health care vital measurement to correct for an anxiety-based impact to generate a corrected health care vital measurement responsive to a determination that the patient health care vital data deviates from the baseline by more than the threshold.

Abstract: A method for knowledge management using concept rules includes receiving event data corresponding to an industrial application and generating at least one inference concept based on the event data. The method also includes obtaining a semantic model having a plurality of inference concepts, a plurality of relationships among the plurality of inference concepts, and a plurality of concept rules representative of domain knowledge. The plurality of concept rules is authored using the plurality of inference concepts and the plurality of relationships. Furthermore, the method includes processing the at least one inference concept based on the semantic model to generate inferential data. The inferential data is representative of an inference corresponding to the event data. In addition, the method includes controlling the industrial application based on the inferential data.

Abstract: Systems, methods, and computer readable mediums are provided for predicting a sequence of actions based on psychographic measures determined from biometric information. Data received from a computing device can include biometric information characterizing a pattern of user interaction during completion of an objective requiring a first set of actions be performed in a first sequence. A psychographic measure characterizing a user's state while performing the objective can be determined. Using the determined psychographic measure and a predictive model, a second sequence associated with the objective can be determined and can include a second set of actions different than the first set of actions. The predictive model can be trained to output the second sequence to affect the psychographic measure of the user completing the objective according to the second set of actions. The second sequence can be transmitted to the computing device for execution.

Abstract: This disclosure provides an apparatus and method for harnessing machine learning and data analytics for a real-time predictive model for a FCC pre-treatment unit. The method includes collecting operating parameters of a pre-treatment unit and fluid catalytic cracking (FCC) unit; evaluating an independent variable of the operating parameters; and adjusting an input to the pre-treatment unit to control the independent variable within specifications in an output of the FCC unit.

Abstract: Methods and apparatus for bit prediction and data compression are provided. In the bit prediction method and apparatus, a specific Suffix Trie is used to maintain a statistical model to predict the next bit in a bit-stream. The statistic model provides the probability for each next or following bit in the bit stream, where an entropy encoder/decoder further encodes/decodes the predicted bit. The algorithm simplicity and its high performance relies on the combination of the specific method of Suffix Trie construction and growing, and the specific way to compute, update and propagate the probability across Suffix Trie nodes. The specific and particular method to grow the Suffix Trie in conjunction with the specific method to compute, update, and propagate the probability across Suffix Trie nodes are key aspects and subject matter of the present invention.

Abstract: An automated method of predictive model development first cleans up raw supervised and unsupervised training data with a step that uses an algorithm to make every field of every record consistent, cohesive, and productive. Then the resulting flat data is given texture in a next step by a data enrichment algorithm that culls fields that do not contribute to predictive model building and that adds new fields computed from data combinations that are tested to add value to later steps that build different types of predictive models. Another late step for building smart-agents and their entity profiles uses another algorithm that benefits greatly from the cleaned and highly enriched training data. The predictive models and smart-agents and their entity profiles are then rendered as deliverable predictive model markup language documents in a final step executed by a specialized algorithm.

Abstract: An online system determines a stability metric that indicates overlap between the set of entities associated with a particular entity when embeddings have been adjusted due to modifications in the input data of an embedding model. The online system generates a stability score for the embedding model by computing a statistic for one or more stability metrics. The online system determines a stability metric for a particular content provider by identifying a first cluster of content providers in a set of first embeddings, and a second cluster of content providers in a set of second embeddings. The second embeddings are generated after modifications have been made to input data. The online system determines the stability metric based on an overlap between the first cluster and the second cluster of content providers. The stability score can be an indicator of model performance that can be used to select embedding models.

Abstract: A machine learning model for predicting a size fit satisfaction for a variable size component is trained using at least sizing profiles of a plurality of items and feedbacks of subjects regarding sizing of the plurality of items. The machine learning model is used to determine a value for the variable size component that corresponds to an optimal predicted size fit satisfaction. The determined value of the variable size component is provided for use in creating a new item with a sizing variation based on the determined value.

Abstract: A processor based system and method of generating new pattern knowledge of a sensory input is disclosed. The system comprises a memory to store abstract patterns as pattern knowledge, a sensor configured to receive a concrete pattern representing a sensory input, an adaption module configured to horizontally blended abstract patterns to create a horizontally blended abstract pattern as the new pattern knowledge if a degree of blending the abstract and concrete patterns exceeds a threshold. In some embodiments, the horizontally blended abstract pattern comprises a selective projection of parameters of two abstract patterns. In some embodiments, the systems and methods further comprise vertically blending the abstract pattern and the concrete pattern by selectively projecting parameters of an abstract pattern with the parameters of the concrete pattern.

Abstract: Systems and methods are provided for performing data mining and statistical learning techniques on a big data set. More specifically, systems and methods are provided for linear regression using safe screening techniques. Techniques may include receiving a plurality of time series included in a prediction hierarchy for performing statistical learning to develop an improved prediction hierarchy. It may include pre-processing data associated with each of the plurality of time series, wherein the pre-processing includes tasks performed in parallel using a grid-enabled computing environment. For each time series, the system may determine a classification for the individual time series, a pattern group for the individual time series, and a level of the prediction hierarchy at which the each individual time series comprises an need output amount greater than a threshold amount.

Abstract: A method and system may provide automated passenger seat assignment procedures for an airline. A server device may determine passengers for a departing flight who are eligible to receive seat assignment notifications. When an eligible passenger is next in a standby or upgrade queue, the server device may transmit an electronic communication such as an email, short message service (SMS) message, push notification, etc., to the passenger's network-enabled device. The electronic communication may indicate that a seat assignment is available and may provide instructions for accepting or declining the available seat assignment. If the user accepts, for example, by transmitting an electronic reply message to the server device accepting the available seat assignment, the server device may transmit an electronic boarding pass to the user corresponding to the available seat assignment.

Abstract: The utility of a portable consumer device is extended by allowing account holders the ability to gain entry into access-controlled venues using a portable consumer device that is associated with an account that was used to purchase the admission or tickets to the event at the access-controlled venue. Merchandise or other goods and services may be pre-purchased by the consumer. Pre-purchasing may occur at the time that the price of admission is paid or before entry into the venue. The portable consumer device may be validated at the entrance to the access-controlled venue. After the consumer's entry into the venue, an arrival notification alert may be triggered to provide additional goods or services. For example, when a consumer successfully enters the venue using the consumer portable device, the pre-purchased items may be delivered to the consumer location at a specific time.

Abstract: Embodiments provide a method for constructing a future-state power grid model and device, including that: a current power grid model, an equipment power-off plan, an equipment retirement plan and an equipment addition plan are acquired; then, equipment is added according to the current power grid model, and an added equipment information set, a retired equipment information set and a powered-off equipment information set are determined according to the equipment addition plan, the equipment retirement plan and the equipment power-off plan respectively; and finally, a state of the added equipment is set to be an operating state, an initial network model of each period is formed according to a time sequence, and a future-state network model is constructed according to the added equipment information set, the retired equipment information set, the powered-off equipment information set and the initial network models. The embodiments further provide construction equipment and a storage medium.

Abstract: Some embodiments of the invention provide systems and methods for deploying employees to roles in a restaurant. For example, some embodiments are directed to basing employee deployment decisions at least in part on a combination of quantitative data and qualitative assessments of performance by various employees in various roles. Qualitative assessments may, for example, relate to an employee's success in promoting customer satisfaction, or to the accuracy of quantitative data relating to a task or function.

Abstract: Systems and methods are provided for predicting a probability of a problem in a service at a service provider based on implementation of a change to the service. One exemplary method includes a risk engine accessing change records for historical changes in services associated with the service provider where each record includes a text description of the implemented change and a problem/no problem result for the change. For each record, the risk engine normalizes the text description of the implemented change and generates a word-count matrix based on the normalized text description. The risk engine then performs a regression analysis of the generated word-count matrices for the records and the corresponding problem/no problem results, thereby providing a regression model, and generates a predictive algorithm based on a score provided from the regression model and at least one change factor associated with the change records.

Abstract: A method for providing operational-efficiency data onboard an aircraft associated with a current flight plan. Prior to flight of the aircraft using the current flight plan, the method obtains and presents pre-flight briefing data, flight advisories, and stored alerts associated with the current flight plan, wherein the pre-flight briefing data comprises at least an operational-efficiency KPI associated with the current flight plan. The method also determines applicability of the operational-efficiency KPI to the current flight plan, based on the pre-flight briefing data, flight plan data, weather data, NOTAM data, and ATC communication data. When the operational-efficiency KPI is applicable to the current flight plan, the method computes and presents fuel savings associated with adherence to the operational-efficiency KPI during the flight; and configures and presents alerts associated with adherence to the operational-efficiency KPI during the flight.

Abstract: Disclosed embodiments relate to a system for monitoring the usage of a functional facility, the system comprising (i) a plurality of detectors for detecting the presence of at least one client device in the functional facility and which detectors are geographically mapped in the functional facility; and (ii) a client device mapping module for determining the location of the at least one client device detected by one or more of the plurality of detectors. Moreover, the invention is directed to a method for monitoring the usage of a functional facility including one or more detectors, comprising the steps (1) detecting an operable engagement made between a detector deployed in the functional facility and a computerized client device; (2) determining a location point of the computerized client device; and (3) providing an output which is descriptive of the detected operable engagement along with information relating to the determined location point.

Abstract: The present disclosure generally relates to a new mass transit system in which cooperative transit vehicles follow dynamic routes and flexible schedules and perform coordinated transfers to provide personalized transit trips which are faster, more convenient, and more efficient than traditional public transportation services with fixed routes and rigid schedules. Passengers use rider devices such as a smart phone to request transit trips and receive personalized transit plans including pickup, transfer, and drop off instructions in real time. A transit coordination center continuously tracks transit vehicles and passengers, process trip requests, offers personalized transit plans, calculates dynamic transit routes and transfer points, and sends routing and ride instructions to coordinate transit vehicles and passengers.

Abstract: Disclosed are various embodiments for facilitating the management of a large network of workers and employers. A schedule management system can includes a registration and verification system that allows a worker to provide his or her resume, certification(s), skills, specialties, experience, contact information, references, and/or any other information that may be needed for creating a worker profile. The schedule management system of the present disclosure can match registered and verified workers with the open shifts of employers.

Abstract: Methods, systems, and apparatuses are provided for enabling an automated handling of information technology incidents in a computing environment. An incident report relating to an incident in a computing environment is received. Based on the incident report, a feature vector is generated and provided as an input to a machine-learning model that may output one or more suggested actions to respond to the incident. For instance, the machine-learning model may be trained based on previous actions performed by a user in response to previous incident reports. A user interface is provided allowing a user to select one or more of the suggested actions. In response to the user's selection, the selected actions may be executed automatically. By orchestrating a set of actions to execute automatically, incident reports may be addressed in a timely and efficient manner.

Abstract: Systems, methods, and other embodiments associated with anomaly detection are described. In one embodiment, a method monitoring an on-going project that comprises a plurality of processes and process activities that occur during the process. A machine learning model is applied that identifies a group of projects that are a similar type as the on-going and generates an expected level of process activities that are expected to occur. Based on a snap shot of the on-going project at a first time period, observed levels of process activities are determined that occurred in each process. The machine learning model compares for each of the processes, the observed levels of process activities to the expected levels of process activities in a corresponding time period. If the observed levels of process activities fail to fall within a range of the expected levels of process activities, an anomaly alert is generated and displayed.

Abstract: A network system provides delivery of items using unmanned aerial vehicles (UAV) or drones. The network system uses an infrastructure of nodes that include landing pads to dock drones, as well as interfaces to provide and receive items from docked drones. Nodes may be stationary (e.g., fixed at a building rooftop or public transit station) or mobile (e.g., mounted to a vehicle). The network system may determine a route for delivery of an item, where a drone transports the item for at least a portion of the route. For example, the route may include multiple waypoints associated with nodes between which drones travel. For other portions of the route, the network system may request a provider to transport the item using a ground-based vehicle.