James J. Rhodes

MINING PATENTS USING MOLECULAR SIMILARITY SEARCH

Text analytics is becoming an increasingly important tool used in biomedical research. While advances continue to be made in the core algorithms for entity identification and relation extraction, a need for practical applications of these technologies arises. We developed a system that allows users to explore the US Patent corpus using molecular information. The core of our system contains three main technologies: A high performing chemical annotator which identifies chemical terms and converts them to structures, a similarity search engine based on the emerging IUPAC International Chemical Identifier (InChI) standard, and a set of on demand data mining tools. By leveraging this technology we were able to rapidly identify and index 3,623,248 unique chemical structures from 4,375,036 US Patents and Patent Applications. Using this system a user may go to a web page, draw a molecule, search for related Intellectual Property (IP) and analyze the results. Our results prove that this is a far more effective way for identifying IP than traditional keyword based approaches.

BIwTL: a business information warehouse toolkit and language for warehousing simplification and automation

Rapidly leveraging information analytics technologies to mine the mounting information in structured and unstructured forms, derive business insights and improve decision making is becoming increasingly critical to todays business successes. One of the key enablers of the analytics technologies is an Information Warehouse Management System (IWMS) that processes different types and forms of information, builds, and maintains the information warehouse (IW) effectively. Although traditional multi-dimensional data warehousing techniques, coupled with the well-known ETL processes (Extract, Transform, Load) may meet some of the requirements in an IWMS, in general, they fall short on several major aspects: 1. They often lack comprehensive support for both structured and unstructured data processing; 2. they are database-centric and require detailed database and data warehouse knowledge to perform IWMS tasks, and hence they are tedious and time-consuming to operate and learn; 3. they are often inflexible and insufficient in coping with a wide variety of on-going IW maintenance tasks, such as adding new dimensions and handling regular and lengthy data updates with potential failures and errors. To cope with such issues, this paper describes an IWMS, called BIwTL (Business Information Warehouse Toolkit and Language), that automates and simplifies IWMS tasks by devising a high-level declarative information warehousing language, GIWL, and building the runtime system components for such a language. BIwTL hides system details, e.g., databases, full text indexers, and data warehouse models, from users by automatically generating appropriate runtime scripts and executing them based on the GIWL language specification. Moreover, BIwTL supports structured and unstructured information processing by embedding flexible data extraction and transformation capabilities, while ensuring high performance processing for large datasets. In addition, this paper systematically studied the core tasks around information warehousing and identified five key areas. In particular, we describe our technologies in three areas, i.e., constructing an IW, data loading, and maintaining an IW. We have implemented such technologies in BIwTL 1.0 and validated it in real world environments with a number of customers. Our experience suggests that BIwTL is light-weight, simple, efficient, and flexible.

BISON: providing business information analysis as a service

We present the architecture of a Business Information Analysis provisioning system, BISON. The service provisioning system combines two prominent domains, namely structured/unstructured data analysis and service-oriented computing. We also discuss open research problems in the area.

Complex Service Management in a Hybrid Cloud

This architectural feasibility study focuses on transforming a complex service management application to take advantage of the economies of scale and convenience afforded by cloud computing. We study a financial analytics application and system and enumerate the following characteristics: (a) highly compute intensive, (b) memory intensive, (c) storage i/o intensive, (d) collaborative, (e) interactive with low latency (propagation of changes must take place at speeds comparable to those of the spreadsheets that the application replaces in order to satisfy user responsiveness requirements) and (f) involves sensitive data for which it must maintain a high level of data security. We see two major challenges to operating such an application from the cloud: (1) achieving low latency interaction and (2) protecting sensitive data. In this paper we will explore alternative potential solutions to each of these challenges. To have any significant economic advantage, we must transform the architecture from heavy client and light server to light client and heavy server, running directly into challenge (1). It appears likely that any solution will involve a hybrid cloud, at least part of which provides high levels of data security (with consequent lowering of the economic advantage) in order to meet challenge (2). In our industrial research context, we are inclined to tackle one challenge at a time and to develop our new system in two phases, the first phase producing a new client server architecture that allows for multiple types of light clients, the second phase moving the server side into a hybrid cloud, from which multiple applications of the same type could be offered as web services to multiple clients.

Impact of cost uncertainty on pricing decisions under risk aversion

This paper studies cost uncertainty in services. Despite the fact that the service sector has become the largest component of gross domestic products in most developed economies, cost uncertainty and its impact on pricing decisions have not received much attention in the literature. In this paper, we first identify the root causes of cost uncertainty in services. Using the distinctive characteristics of services defined in the literature, we show why cost uncertainty, which has been widely neglected in the manufacturing dominated literature, is pervasive in services. Next, we investigate how cost uncertainty affects a risk-averse service provider’s pricing decisions in a make-to-order setting. Using the expected utility theory framework, we show that cost uncertainty increases the optimal price, whereas demand uncertainty reduces it. As a result of the countervailing impacts, the optimal price under risk aversion may be larger or smaller than the optimal risk-neutral price. Next, we study the problem of optimizing cost contingency in service contract pricing. We show that the optimal cost contingency decreases as the profit of the contract increases even when the utility function exhibits an increasing absolute risk aversion. Finally, we introduce various strategies to mitigate the risk of cost uncertainty observed in practice, and propose new research problems.

An analysis of a service system supporting IBM's global service delivery

The world economy continues to evolve away from the traditional producer-consumer roles prevalent in the industrial era, towards more dynamical, collaborative business relationships characteristic of the emerging global services economy. New challenges are emerging for businesses engaged in service delivery, particularly concerning process normalization and cost estimation. Service standardization is being addressed at IBM through careful analysis of historical service deals and the development of “best practice” standards for service solutions. The Solution Definition Manager (SDM) is a platform developed at IBM through a close collaboration among research and business representatives, designed to support the flexible creation of global service solutions within standards, as well as subsequent cost estimation and pricing of those solutions at fine level of granularity. In this paper, the SDM is discussed in terms of how and why it is being developed as well as the business transformation it represents. An analysis is presented of how research and business groups within IBM have co-created value by transforming internal processes, and how the SDM in turn allows IBM to be a better partner for value co-production while delivering services to external clients.

Optimal Asynchronous Service Scheduling in BISON

In this paper, we present some of our experiences from the architectural design phase of business information analysis service provisioning system, BISON. More specifically we discuss the issues, challenges, and our solution approaches to asynchronously scheduling the service requests in an optimal manner to enhance the user experience in the system

Enhancing Collaboration with IBM's Rational Jazz tm .

This paper describes our experience with IBM’s Rational Jazz tm platform for collaboration and for coordinating software development in the context of a medium sized service research and development project. We discuss the observed advantages of Jazz in systematizing the development process, especially when we are operating with extreme agility and the team is widely distributed around the world. We cover both narrative observations and quantitative measurements of Jazz usage. We demonstrate an objective measure of the value of such a software development management system. And we study the extent to which Jazz interfaces can replace ad hoc communication. While Jazz provides sufficient structure to replace all other communication within a geographically distributed research and development team, we conclude that redundant team communication in the forms of email and telephone meetings is necessary to maintain team motivation.