International Business Machines Corporation

University of Illinois

Rensselaer Polytechnic Institute

Kumar Bhaskaran

Yada Zhu

Santhosh Babu Kumaran

Yaoping Ruan

Milton H Hernandez

Ying Huang

Sambit Sahu

Lawrence David Bergman

Hai Huang

Pin-Yu Chen

Robert Alan Flavin

Fenno F Heath, Iii

Winnie W Cheng

Prabir Nandi

Ravi B Konuru

Nitin R Nayak

Jian Wang

Maja Vukovic

Jun Zhu

Jonathan Lenchner

Nitin Gaur

Wei Zhang

Jinjun Xiong

Gopal Sarma Pingali

Yang Zhang

Rong Nickle Chang

Frederick Y Wu

Xiaodong Cui

Ying Tat Leung

Lucian Popa

Sijia Liu

Robert Allen Hood

Girish Bhimrao Chafle

Zhong Tian

Toby Kurien

Daniel Patrick Connors

Thomas Dean Rosinski

John Anthony Bellacicco

Upendra D Chitnis

Igor Anatolyevich Naumov

Gerhard Sigl

Keith C Houck

Reha Yurdakul

Sanjay Elathur Ramaswamy

Shiwa S Fu

Guangnan Ye

Stephen K Kibby

Tian J Chao

Paul Driscoll

Rajesh Chopra

Rainer Kerth

Koushik Kar

Giovanni Mariani

Bo Li

Tirthankar Bandyopadhyay

Christina Derra

Guillermo Gallego

Di Chen

Yunan Ye

Aparna Gupta

Sai Radhakrishna Manikant Sarma Palepu

Toyotaro Suzumura

A graph neural network can be built and trained to predict a risk of an entity. A multi-relational graph network can include a first graph network and a second graph network. The first graph network can include a first set of nodes and a first set of edges connecting some of the nodes in the first set. The second graph network can include a second set of nodes and a second set of edges connecting some of the nodes in the second set. The first set of nodes and the second set of nodes can represent entities, the first set of edges can represent a first relationship between the entities and the second set of edges can represent a second relationship between the entities. A graph convolutional network (GCN) can be structured to incorporate the multi-relational graph network, and trained to predict a risk associated with a given entity.

Multi-relational graph convolutional network (GCN) in risk prediction

A method for improving computing efficiency of a computing device for temporal directed cycle detection in a transaction graph includes preparing the transaction graph based on a plurality of transactions, the transaction graph including nodes indicating transaction origination points and transaction destination points, and edges indicating interactions between the nodes. Irrelevant nodes in the transaction graph are identified and pruned to provide a pruned, preprocessed transaction graph which can be partitioning into sections, where each section includes selected nodes that are linked to other linked nodes therein. Each of the sections having non-cyclic nodes can be trimmed prior to performing cycle detection on the resulting pruned transaction graph. Postprocessing pruning can be performed to further reduce the number of detected cycles that may be of interest to a particular application, such as in anti-money laundering.

Temporal directed cycle detection and pruning in transaction graphs

A method, computer system, and a computer program product for energy efficient data exchange is provided. The present invention may include a first electronic card device (ECD) including a first switch and a second switch. The present invention may include the first switch being configured to power on the first ECD responsive to the first ECD engaging a second ECD. The present invention may include the first ECD being configured to exchange data with the second ECD. The present invention may include a docking component configured to receive the first ECD. The present invention may include the docking component including an actuator configured to engage the second switch to power on the first ECD when the first ECD is received by the docking component. The present invention may include the first ECD configured to transfer received data from the second ECD to a mobile device associated with the docking component.

Energy efficient electronic card

A processor training a reinforcement learning model can include receiving a first dataset representing an observable state in reinforcement learning to train a machine to perform an action. The processor receives a second dataset. Using the second dataset, the processor trains a machine learning classifier to make a prediction about an entity related to the action. The processor extracts an embedding from the trained machine learning classifier, and augments the observable state with the embedding to create an augmented state. Based on the augmented state, the processor trains a reinforcement learning model to learn a policy for performing the action, the policy including a mapping from state space to action space.

State-augmented reinforcement learning

Methods and systems for explainable machine learning are described. In an example, a processor can receive a data set from a plurality of data sources corresponding to a plurality of domains. The processor can train a machine learning model to learn a plurality of vectors that indicate impact of the plurality of domains on a plurality of assets. The machine learning model can be operable to generate forecasts relating to performance metrics of the plurality of assets based on the plurality of vectors. In some examples, the machine learning model can be a neural attention network with shared hidden layers.

Explainable machine learning based on heterogeneous data

A task-based learning using task-directed prediction network can be provided. Training data can be received. Contextual information associated with a task-based criterion can be received. A machine learning model can be trained using the training data. A loss function computed during training of the machine learning model integrates the task-based criterion, and minimizing the loss function during training iterations includes minimizing the task-based criterion.

Task-oriented machine learning and a configurable tool thereof on a computing environment

A deep-learning neural network can be trained to model a probability distribution of the asset-price trends for a future time period using a training data set, which can include asset-price trends of a plurality of assets over a past time period and a latent vector sampled from a prior distribution associated with the asset-price trends of a plurality of assets. The training data set can represent a time series data. A portfolio optimization can be executed on the modeled probability distribution to estimate expected risks and returns for different portfolio diversification options.

Generative network based probabilistic portfolio management

Access is obtained to a plurality of intermediately transformed electronic documents (with a plurality of sections and subsections) which have been transformed, by topical analysis and text summarization techniques, from a plurality of original electronic documents comprising at least some unstructured electronic documents. Audit and retrieval agent code is appended to the sections and subsections to create a plurality of finally transformed electronic documents. Users are allowed to access the finally transformed electronic documents. The users are provided with accountability reminders contemporaneous with the access. The access of the users to the sections and subsections of the finally transformed electronic documents is logged. An audit report is provided based on the logging. Also provided is a cloud service for enterprise-level sensitive data protection with variable data granularity, using one or more one guest virtual machine images.

Enterprise-level data protection with variable data granularity and data disclosure control with hierarchical summarization, topical structuring, and traversal audit

A method, system and computer program product for autonomic security configuration may include controlling a security configuration of at least one resource forming a solution based on a plurality of security requirements. The method may further include applying the plurality of security requirements across a plurality of resources independent of a resource type.

Method and system for autonomic security configuration

Auditing system logs of a remote client device is provided. Login session information entered at a workstation device accessing the remote client device to perform an activity associated with a work request is received. An access token is generated based on the login session information and information associated with the work request on the remote client device. The access token is compared with an audit log report of the remote client device that includes the activity associated with the work request performed by the workstation device on the remote client device. It is determined whether information in the access token matches information in the audit log report of the remote client device. In response to determining that the information in the access token does not match the information in the audit log report of the remote client device, an action alert is sent.

Service compliance enforcement using user activity monitoring and work request verification

Process Brokering Services (PBS) are implemented though the concept of Adaptive Documents to facilitate electronic commerce (e-commerce). PBS provides a single point of process control over the various fragmented execution flows and brings together the elements for process integration (views, content, flows) in a unified, scalable architecture on an industry standard platform. The two principal functions of the PBS are brokering of multiple business processes encapsulated in various back-end systems including workflow engines and business applications, and aggregating content from multiple enterprise information systems in the business context and managing the shared access to this based on the roles of the participants. The dynamic services provided by PBS are accessible to clients through the PBS Interface.

Method and system for process brokering and content integration for collaborative business process management

An electronic chronicling method and system that automatically captures, stores, and organizes various aspects of business processes in real-time. The system allows different role players to analyze, mine, and share selected performance metrics of business processes in the context of various measures involved, including users, systems, applications, processes, etc. The system enables users to analyze and utilize sortable performance metrics with varying degrees of specificity and detail. This analysis and utilization provides an efficient means for the real-time streamlining and optimization of business processes.

Chronicling for process discovery in model driven business transformation

A method for configuring a software package for an enterprise includes gathering desired To-be configuration requirements for a software package through at least one questionnaire; gathering current As-Is raw data for the enterprise; displaying the To-Be configuration requirements and As-Is raw data; comparing the To-Be configuration requirements and As-Is raw data; selecting final To-Be configuration requirements; and generating and executing configuration code for the software package to pre-configure the software package with the final To-Be configuration requirements.

Methods for configuring software package

Systems and methods are provided for the management of assets within an enterprise. Management of assets includes the creation of assets, the derivation and lineage of assets and asset composition. In addition, the creation of artifacts by the assets is provided. Assets and artifacts are maintained within a repository that is accessible throughout the enterprise. In addition, rules are provided for the composition and usage of all assets within the enterprise.

Model, design rules and system for asset composition and usage

Growing community of inventors

Englewood Cliffs, NJ, United States of America

Kumar Bhaskaran