Abstract: The disclosure relates to a self-adaptive leakage value neuron information processing method and system. The method includes: receiving front end pulse neuron output information; reading current pulse neuron information, wherein the current pulse neuron information includes self-adaptive membrane potential leakage information; calculating current pulse neuron output information according to the front end pulse neuron output information and the current pulse neuron information; updating the self-adaptive membrane potential leakage information according to the current pulse neuron output information; outputting the current pulse neuron output information.

Abstract: Systems and methods are provided for implementing hardware optimization for a hardware accelerator. The hardware accelerator emulates a neural network. Training of the neural network integrates a regularized pruning technique to systematically reduce a number of weights. A crossbar array included in hardware accelerator can be programmed to calculate node values of the pruned neural network to selectively reduce the number of weight column lines in the crossbar array. During deployment, the hardware accelerator can be programmed to power off periphery circuit elements that correspond to a pruned weight column line to optimize the hardware accelerator for power. Alternatively, before deployment, the hardware accelerator can be optimized for area by including a finite number of weight column line. Then, regularized pruning of the neural network selectively reduces the number of weights for consistency with the finite number of weight columns lines in the hardware accelerator.

Abstract: A method of machine learning model development includes building an autoencoder including an encoder trained to map an input into a latent representation, and a decoder trained to map the latent representation to a reconstruction of the input. The method includes building an artificial neural network classifier including the encoder, and a classification layer partially trained to perform a classification in which a class to which the input belongs is predicted based on the latent representation. Neural network inversion is applied to the classification layer to find inverted latent representations within a decision boundary between classes in which a result of the classification is ambiguous, and inverted inputs are obtained from the inverted latent representations. Each inverted input is labeled with a class that is its ground truth, and thereby producing added training data for the classification, and the classification layer is further trained using the added training data.

Abstract: Systems and methods are provided for estimating travel time and distance. Such method may comprise obtaining a vehicle trip dataset comprising an origin, a destination, a time-of-day, a trip time, and a trip distance associated with each of a plurality of trips, and training a neural network model with the vehicle trip dataset to obtain a trained model. The neural network model may comprise a first module and a second module, the first module may comprise a first number of neuron layers, the first module may be configured to obtain the origin and the destination as first inputs to estimate a travel distance, the second module may comprise a second number of neuron layers, and the second module may be configured to obtain the information of a last layer of the first module and the time-of-day as second inputs to estimate a travel time.

Abstract: A mechanism is described for facilitating slimming of neural networks in machine learning environments. A method of embodiments, as described herein, includes learning a first neural network associated with machine learning processes to be performed by a processor of a computing device, where learning includes analyzing a plurality of channels associated with one or more layers of the first neural network. The method may further include computing a plurality of scaling factors to be associated with the plurality of channels such that each channel is assigned a scaling factor, wherein each scaling factor to indicate relevance of a corresponding channel within the first neural network. The method may further include pruning the first neural network into a second neural network by removing one or more channels of the plurality of channels having low relevance as indicated by one or more scaling factors of the plurality of scaling factors assigned to the one or more channels.

Abstract: Methods, apparatus, systems and articles of manufacture are disclosed to map workloads. An example apparatus includes a constraint definer to define performance characteristic targets of the neural network, an action determiner to apply a first resource configuration to candidate resources corresponding to the neural network, a reward determiner to calculate a results metric based on (a) resource performance metrics and (b) the performance characteristic targets, and a layer map generator to generate a resource mapping file, the mapping file including respective resource assignments for respective corresponding layers of the neural network, the resource assignments selected based on the results metric.

Abstract: Provided is an electronic equipment offering a plurality of controllable functionalities and that receives information designating a functionality selected amongst this plurality of controllable functionalities. The electronic equipment also receives a plurality of recordings of a same group of at least one user input for this selected functionality. The electronic equipment also applies a trained neural network model on this received plurality of recordings to recognize an input pattern. The electronic equipment also associates this recognized input pattern with this selected functionality, and stores this association in a memory. The electronic equipment also controls this selected functionality based on this stored association.

Abstract: A neuromorphic neuron apparatus includes an accumulation block and an output generation block. The apparatus has a current state variable corresponding to previously received one or more signals. The output generation block is configured to use an activation function for generating a current output value based on the current state variable. The accumulation block is configured to repeatedly: compute an adjustment of the current state variable using the current output value and a correction function indicative of a decay behaviour of a time constant of the apparatus; receive a current signal; update the current state variable using the computed adjustment and the received signal, the updated state variable becoming the current state variable; and cause the output generation block to generate a current output value based on the current state variable.

Abstract: A system and method are disclosed for running a plurality of simulation tests on a first machine learning model to obtain a plurality of results that are each produced during a respective simulation test, the first machine learning model gradually trained using first training data historically collected over a period of time, the first training data comprising a plurality of first training data sets each including a subset of first training inputs and first target outputs associated with one of a plurality of points in time during the period of time, determining a simulation test of the plurality of simulation tests at which corresponding results of the first machine learning model satisfy a threshold condition, wherein the threshold condition is based on historical data at a first point in time of the plurality of points in time, identifying a first training data set of the plurality of first training data sets on which the first machine learning model used during the determined simulation test was trained,

Abstract: An integrated optical circuit for an optical neural network is provided. The integrated optical circuit is configured to process a phase-encoded optical input signal and to provide a phase-encoded output signal depending on the phase-encoded optical input signal. The phase-encoded output signal emulates a synapse functionality with respect to the phase-encoded optical input signal. A related method and a related design structure are further provided.

Abstract: Synaptic resistors (synstors), and their method of manufacture and integration into exemplary circuits are provided. Synstors are configured to emulate the analog signal processing, learning, and memory functions of synapses. Circuits incorporating synstors are capable of performing signal processing and learning concurrently in parallel analog mode with speed, energy efficiency, and functions superior to computers.

Abstract: A method of forming a neural network includes specifying layers of neural network neurons. A parameter genome is defined with numerical parameters characterizing connections between neural network neurons in the layers of neural network neurons, where the connections are defined from a neuron in a current layer to neurons in a set of adjacent layers, and where the parameter genome has a unique representation characterized by kilobytes of numerical parameters. Parameter genomes are combined into a connectome characterizing all connections between all neural network neurons in the connectome, where the connectome has in excess of millions of neural network neurons and billions of connections between the neural network neurons.

Abstract: Various implementations are related to an apparatus with memory cells arranged in columns and rows, and the memory cells are accessible with a column control voltage for accessing the memory cells via the columns and a row control voltage for accessing the memory cells via the rows. The apparatus may include neural network circuitry having neuronal junctions that are configured to receive, record, and provide information related to incoming voltage spikes associated with input signals based on resistance through the neuronal junctions. The apparatus may include stochastic re-programmer circuitry that receives the incoming voltage spikes, receives the information provided by the neuronal junctions, and reconfigure the information recorded in the neuronal junctions based on the incoming voltage spikes associated with the input signals along with a programming control signal provided by the memory circuitry.

Abstract: In example implementations described herein, there are systems and methods for processing sensor data from an equipment over a period of time to generate sensor time series data; processing the sensor time series data in a kernel weight layer configured to generate weights to weigh the sensor time series data; providing the weighted sensor time series data to fully connected layers configured to conduct a correlation on the weighted sensor time series data with predictive maintenance labels to generate an intermediate predictive maintenance label; and providing the intermediate predictive maintenance label to an inversed kernel weight layer configured to inverse the weights generated by the kernel weight layer, to generate a predictive maintenance label for the equipment.

Abstract: A computer-implemented method for training a random matrix network is presented. The method includes initializing a random matrix, inputting a plurality of first vectors into the random matrix, and outputting a plurality of second vectors from the random matrix to be fed back into the random matrix for training. The random matrix can include a plurality of two-terminal devices or a plurality of three-terminal devices or a film-based device.

Abstract: A distributed industrial energy operation optimization platform which is capable of automatically constructing intelligent models and algorithms, is divided into three parts: a modeling terminal, a background service and a human-computer interface. The models like data pre-processing, energy generation-consumption-storage trend forecasting and optimal scheduling decision models are encapsulated in the modeling terminal as different visualization modules facing with multiple categories production scenarios, by dragging which the complex functional models can be realized conveniently. The background service is capable of automatically constructing the training samples and the production plans/manufacturing signals series according to the device model requirements of each edge side, interacts with the trained intelligent models through corresponding interfaces, and the computing results are saved in the specified relational database.

Abstract: The present disclosure provides an operation apparatus and method for an acceleration chip for accelerating a deep neural network algorithm. The apparatus comprises: a vector addition processor module and a vector function value arithmetic unit and a vector multiplier-adder module wherein the three modules execute a programmable instruction, and interact with each other to calculate values of neurons and a network output result of a neural network, and a variation amount of a synaptic weight representing the interaction strength of the neurons on an input layer to the neurons on an output layer; and the three modules are all provided with an intermediate value storage region and perform read and write operations on a primary memory.

Abstract: A side information calculating unit (110) calculates side information for assisting either identification processing or classification processing. When there is a discrepancy between a processing result of either the identification processing or the classification processing, and the side information, the multilayer neural network (120) changes an output value of an intermediate layer (20) and performs either the identification processing or the classification processing again.

Abstract: A neural array may include an array unit, a first processing unit, and a second processing unit. The array unit may include synaptic devices. The first processing unit may input a row input signal to the array unit, and receive a row output signal from the array unit. The second processing unit may input a column input signal to the array unit, and receive a column output signal from the array unit. The array unit may have a first array value and a second array value. When the first processing unit or the second processing unit receives an output signal based on the first array value from the array unit which has selected the first array value and then the array unit selects the second array value, it may input a signal generated based on the output signal to the array unit which has selected the second array value.

Abstract: According to an embodiment, a neural network device includes a plurality of cores, and a plurality of routers. Each of the plurality of routers includes an input circuit and an output circuit. Each of the plurality of cores transmits at least one of forward direction data propagating in the neural network in a forward direction and reverse direction data propagating in the neural network in a reverse direction. The input circuit receives the forward direction data and the reverse direction data from any one of the plurality of cores and the plurality of routers. The output circuit or the input circuit selectively deletes the reverse direction data stored based on a request signal for requesting reception of data.

Abstract: A network-based work assignment platform is provided. Employers and workers are registered on the platform. A registered employer posts a temporary job on the platform. The platform matches a registered worker to the temporary job. The platform facilitates, coordinates, and monitors performance of the temporary job by the registered worker and processes payment to the registered worker for performance of the temporary job on behalf of the registered employer.

Abstract: An autonomic function is caused to execute in an artificial intelligence environment to detect a new problem space. Using the autonomic function, a first model is selected. The first model includes a first trained neural network corresponding to a first ontology. A second model is automatically identified. the second model includes a second trained neural network corresponding to a second ontology. A layer is autonomically extracted from the second model and inserted into a location in the first model. A vector transformation is automatically constructed to transform an output vector of a previous layer in an immediately previous location in the model relative to the location. The layer is automatically fused in the first model using the transformed output vector as input to the layer, the fusing forming a fused model that is operable on an ontology of the new problem space.

Abstract: The present disclosure describes aspects of health management for magnetic storage media. In some aspects, a media health manager determines, with a read channel, read metrics for a sector of magnetic storage media that resides in a zone of magnetic storage media. The media health manager accesses read metrics of the zone and updates the read metrics of the zone based on the read metrics determined for the sector to provide updated read metrics for the zone of magnetic storage media. A health score for the zone of magnetic storage media is then determined with a neural network based on the updated read metrics of the zone of magnetic storage media. By so doing, gradual wear of the magnetic storage media may be predicted using the health score, enabling replacement of a magnetic storage media device before failure to improve reliability or availability of data stored to the device.

Abstract: Embodiments of the invention provide a neural network comprising multiple functional neural core circuits, and a dynamically reconfigurable switch interconnect between the functional neural core circuits. The interconnect comprises multiple connectivity neural core circuits. Each functional neural core circuit comprises a first and a second core module. Each core module comprises a plurality of electronic neurons, a plurality of incoming electronic axons, and multiple electronic synapses interconnecting the incoming axons to the neurons. Each neuron has a corresponding outgoing electronic axon. In one embodiment, zero or more sets of connectivity neural core circuits interconnect outgoing axons in a functional neural core circuit to incoming axons in the same functional neural core circuit.

Abstract: Provided are methods for learning to identify safety-critical scenarios for autonomous vehicles. First state information representing a first state of a driving scenario is received. The information includes a state of a vehicle and a state of an agent in the vehicle's environment. The first state information is processed with a neural network to determine at least one action to be performed by the agent, including a perception degradation action causing misperception of the agent by a perception system of the vehicle. Second state information representing a second state of the driving scenario is received after performance of the at least one action. A reward for the action is determined. First and second distances between the vehicle and the agent are determined and compared to determine the reward for the at least one action. At least one weight of the neural network is adjusted based on the reward.

Abstract: Techniques and mechanisms to facilitate a data classification functionality by communicating feedback signals with a spiked neural network. In an embodiment, input signaling, provided to the spiking neural network, results in one or more output spike trains which are indicative of that the input signaling corresponds to a particular data type. Based on the one or more output spike trains, feedback signals are variously communicated each to a respective node of the spiking neural network. The feedback signals variously control signal response characteristics of the nodes. Subsequent output signaling by the spiking neural network, in further response the input signaling, is improved based on the feedback control of nodes' signal responses. In another embodiment, the feedback signals are used to adjust synaptic weight values during training of the spiking neural network.

Abstract: The method may include processing, by using a neural network, input feature maps of an image to obtain output feature maps of the image. The neural network may include a convolution part and/or a pooling part, and an aggregation part. The convolution part may include at least one parallel unit each of which contains two parallel paths, each path of the two parallel paths contains two cascaded convolution layers. The kernel sizes are 1 dimension and are different in different units. The pooling part includes at least one parallel unit each of which contains two parallel paths, each path of the two parallel paths contains two cascaded pooling layers. The size of filters of pooling is 1 dimension and is different in different units. The aggregation part is configured to concatenate results of the convolution part and/or the pooling part to obtain the output feature maps of the image.

Abstract: A method, system and computer-readable storage medium for performing a cognitive information processing operation.

Abstract: A method of operating an apparatus using a control system that includes at least one neural network. The method includes receiving an input value captured by the apparatus, processing the input value using the at least one neural network of the control system implemented on first one or more solid-state chips, and obtaining an output from the at least one neural network resulting from processing the input value. The method may also include processing the output with another neural network implemented on solid-state chips to determine whether the output breaches a predetermined condition that is unchangeable after an initial installation onto the control system. The aforementioned another neural network is prevented from being retrained. The method may also include the step of using the output from the at least one neural network to control the apparatus unless the output breaches the predetermined condition. Similar corresponding apparatuses are described.

Abstract: Systems and methods are provided for performing medical imaging analysis. Input medical imaging data is received for performing a particular one of a plurality of medical imaging analyses. An output that provides a result of the particular medical imaging analysis on the input medical imaging data is generated using a neural network trained to perform the plurality of medical imaging analyses. The neural network is trained by learning one or more weights associated with the particular medical imaging analysis using one or more weights associated with a different one of the plurality of medical imaging analyses. The generated output is outputted for performing the particular medical imaging analysis.

Abstract: A signal processing method and apparatus includes determining a first signal F1(t) output by a first neuron, processing the first signal F1(t) using q orders of synapse weight parameters wq(t), wq?1(t), . . . , w1(t) to obtain a second signal F2(t), and inputting the second signal F2(t) to a second neuron, where the second neuron is in a layer immediately subsequent to that of the first neuron.

Abstract: A pixel compensation circuit includes: an integration circuit, a comparison circuit, a timing circuit, and a processor, wherein the integration circuit is configured to integrate driving currents of a pixel circuit, and then output a first voltage; the comparison circuit is configured to receive the first voltage, compare the first voltage with a first reference voltage, and then output a first logic control signal; the timing circuit is configured to acquire a first working duration; and the processor is configured to acquire the first working duration, obtain, according to correlations between the pre-obtained working duration and the pixel driving currents, a target driving current, corresponding to the first working duration, of the pixel circuit, and obtain a compensation parameter according to the target driving current.

Abstract: A system configured to emulate a correlithm object processing system includes an input node, a first output node, and a second output node. The input node receives a real-world numeric value comprising a plurality of numerical digits, and a flag indicating a numeric system associated with the numeric value. The first output node receives a first one of the plurality of numerical digits and generates a first correlithm object associated with the first numerical digit. The second output node receives a second one of the plurality of numerical digits and generates a second correlithm object associated with the second numerical digit. A string correlithm object engine maps the first correlithm object to a first sub-string correlithm object of a string correlithm object, and maps the second correlithm object to a second sub-string correlithm object of the string correlithm object.

Abstract: A system and method may generate code to be used when executing neural networks (NNs), for example convolutional neural networks (CNNs) which may include one or more convolutional layers. For at least one convolutional layer, for each non-zero element in a kernel tensor or matrix associated with the convolutional layer, instructions may be generated or issued. For example, for each non-zero element, a vector broadcast instruction may be generated, and a fused multiply-add (FMA) instruction may be generated, having as parameters a register representing a portion of the output for the convolutional layer, a register storing input data for the convolutional layer, and a register or reference to memory storing the non-zero element. The software or code produced may be executed during convolutional operations, for example as part of a larger application such as a NN inference application.

Abstract: Disclosed is a data processing system that includes compile time logic configured to section a graph into a sequence of sections, and configure each section of the sequence of sections such that an input layer of a section processes an input, one or more intermediate layers of the corresponding section processes corresponding one or more intermediate outputs, and a final layer of the corresponding section generates a final output. The final output has a non-overlapping final tiling configuration, the one or more intermediate outputs have corresponding one or more overlapping intermediate tiling configurations, and the input has an overlapping input tiling configuration. The compile time logic is further to determine the various tiling configurations by starting from the final layer and reverse traversing through the one or more intermediate layers, and ending with the input layer.

Abstract: Disclosed are a method and an apparatus for detecting spike event or transmitting spike event information generated in a neuromorphic chip. The apparatus for detecting spike event generated in a neuromorphic chip may detect spike event information for a plurality of neurons included in the neuromorphic chip based on a neuron group.

Abstract: A system and computer program enable a candidate to anonymously apply for a job position at an organization. In response to submitting a resume or other data for the purpose of applying anonymously for a job, the system generates an anonymous profile for the candidate and provides it to the organization to which the candidate is applying. The system excludes the candidate's name from the anonymous profile. In certain embodiments, the system also excludes one or more of the following: the candidate's address, data that is indicative of a candidate's race, age, or gender, and data that is not relevant for the job role for which the candidate is applying. After reviewing the anonymous profile, the organization has the option to reject the candidate or explore the candidate further. In response to the organization rejecting the candidate, the system notifies the candidate of the rejection without revealing the candidate's identity to the organization.

Abstract: A static random-access memory (SRAM) system includes SRAM cells configured to perform exclusive NOR operations between a stored binary weight value and a provided binary input value. In some embodiments, SRAM cells are configured to perform exclusive NOR operations between a stored binary weight value and a provided ternary input value. The SRAM cells are suitable for the efficient implementation of emerging deep neural network technologies such as binary neural networks and XNOR neural networks.

Abstract: At least a subset of first processing units of a first arrangement of an array of processing units is assigned to perform computations of a first layer of a neural network and at least a subset of second processing units of a last arrangement is assigned to perform computations of a second layer. At least a second subset of the first processing units of the first arrangement is re-assigned to perform computations of a third layer. In one aspect, it is determined that a number of layers exceeds a number of arrangements of processing units of a systolic processing chip. A first arrangement of processing units of the number of arrangements is assigned to perform computations according to a first layer for a first set of forward propagations, and the first arrangement is assigned to perform computations according to a different layer for a second set of forward propagations.

Abstract: An integrated optical circuit for an optical neural network is provided. The optical circuit is configured to process a plurality of phase-encoded optical input signals and to provide a phase-encoded optical output signal depending on the phase-encoded optical input signals. The phase-encoded optical output signal emulates a neuron functionality with respect to the plurality of phase-encoded optical input signals. Such an embodied optical circuit uses the phase to encode information in the optical domain. A related method and a related design structure are further provided.

Abstract: Methods, systems, and apparatuses for adapting a predictive model for a manufacturing process. One method includes receiving, with an electronic processor, a plurality of data points for a plurality of manufactured parts and the predictive model. The predictive model outputs a label for a manufactured part provided by the manufacturing process indicating whether the manufactured part is accepted or rejected. The method also includes estimating, with the electronic processor, a drift for each of the plurality of data points and generating, with the electronic processor, an adapted version of the predictive model based on the predictive model and the drift for each of the plurality of data points. In addition, the method includes outputting, with the electronic processor, a label for each of the plurality of manufactured parts using the adapted version of the predictive model.

Abstract: A computer processor includes an on-chip network and a plurality of tiles. Each tile includes an input circuit to receive a voltage signal from the network, and a crossbar array, including at least one neuron. The neuron includes first and second bit lines, a programmable resistor connecting the voltage signal to the first bit line, and a comparator to receive inputs from the two bit lines and to output a voltage, when a bypass condition is not active. Each tile includes a programming circuit to set a resistance value of the resistor, a pass-through circuit to provide the voltage signal to an input circuit of a first additional tile, when a pass-through condition is active, a bypass circuit to provide values of the bit lines to a second additional tile, when the bypass condition is active; and at least one output circuit to provide an output signal to the network.

Abstract: The present disclosure relates to a communication method and system for converging a 5th-Generation (5G) communication system for supporting higher data rates beyond a 4th-Generation (4G) system with a technology for Internet of Things (IoT). The present disclosure may be applied to intelligent services based on the 5G communication technology and the IoT-related technology, such as smart home, smart building, smart city, smart car, connected car, health care, digital education, smart retail, security and safety services. A Processing Element (PE) implemented in an accelerator in a convolutional neural network, which includes a first buffer configured to transfer input data to one other PE, and a second buffer configured to transmit to an outside output data that is processed on the basis of the input data; and an operation unit configured to generate output data.

Abstract: In an autonomous driving control apparatus, a drive control unit determines, based on first and second images respectively captured by first and second cameras and an autonomous driving condition, a value of at least one controlled variable for autonomous driving of a vehicle, and outputs, to a vehicle control unit, the value of the at least one controlled variable to thereby cause the vehicle control unit to carry out a task of autonomously driving the vehicle. A camera monitor unit determines whether at least one of the first and second cameras has malfunctioned. The camera monitor unit limits, based on the at least one of the first and second directional regions corresponding to the at least of the first and second cameras, the autonomous driving condition when determining that the at least one of the first and second cameras has malfunctioned.

Abstract: A method for retraining an artificial neural network trained on data from an old task includes training the artificial neural network on data from a new task different than the old task, calculating a drift, utilizing Sliced Wasserstein Distance, in activation distributions of a series of hidden layer nodes during the training of the artificial neural network with the new task, calculating a number of additional nodes to add to at least one hidden layer based on the drift in the activation distributions, resetting connection weights between input layer nodes, hidden layer nodes, and output layer nodes to values before the training of the artificial neural network on the data from the new task, adding the additional nodes to the at least one hidden layer, and training the artificial neural network on data from the new task.

Abstract: Disclosed is a method for training memristive learning systems (MLSs) using stochastic learning algorithms and the training system apparatus designed to implement the stochastic learning algorithms.

Abstract: A system and method for controlling a nodal network. The method includes estimating an effect on the objective caused by the existence or non-existence of a direct connection between a pair of nodes and changing a structure of the nodal network based at least in part on the estimate of the effect. A nodal network includes a strict partially ordered set, a weighted directed acyclic graph, an artificial neural network, and/or a layered feed-forward neural network.

Abstract: A device includes a memory and a node engine. The memory stores first and second tables that values represented by correlithm objects. The first table provides an intermediate output value based on the difference between two input values. The second table provides an output value based on an amplification of the intermediate output value from the first table. The node engine receives a first input correlithm object and a second input correlithm object, and identifies an amplification correlithm object. The node engine determines an output correlithm object by using the first and second tables in conjunction with the first input correlithm object, the second input correlithm object, and the amplified value correlithm object. The output correlithm object is based on the amplified difference between the first input correlithm object and the second input correlithm object.

Abstract: A data processing method comprises: if detecting that a number of image data to be transferred is greater than zero wherein the number of image data is a product of a number of input image data and a number of output image data, and a first available storage space of a FIFO memory is greater than or equal to a storage space occupied by an N number of input image data, transferring the N input image data in a first memory to the first FIFO memory; if detecting that a number of weight data to be transferred is greater than zero wherein the number of weight data is a product of the number of input image data and the number of output image data, and a second available storage space of a second FIFO memory is greater than or equal to a storage space occupied by an M number of weight data, transferring the M weight data in a second memory to the second FIFO memory; when the number of input image data cached in the first FIFO memory and the number of weight data cached in the second FIFO memory are greater than or e

Abstract: In one embodiment, a method includes establishing access to first and second different computing systems. A machine learning model is assigned for training to the first computing system, and the first computing system creates a check-point during training in response to a first predefined triggering event. The check-point may be a record of an execution state in the training of the machine learning model by the first computing system. In response to a second predefined triggering event, the training of the machine learning model on the first computing system is halted, and in response to a third predefined triggering event, the training of the machine learning model is transferred to the second computing system, which continues training the machine learning model starting from the execution state recorded by the check-point.