International Business Machines Corporation

Ta-Hsin Li

Wen Jun Yin

Jin Dong

Bin Zhang

Xin Xin Bai

Ming Xie

Nianjun Zhou

Ying Li

Richard B Segal

Pu Huang

Kaan Kudsi Katircioglu

Axel Martens

Hai Rong Lv

Ashish Jagmohan

Krishna Chaitanya Ratakonda

Vadim Sheinin

Rakesh Mohan

Elham Khabiri

Pietro Mazzoleni

Wesley M Gifford

Aditya Vempaty

Li Jun Xia

Pavankumar Murali

Jin Yan Shao

Swapna Buccapatnam Tirumala

Matthew Daniel Riemer

T K Balachandran

Lisa A Smith

Cindy Q Zhang

Feng Li

Tamar Eilam

Jianying Hu

Ming Zhu

Jing Gao

Bonnie Kathryn Ray

German Sergio Goldszmidt

Heng Cao

Chidanand V Apte

David Hoffman

Gigi Y C Yuen-Reed

Kevin N Tran

Huijing Jiang

Zhong Lin Lin

Hua Liang

Bob Kelley

Thomas Halvorson

Giuseppe Andrea Palcologo

Stephen Scott Briggs

Mark A Eaton

Meng-Chen Hsieh

Techniques for mapping policy documents to regulatory documents to check for compliance between the policies and documents are provided. In one example, a computer-implemented method determining, by a system operatively coupled to a processor, an information input, a control framework, and a document from a first group consisting of a regulatory document and a policy document, wherein the information input is a corpora from a second group consisting of a domain corpora and a global corpora. The computer-implemented method can also comprise mapping, by the system, the received regulatory document or the received policy document to the control framework using a supervised machine learning technique.

Facilitating mapping of control policies to regulatory documents

A computer-implemented method for providing interpretable predictions from a machine learning model includes receiving a data structure that represents a hierarchical structure of a set of features (X) used by one or more predictive models to generate a set of predictions (Y). An interpretability model is built corresponding to the predictive models, by assigning an interpretability to each prediction Ybased on the hierarchical structure. Assigning the interpretability includes decomposing X into a plurality of partitions Xusing the hierarchical structure, wherein X=UX, N being the number of partitions. Further, each partition is decomposed into a plurality of sub-partitions using the hierarchical structure until atomic sub-partitions are obtained. A score is computed for each partition as a function of the predicted scores of the sub-partitions, wherein the predicted scores represent interactions between the sub-partitions. Further, an interpretation of a prediction is outputted.

Predictive models having decomposable hierarchical layers configured to generate interpretable results

The disclosure relates to extraction of rationales for studied outcome. A method comprises: grouping features as expert to align with a set of operating practices; generating interpretable features using operating rules, combining with statistical dependence analysis to bin selected features to generate favorite practice actions; grouping features as expert that combine a subset of the interpretable features to align with a set of operating practices. The method can also comprise: using a neural network or deep learning component to quantify contribution of respective experts at a consumer level applying a generic additive approach; extracting feature importance at an individual consumer-level decomposed from expert level importance; evaluating alternative, what-if, scenarios through sensitivity analysis to identify favorite practice actions; consolidating a subset of the practice actions at client or stakeholder levels; and routing respective practice actions as a function of responsibility for the set of operating practices to stakeholders or consumers.

Facilitating extraction of individual customer level rationales utilizing deep learning neural networks coupled with interpretability-oriented feature engineering and post-processing

Computer-implemented methods, systems, and computer readable media are provided for identifying and optimizing performance drivers of a healthcare related system. Healthcare related data may be analyzed to produce performance information pertaining to performance indicators for performance drivers of the healthcare related system. From the performance information, changes in the sets of performance indicators over time for the performance drivers may be determined and performance drivers with determined changes satisfying a threshold may be identified. An impact of the determined changes in the performance indicators to the identified performance drivers and contributions to the determined impact from one or more factors may be identified. Factors of the identified performance drivers with opposing utilization trends may be identified and an impact of the identified factors on the performance drivers may be determined. The identified performance drivers may be ranked and use of the performance drivers may be adjusted based on the ranking.

Searching and detecting interpretable changes within a hierarchical healthcare data structure in a systematic automated manner

A method for simulating participation patterns in a plurality of events from a pool of qualified participants includes selecting an event whose preference rank r is equal to a predetermined preference rank R(i,j) for an event of type j for individual i; sampling a random number Z that represents a number of events in which to participate when a number N(j) of available events of type j is greater than 0 and an expected participation rate P(i, j) of individual i for events of type j is greater than 0; selecting a random subset S with Z elements that indicates which Z events to participate in; and looping over k=1 to N(j) and setting a participation array M(i,j,k,l) to 1 iff k is contained in S, where participation array M(i,j,k,l) indicates whether or not individual i participates in a k-th event of type j in an l-th simulation.

Simulating selective participation of multiple events from a pool of qualified participants

A service engagement map may be generated based on data collected associated with the service transition and delivery processes. The service engagement map may be refined iteratively by discovering additional data associated with the service transition and delivery processes and updating the service engagement map according to the additional data. Engagement metrics may be computed based on the service engagement map and presented. The service engagement map may also be presented visually.

Metrics management and monitoring system for service transition and delivery management

Assignment scheduling for service projects, in one aspect, may comprise preparing input parameter data for servicing a client service request; generating a schedule for servicing the client service request by executing an optimization algorithm with the input parameter data; determining whether the schedule is acceptable by the client; and repeating automatically the preparing, the generating, the transmitting and the determining until it is determined that the schedule is acceptable by the client, wherein each iteration automatically prepares different input parameter data for inputting to the optimization algorithm and generates a different schedule based on the different input parameter data.

Scheduling for service projects via negotiation

A computer searching technique identifies high quantitative patterns in data. A spatial indexing technique, such as an R-tree is used to represent the data. Then a pattern searching algorithm is used to identify anchor points that define the componentwise minimum patterns. High quantitative patterns are found responsive to the componentwise minimum patterns. The search strategy is demonstrated relevant to the problem of finding suitable locations for a retail business with reference to environments of prior successful retail businesses.

High quantitative pattern searching using spatial indexing

A data integration module is operable to integrate a plurality of data sources, a customer preference module builds a model representing preference to different channels in merchandise category for each customer segment. A customer satisfaction module creates a model representing customer satisfaction metrics. A joint multi-channel optimization module is operable to use an optimization model that utilizes the customer preference model and the customer satisfaction model and maximize retailer's profit and customer satisfaction.

Maximizing retailer profit and customer satisfaction using multi-channel optimization

A method, and system employing the method, for service channel reconfiguration at a service outlet includes generating service transaction data of a service outlet, generating queue management system (QMS) data of the service outlet, and generating cost and profit data for the service outlet. Data is extracted from the service transaction data and the QMS data relating to specified parameters including customer experience data, and customer demand data. The service transaction data and the QMS data is integrated with the cost and profit data providing a unified objective function. Stochastic service processes and customer behavior data are modeled. The unified objective function is evaluated using the stochastic service processes and customer behavior data model, and the service channel function of the service outlet is reconfigured using the unified objective function.

Method and apparatus for optimal service channel reconfiguration based on multi-agent simulation

A method of analyzing a spectrum of one-dimensional or multi-dimensional signal X(t) requires a number of steps including deriving coefficients [A(&#x3c9;), B(&#x3c9;)] of an Lp-norm harmonic regression of the signal with 0&#x3c;p&#x2266;&#x221e; and p&#x2260;2, squaring the coefficients, summing the squared coefficients, and scaling the summed, squared coefficients with a constant c to realize a periodogram of X(t) as L(&#x3c9;)=c{[A(&#x3c9;)]+[B(&#x3c9;)]}. The method may include receiving the signal X(t), storing the received signal X(t), and outputting the periodogram L(&#x3c9;). The method may still further include scanning to maximize the periodogram L(&#x3c9;) by identifying its largest peak(s) and comparing the amplitude of the identified largest peak(s) with a threshold to determine if the largest peak(s) is(are) attributable to a presence of a periodic signal.

Robust spectral analyzer for one-dimensional and multi-dimensional data analysis

A system and method of determining at least one location for a retail outlet in a region are described. The system and method use clustering technology for partitioning the region into a fixed number of sub-regions. Then, the system and method compute marginal increments from input data for each sub-region. The system and method choose a sub-region having a maximal marginal increment for a location of a first retail outlet.

Method and apparatus for outlet location selection using the market region partition and marginal increment assignment algorithm

A system, method and computer program product for providing the ability for retailers to devise a current channel strategy (e.g., adaptive price setting) that considers competitors in a dynamic competing environment, and that enables computing a competitive advantage of a channel. To estimate a price for selling a product j in a commerce channel comprises: a) receiving, at a processor device, real market data including sales and price history data of a product j sold by one or more retailers over one or alternate sales channels t; generating, by the processor device, a competitive advantage parameter value based on the sales and price history data; and, computing, utilizing the competitive advantage parameter value, an optimum price for a particular product to be marketed in one of the one or alternate sales channel.

Competing simulator in multi-channel retailing environment among multiple retailers

Growing community of inventors

Danbury, CT, United States of America

Ta-Hsin Li