ESG recently completed testing of IBM Decision Optimization for Watson Studio, which is designed to enable organizations to accelerate the value they can extract from AI more easily. Testing examined how IBM Watson Studio with Decision Optimization collects data, organizes an analytics foundation, and analyzes insights at scale—with a focus on the ease of operationalizing AI and data science to improve trust, simplify compliance, and speed monetization.
ESG’s research indicates that the over-arching theme driving interest and adoption in AI/ML technology is related to improving operations. When looking at the top line data that combines responses of enterprise and midmarket organizations, 60% of respondents indicate that improving operational efficiency is one of the most important objectives their organization expects to achieve from their AI/ML investments (see Figure 1). When examined by company size, enterprises (organizations with 1,000 or more employees) view faster tactical response to shifting customer requirements, quicker time to market for product/services, and more predictive insight into future scenarios as more important objectives compared to midmarket organizations (organizations with 100 to 999 employees).1 For enterprises, the emphasis on quicker response times with more predictive insight is critical as they lack the flexibility to change their environment as quickly as midmarket firms.
While machine learning can be used across an organization, the current use is still focused within IT itself, and IT is leading the way both in terms of using ML and defining the ML initiative strategy. However, as Figure 2 shows, when ESG asked respondents how they measured the activity and effectiveness of AI and ML, three of the top five responses related to business metrics, key performance indicators (KPI), and/or responsiveness to business requests or questions.2 Clearly, AI and ML are becoming more strategic to the business.
As companies expand AI and ML beyond IT and into the lines of business and executive teams, they need a solution that can provide prescriptive analytics to satisfy the business, spanning on-premises, hybrid, and multi-cloud environments.
IBM Decision Optimization for Watson Studio
IBM describes a prescriptive approach—called the AI Ladder—to accelerate organizations’ journeys to AI. The AI Ladder is IBM’s guiding principle for organizations to embrace and transform their business by connecting data and AI. The AI Ladder provides companies four key areas designed to help businesses get started, from how they collect data, organize data, analyze data, and then ultimately infuse AI into their organization. Prebuilt app services and built-in expertise help accelerate time to value. ESG validated IBM Watson Studio and IBM Watson Machine Learning in 2018.3
As part of IBM’s prescriptive approach to AI, IBM Watson Studio and Watson Machine Learning are designed to bring hidden intelligence to the surface to help organizations transform business operations, on any cloud. This is critically important for organizations that hope to leverage AI and machine learning for prescriptive analytics, or decision optimization. Prescriptive analytics refers to the branch of advanced analytics that—as the name implies—helps prescribe best courses of action that decision-makers can take when faced with lots of alternative approaches. Optimization techniques are at the core of prescriptive analytics solutions.
Predictive and Prescriptive Analytics
Predictive analytics—at the most basic level—is supervised machine learning, where the answer is known, and the machine is taught how to find it. More advanced versions of predictive analytics include unsupervised learning, where the machine discovers patterns; reinforcement learning, where cumulative rewards indicate success; and deep learning, which is part of the broader family of machine learning methods based on artificial neural networks. These versions are used for pattern, image, object, and speech recognition, as well as text processing and time series analysis.
Prescriptive analytics applies mathematical and computational sciences to suggest optimal decision options to take advantage of the results of predictive analytics. In other words: You use mathematical algorithms to sift through all possible solutions and recommend the best one that optimizes a business goal, taking into consideration all decision variables, constraints, and trade-offs.
There are many parallels between machine learning and decision optimization. Both involve complex mathematics and algorithms; both require experts to effectively build and deploy solutions; both involve exponential network growth due to mixed integer program (MIP) trees and deep nets; and both often leverage weights and probabilities in stochastic processes.
Further, both technologies have been around for decades; the quality of results for both depend on the quality of the data; and both require analysis to verify the accuracy of results such as training for machine learning and analysis of what-if scenarios for decision optimization.
IBM Decision Optimization for Watson Studio is designed to bring the power of prescriptive analytics within reach of data science teams that want to drive operational efficiency and business impact across businesses and complete their ML approaches with optimization.
IBM Decision Optimization for Watson Studio is integrated with IBM Watson Studio, which lets data science teams combine optimization and machine learning techniques with model management and deployment capabilities—and other data science capabilities—to develop solutions that can help improve operational efficiency. Decision Optimization for Watson Studio contains multiple examples that show how ML and optimization can be combined to help data science teams learn and adopt prescriptive analytics faster. Data scientists can build prescriptive models, analyze scenarios, and solve optimization problems in the same platform and with similar tools that they currently use to build their machine learning models.
Data scientists can use IBM Decision Optimization for Watson Studio to speed the end-to-end process of building and deploying optimization models. Using the platform features, it is easy to connect to the same data sources that were used to create the machine learning models. What follows is the process of creating optimization models by coding in Python or using visual workflows and a modelling assistant for specific domains like scheduling. Graphical dashboards and the generation of multiple what-if scenarios makes it easy to test models and share results with other data scientists and business analysts. Once the model has been verified, the data scientists can deploy the model and optimization engine as a microservice. This microservice can be then be accessed from an application by business users.
IBM Decision Optimization for Watson Studio is designed to simplify the process of building optimization models and incorporates powerful optimization solvers like CPLEX to solve these models. Several application programming interfaces (APIs) allow IT to connect the model and engine to data sources and export the output to other systems.
Once the models are built using either Decision Optimization for Watson Studio or CPLEX Optimization Studio, the optimization models can be deployed on IBM Watson Machine Learning so that business users can access the optimization models from their applications.
IBM Cloud Pak for Data is an open, cloud-native information architecture for AI. Designed as an integrated, fully governed team platform, organizations can keep data secure at its source and add preferred data and analytics microservices as needed. Cloud Pak for Data can also be augmented with a Watson Studio Premium Add-on, a consumption-based value-added offering that includes IBM SPSS Modeler and Decision Optimization to further accelerate time to results and combine predictive and optimization models.
ESG Technical Validation
ESG performed evaluation and testing of IBM Decision Optimization for Watson Studio and IBM Cloud Pak for Data. Testing was designed to demonstrate how IBM Decision Optimization can help organizations make complex business decisions that typically involve multiple decision variables, trade-off possibilities, and complex constraints, with a goal of improving business operations and generating significant return on investment.
IBM Decision Optimization
To support the needs of organizations making critical business decisions involving thousands of decision variables and millions of alternatives, IBM Decision Optimization helps drive business results by enabling data science teams to solve complex problems using a combination of optimization technology and other data science techniques like machine learning within the unified IBM Watson Studio environment.
ESG looked at how IBM Decision Optimization can be used to accelerate and optimize business decisions using predictive and prescriptive analytics. As seen in Figure 4, we logged into IBM Watson Studio and opened a notebook that contained a detailed walkthrough of IBM Decision Optimization CPLEX modeling for Python (DOCplex) using Python and coding to the APIs. This notebook is based on a real-world use case where an organization plans to open new retail coffee shop locations and needs help deciding on optimal locations.
This use case leverages a combination of existing data, such as available capital, sales volume, relative performance of stores, and location; external factors, like the locations of public libraries; and business constraints, like locations of commissary kitchens and delivery routes. The model contains embedded libraries and links to open source libraries to enrich and visualize data, and prewritten code that users can plug into their model. Organizations can save their own notebooks to preserve and document successful models for future use.
Figure 5 shows two data visualizations. On the left are the locations of all public libraries in the Chicago area, and on the right are the optimal locations for the new coffee shops.
While this method was fast and efficient, it may be challenging for organizations that don’t have the skill set in house to code their own models. IBM provides a modeling assistant —a simpler method that enables building Decision Optimization inside IBM Watson Studio without coding. We began by creating a new project, called my_project, and created a new Decision Optimization model.
Next, we were prompted to select a method to formulate the model. Organizations can write their own code using general programming language APIs or Optimization Programming Language (OPL), import an existing model from a notebook or external file, or use the IBM Modeling Assistant to procedurally generate a model. We selected the Modeling Assistant.
Using this method, users are walked through the process of defining the model’s objectives and constraints. First, we imported the data set to work with. The data set we used represented resources and activities for a construction company. IBM Watson Studio integrates with over 120 data sources via built-in connectors to enable organizations to apply AI to their data wherever it lives.
The next step is to prepare the data. We visualized and edited the data directly in IBM Watson Studio, then selected the model type. We chose Scheduling. In its simplest form, a scheduling problem requires assigning tasks to resources. The Modeling Assistant presented us with pull-down lists for tasks and resources, which were automatically populated from our data set. We selected bridge_activity as the task to schedule and bridge_equipment as the resources to assign those activities to.
The IBM modeling assistant then displayed a visual representation of the model, including the objective and constraints. All data is presented in plain English, making it easier for users to understand exactly what the model’s goals, resources, and constraints are. On the right are suggestions. We searched for predecessors, precedence constraints, and conditions that needed to be satisfied before a task could start. At this point, we were ready to run our model.
We clicked Run to execute the model. Within a couple of minutes, the model returned our results, indicating that it had found an optimal model. Results are presented in table form, but most users will find the automatically-generated Gantt chart, seen in Figure 11, to be more useful.
The Gantt chart is interactive and can show predecessors for a task with a single click. This view makes it easy to create multiple scenarios for comparison, increasing or decreasing the amount of resources or time needed to complete, for example. Once the model is complete and tested, it can be deployed using IBM Watson Machine Learning. In the real world, an organization will want to run a model more than once, and IBM Watson Studio provides the ability to save a model and revisit it at any time.
Why This Matters
According to ESG research, 60% of organizations cited improving operational efficiency as one of the most important objectives their organization expects to achieve from their AI/ML investments. In the same survey, respondents cited business metrics or key performance indicator (42%) and/or responsiveness to business requests or questions (32%) as measures of success for their AI/ML initiatives.4 Clearly, AI and ML are becoming more strategic to the business.
As companies expand AI and ML beyond IT and into the line-of-business and executive teams, they need a solution that can provide prescriptive analytics to satisfy the business that can span on-premises, hybrid, and multi-cloud environments without the need for complex coding.
ESG testing revealed that IBM Watson Studio and IBM Decision Optimization provide simple tools that data scientists, application developers, and subject matter experts can use to work together to develop and build prescriptive analytics models at scale. Integrated data sources give the flexibility to build and train models where the data resides, and pretrained activities and notebooks accelerate the process. Extensive help and training resources keep it simple for non-experts.
IBM Watson Machine Learning enables organizations to deploy optimization engines to the cloud or data center environment with anytime, anywhere access—without having to worry about cloud cost overages for private data center deployment or about significant hardware investment for public clouds.
ESG testing revealed an optimized user experience with the right tools for all roles involved with Decision Optimization—data scientist, app developers, and IT admins. Organizations can deploy applications anywhere in a hybrid environment to operationalize prescriptive analytics faster and solve complex business problems at scale. ESG found that Watson Studio and Watson Machine Learning provide an end-to-end environment that helps organizations apply learning from production and quickly iterate while ensuring visibility across data science, application development, and business teams.
IBM Cloud Pak for Data
The pace of innovation in AI is accelerating, and while governance of data assets and compliance with policies get a lot of attention, little attention has been given to governing the complexity associated with the lifecycle management of data science and machine learning. The range of open source frameworks in data science makes governance difficult for the typical enterprise. IBM Cloud Pak for Data can provide seamless governance to asset curation with quality assessments (governance for insight) and policy-based enforcements (governance for compliance) through its integrated platform, encompassing key services from IGC, Data Stage, and Watson Studio Local. Cloud Pak for Data is designed to bring the power of IBM analytics to a simple validated package, repackaging proven technologies. As a foundational component of the AI Ladder, IBM Cloud Pak for Data helps to operationalize collection, organization, analytics, and infusion of AI seamlessly.
With IBM Cloud Pak for Data, organizations can modernize their information architecture (IA) and start their journey to becoming AI-driven. Cloud Pak for Data supports and governs the end-to-end AI workflow, as shown in Figure 12.
The Enterprise Catalog enables the right users to find the right data and analytics assets—indexed for search, with lineage, usage metrics, and quality profiles. Users can pull these assets into their analytics projects, where they can cleanse, shape, understand, and model their data. IBM Cloud Pak for Data supports open source and IBM frameworks—Spark, TensorFlow, IBM SPSS Modeler, CPLEX, etc.—with model management and deployment capabilities providing governance across dev, test, staging, and production. Models are versioned and scaled automatically through load balancing to meet SLAs. Model performance is automatically monitored and can trigger model retraining and redeployment to be released as rolling upgrades.
Figure 13 shows the project we created in Watson Studio. Watson Studio and Watson Machine Learning—along with all their assets—are embedded in IBM Cloud Pak for Data. Organizations can accelerate time to value and increase productivity of data science and business teams by purchasing a Watson Studio Premium Add-on to Cloud Pak for Data. This Watson Studio Premium Add-on includes SPSS Modeler, Decision Optimization, and Data Refinery and is designed to enable the business to mix and match capabilities based on the consumption model.
Why This Matters
In addition to the plethora of disparate tools, a lack of experienced, trained personnel presents an obstacle to organizations working toward operationalizing AI and ML. IBM Cloud Pak for Data is an open, cloud-native information architecture for AI.
ESG validated IBM Cloud Pak for Data’s cloud-native, fully governed platform, verifying that organizations can keep data secure at the source and add preferred data and analytics microservices as needed. Cloud Pak for Data’s consumption-based model enables organizations to move among tools easily—IBM SPSS to Decision Optimization with Watson Studio Premium Add-on.
The Bigger Truth
Organizations need to get more out of AI/ML than simply great insights. Businesses need to execute against their insights, and they need recommendations around how scarce resources should be allocated, how to schedule tasks, and how to deal with constraints for complex business projects. Enterprises view faster tactical response to shifting customer requirements, quicker time to market for product/services, and more predictive insight in future scenarios as more important objectives for their AI/ML investments compared to midmarket organizations.5
Businesses are faced with decisions that involve thousands of variables and operational constraints and tradeoffs to find solutions that optimize a specific objective like increased revenue or reduced cost. Relying on spreadsheets or simple analytic tools to address these challenges is more likely to result in suboptimal results as the complexity of the problems to be solved increases. Organizations need to run sophisticated mathematical models that factor in business resources, constraints, and trade-offs, evaluating scenarios to find the best course of action from millions of alternatives.
IBM Decision Optimization solutions are designed to enable business decision-making processes such as operational, tactical, or strategic planning and scheduling use cases by making it easy for decision makers to choose the optimal course of action from millions of alternatives while factoring in business constraints and compromises. Organizations can achieve faster time to value while making complex decisions using optimization solvers for large, real-world problems, at the speed required for today’s interactive decision optimization applications.
IBM Decision Optimization for Watson Studio helps organizations drive business results by using a combination of optimization technology and other data science techniques like machine learning within a unified IBM Watson Studio environment.
ESG testing validated that IBM Decision Optimization for Watson Studio can simplify and accelerate the process of developing and deploying optimization models. Users can easily create an optimization model by coding to the APIs or using the built-in Modeling Assistant. Using visual dashboards and point-and-click tools, users can simplify the process of testing multiple scenarios. The models can be easily deployed on Watson Machine Learning as a microservice—which business users can then access from their applications. Further, organizations can accelerate time to value and increase productivity of data science and business teams with Watson Studio Premium Add-on to Cloud Pak for Data. This Watson Studio Premium Add-on includes SPSS Modeler, Decision Optimization, and Data Refinery and is designed to enable the business to mix and match capabilities based on the consumption model.
Leveraging data and driving better business decisions and outcomes is a task that has traditionally been difficult to achieve. Choosing how and where to start, and the right tools for the job, are often difficult decisions. IBM’s analytics portfolio has been designed to support all organizations’ analytics needs, including descriptive, predictive, and prescriptive solutions.
With IBM Decision Optimization, organizations can model business issues mathematically and solve them with powerful algorithms to produce precise and logical decisions for improving efficiency, reducing costs, and increasing profitability. If you are seeking to drive improvements in operational efficiency, optimize business decisions, transform planning processes, and perform powerful scenario modeling to evaluate and compare potential outcomes, you’d be smart to take a close look at IBM Decision Optimization for Watson Studio.
1. Source: ESG Master Survey Results, Artificial Intelligence and Machine Learning: Gauging the Value of Infrastructure, March 2019.↩
3. Source: ESG Technical Validation, Hybrid Multi-cloud Artificial Intelligence (AI): IBM Watson Studio and Watson Machine Learning, February 2019.↩
4. Source: ESG Master Survey Results, Artificial Intelligence and Machine Learning: Gauging the Value of Infrastructure, March 2019.↩
ESG Technical Validations
The goal of ESG Technical Validations is to educate IT professionals about information technology solutions for companies of all types and sizes. ESG Technical Validations are not meant to replace the evaluation process that should be conducted before making purchasing decisions, but rather to provide insight into these emerging technologies. Our objectives are to explore some of the more valuable features and functions of IT solutions, show how they can be used to solve real customer problems, and identify any areas needing improvement. The ESG Validation Team’s expert third-party perspective is based on our own hands-on testing as well as on interviews with customers who use these products in production environments.