Gartner predicts that by 2028, 25% of chief data and analytics officer vision statements will become “decision-centric”, surpassing “data-driven” slogans, as human decision-making behaviors are modeled to improve Data &Analytics value. Link
Arash Aghlara goes further suggesting that Data & Analytics leaders should STOP thinking about “data” and “analytics” altogether. Link
Harshil Patel describes how to align Data-Driven and Decision-Centric approaches using Machine Learning. Link
Richard Susskind writes “about the use of AI to draft legislation and regulation and then to generate legal rules in the form of executable code.” He states that AI will draft complex legislations within five years and new tools will make breaking the law harder.” Link
Business decision models frequently use lookup tables that may contain thousands of data rows. There are at least two issues to be addressed when representing lookup tables: 1) users need to be able to modify the table data without modifications in the decision model; 2) the tables may contain hundreds of thousands of rows and still should be highly efficient when used by decision services. So, this month’s challenge provides a concrete example of lookup tables from the claim processing domain. Each claim includes lists of procedures and diagnoses with their types, codes, and other information. Your decision model needs to validate each claim against the standard lists of compatible and incompatible pairs of procedures and diagnoses represented as ICD-10 codes. Link
From an expert working on cutting edge neuro symbolic AI solutions, integrating LLM and decision management systems to bring real AI value to the enterprises: “Enterprise should not consider Gen AI as the solution to implement their interactive, customer facing, conversation agents. This is too risky business as of today. Rule-based systems have been deployed in the industries for more than 25 years, and are taking millions of decisions per day. Symbolic AI based on deep learning models, such as Generative AI, should be combined with knowledge graph, inference rules and ontologies as they make precise, reasoned, and explainable decisions.” Link Related articles: “Neuro-Symbolic AI” “How does neuro-symbolic AI actually work?“
Daniel Kahneman, a psychologist who won a Nobel Prize in economics for his insights into how ingrained neurological biases influence decision making, died Wednesday at the age of 90. Link Read “Thinking, Fast and Slow” for free.
Decision models (similarly to databases) frequently deal with analysis of collections of objects. This month’s Challenge deals with such a situation: you need to help an HR office create a rules-based service to analyze its employees. Each employee has a unique name, age, gender, marital status, locations (places of residence), number of children, salary, and more attributes. This information is coming to the service as a JSON request such as in this file. Your service should find answers to questions such as:
Arash Aghlara started a good discussion at LinkedIn: “Many times, based on the decisions that we execute on a specific case, we influence the future of the case. Although the case is the same but, it belongs to an altered situation where the decision we had executed has changed the situation of the case. Now, in this new state of the case, the already applied decision is either still applicable or we need a set of new decision models based on the new states of the case and the new situations the case belongs to.
The fact is that the case’s final outcome has not yet been determined by the first cycle of decision execution. We still need to continuously execute more decisions on the case until the final outcome is conclusive. Or we determine the outcome is inconclusive, so we need an alternative approach (manual review, domain expert input, marking the case infeasible, etc.) to finalize the case and find the best outcome for the case.” Link
Meinoff Sellmann, a well-known expert in Operation Research (OR), gives this advice: “No surfer has ever surfed a wave by paddling to a place where others are already surfing the tube. In 2024, generative AI is the latest fashion. And guess what many OR researchers are looking at right now. OR, focus on your craft. Focus on decision-making under uncertainty. And be prepared for the next wave. It is coming.” Link
Continue reading
George Keller Hart: “Large language models are potentially our civilization’s self blinding. How? By undermining our 5,000+ year heritage of written language – specifically because LLMs substitute memetics for attribution, breaking the first rule of shared, collective knowledge. Our vast corpus of recorded language has always rested on the provenance of each text, whether books, laws, correspondence, etc. – and across media formats as well. Attributed provenance is the foundation for evaluation and discernment, as opposed to LLMs that generate probabilistic ‘mimema’ – Greek for ‘imitations’.” Link
Tim Berners-Lee, the inventor of the web, marked the 35th anniversary of its creation (March 12, 1989) with a letter about the big challenges web technology faces as it moves towards its fourth decade of existence. He sees monopoly power and the exploitation of our personal data as the major issues currently keeping the web from reaching its potential. “The future hinges on our ability to both reform the current system and create a new one that genuinely serves the best interests of humanity.” Link
Decision Management Community 
You must be logged in to post a comment.