Nec Corporation

Nec Laboratories America, Inc.

Erik Kruus

Asim Kadav

Cristian Ungureanu

Renqiang Min

Biplob Kumar Debnath

Cezary Dubnicki

Srimat T Chakradhar

Hui Joyce Zhang

Hans Peter Graf

Masayuki Sakata

Murugan Sankaradas

Iain Melvin

Yuki Chiba

Hao Li

Krzysztof Lichota

Asako Fujii

Wen Xia

Yiren Jian

Chris White

Systems and methods for predicting T-Cell receptor (TCR)-peptide interaction, including training a deep learning model for the prediction of TCR-peptide interaction by determining a multiple sequence alignment (MSA) for TCR-peptide pair sequences from a dataset of TCR-peptide pair sequences using a sequence analyzer, building TCR structures and peptide structures using the MSA and corresponding structures from a Protein Data Bank (PDB) using a MODELLER, and generating an extended TCR-peptide training dataset based on docking energy scores determined by docking peptides to TCRs using physical modeling based on the TCR structures and peptide structures built using the MODELLER. TCR-peptide pairs are classified and labeled as positive or negative pairs using pseudo-labels based on the docking energy scores, and the deep learning model is iteratively retrained based on the extended TCR-peptide training dataset and the pseudo-labels until convergence.

T-cell receptor repertoire selection prediction with physical model augmented pseudo-labeling for personalized medicine decision making

A classification apparatus according to the present disclosure includes an input unit configured to receive an operation performed by a user, an extraction unit configured to extract moving image data by using a predetermined rule, a display control unit configured to display an icon corresponding to the extracted moving image data on a screen of a display unit, a movement detection unit configured to detect a movement of the icon on the screen caused by the operation performed by the user, and a specifying unit configured to specify a classification of the moving image data corresponding to the icon based on a position of the icon on the screen.

Classification apparatus, classification method and computer readable medium

A method is provided for sparse communication in a parallel machine learning environment. The method includes determining a fixed communication cost for a sparse graph to be computed. The sparse graph is (i) determined from a communication graph that includes all the machines in a target cluster of the environment, and (ii) represents a communication network for the target cluster having (a) an overall spectral gap greater than or equal to a minimum threshold, and (b) certain information dispersal properties such that an intermediate output from a given node disperses to all other nodes of the sparse graph in lowest number of time steps given other possible node connections. The method further includes computing the sparse graph, based on the communication graph and the fixed communication cost. The method also includes initiating a propagation of the intermediate output in the parallel machine learning environment using a topology of the sparse graph.

Communication efficient sparse-reduce in distributed machine learning

A computer-implemented method for emulating an object recognizer includes receiving testing image data, and emulating, by employing a first object recognizer, a second object recognizer. Emulating the second object recognizer includes using the first object recognizer to perform object recognition on a testing object from the testing image data to generate data, the data including a feature representation for the testing object, and classifying the testing object based on the feature representation and a machine learning model configured to predict whether the testing object would be recognized by a second object recognizer. The method further includes triggering an action to be performed based on the classification.

Object recognizer emulation

Systems and methods for building a distributed learning framework, including generating a sparse communication network graph with a high overall spectral gap. The generating includes computing model parameters in distributed shared memory of a cluster of a plurality of worker nodes; determining a spectral gap of an adjacency matrix for the cluster using a stochastic reduce convergence analysis, wherein a spectral reduce is performed using a sparse reduce graph with a highest possible spectral gap value for a given network bandwidth capability; and optimizing the communication graph by iteratively performing the computing and determining until a threshold condition is reached. Each of the plurality of worker nodes is controlled using tunable approximation based on available bandwidth in a network in accordance with the generated sparse communication network graph.

System and method for communication efficient sparse-reduce

Systems and methods for implementing content-level anomaly detection for devices having limited memory are provided. At least one log content model is generated based on training log content of training logs obtained from one or more sources associated with the computer system. The at least one log content model is transformed into at least one modified log content model to limit memory usage. Anomaly detection is performed for testing log content of testing logs obtained from one or more sources associated with the computer system based on the at least one modified log content model. In response to the anomaly detection identifying one or more anomalies associated with the testing log content, the one or more anomalies are output.

Content-level anomaly detector for systems with limited memory

A computer-implemented method and computer processing system are provided. The method includes synchronizing, by a processor, respective ones of a plurality of data parallel workers with respect to an iterative process. The synchronizing step includes individually continuing, by the respective ones of the plurality of data parallel workers, from a current iteration to a subsequent iteration of the iterative process, responsive to a satisfaction of a predetermined condition thereby. The predetermined condition includes individually sending a per-receiver notification from each sending one of the plurality of data parallel workers to each receiving one of the plurality of data parallel workers, responsive to a sending of data there between. The predetermined condition further includes individually sending a per-receiver acknowledgement from the receiving one to the sending one, responsive to a consumption of the data thereby.

Fine-grain synchronization in data-parallel jobs

A computer-implemented method and computer processing system are provided. The method includes synchronizing, by a processor, respective ones of a plurality of data parallel workers with respect to an iterative distributed machine learning process. The synchronizing step includes individually continuing, by the respective ones of the plurality of data parallel workers, from a current iteration to a subsequent iteration of the iterative distributed machine learning process, responsive to a satisfaction of a predetermined condition thereby. The predetermined condition includes individually sending a per-receiver notification from each sending one of the plurality of data parallel workers to each receiving one of the plurality of data parallel workers, responsive to a sending of data there between. The predetermined condition further includes individually sending a per-receiver acknowledgement from the receiving one to the sending one, responsive to a consumption of the data thereby.

Fine-grain synchronization in data-parallel jobs for distributed machine learning

A network device, system, and method are provided. The network device includes a processor. The processor is configured to store a local estimate and a dual variable maintaining an accumulated subgradient for the network device. The processor is further configured to collect values of the dual variable of neighboring network devices. The processor is also configured to form a convex combination with equal weight from the collected dual variable of neighboring network devices. The processor is additionally configured to add a most recent local subgradient for the network device, scaled by a scaling factor, to the convex combination to obtain an updated dual variable. The processor is further configured to update the local estimate by projecting the updated dual variable to a primal space.

System and method for fault-tolerant parallel learning over non-iid data

Systems and methods are disclosed for providing distributed learning over a plurality of parallel machine network nodes by allocating a per-sender receive queue at every machine network node and performing distributed in-memory training; and training each unit replica and maintaining multiple copies of the unit replica being trained, wherein all unit replicas train, receive unit updates and merge in parallel in a peer-to-peer fashion, wherein each receiving machine network node merges updates at later point in time without interruption and wherein the propagating and synchronizing unit replica updates are lockless and asynchronous.

Parallelized machine learning with distributed lockless training

Methods and systems for adaptive compression include compressing input data according to a first compression ratio; pausing compression after a predetermined amount of input data is compressed; estimating which of a set of ranges a compressed output size will fall within using current settings; and performing compression on a remainder of the input data according to a second compression ratio based on the estimated range.

Adaptive compression supporting output size thresholds

Systems and methods for generating and storing a data structure for maintaining cache supporting compression and cache-wide deduplication, including generating data structures with fixed size memory regions configured to hold multiple signatures as keys, wherein the number of the fixed size memory regions is bounded. A first mapping is generated from short-length signatures to a storage location and a quantized length measure on a cache storage device; and unused contiguous regions on the cache device are allocated. Metadata and cache page content is retrieved using a single input/output operation; a correctness of a full value of hash functions of uncompressed cache page content is validated; a second mapping is generated from short-length signatures to entries in the first mapping; and verification of whether the cached page content corresponds to a full-length original logical block address using the metadata is performed.

Bucketized multi-index low-memory data structures

System(s) and method(s) are provided for data management and data processing. For example, various embodiments may include systems and methods relating to relatively larger groups of data being selected with comparable or better performing selection results (e.g., high data redundancy elimination and/or average chunk size). In various embodiments, the system(s) and method(s) may include, for example a data group, block, or chunk combining technique or/and a data group, block, or chunk splitting technique. Various embodiments may include a first standard or typical data grouping, blocking, or chunking technique and/or data group, block, or chunk combining technique or/and a data group, block, or chunk splitting technique. Exemplary system(s) and method(s) may relate to data hashing and/or data elimination. Embodiments may include a look-ahead buffer and determine whether to emit small chunks or large chunks based on characteristics of underlying data and/or particular application of the invention (e.g., for backup).

Methods and systems for quick and efficient data management and/or processing

Systems and methods for data management and data processing are provided. Embodiments may include systems and methods relating to fast data selection with reasonably high quality results, and may include a faster data selection function and a slower data selection function. Various embodiments may include systems and methods relating to data hashing and/or data redundancy identification and elimination for a data set or a string of data. Embodiments may include a first selection function is used to pre-select boundary points or data blocks/windows from a data set or data stream and a second selection function is used to refine the boundary points or data blocks/windows. The second selection function may be better at determining the best places for boundary points or data blocks/windows in the data set or data stream. In various embodiments, data may be processed by a first faster hash function and slower more discriminating second hash function.

Growing community of inventors

Hillsborough, NJ, United States of America

Erik Kruus