We start on a premise that the brain of all living species is a language-independent system for acquiring knowledge about objects and with finite space, where each bit of information placed there needs to be chosen carefully, organized and stored in a way that is both useful and accessible – and use that knowledge to discover new knowledge, reasoning and inferences through acquisition of symbolic as well as situated learning and build capabilities. These objects and concepts give meaning to the words we learn later. Individuals often do multiple activities at once, fragment a “major” decision or an objective (a particular macrostate) into a series of multiple “smaller” decisions(the number of microstates) – multiplicity as a function in the decision journey – and actions to measure their progress accurately. However, some learning methods, where the result is the final reward or pay-off, are awfully hard to untangle the future information to foresee the sequence of actions that will benefit the user at some point in future. Some of these infrequent and delayed rewards or learnings limit decisions making process. For some combinatorial problems, where all rules and information are known to all parties, one may set up intermediate positions for them to achieve the optimal results where, as in real-life conditions, such learning success depends on how well one would fragment this “major” decision.

What Holmes offers isn’t just a way of solving a crime, it is an entire way of thinking or acquisition of knowledge. … It is an approach born out of the scientific method that transcends science and crime both and can serve as a model for cognition, a way of being, even, just as powerful in our time as it was in Conan Doyle’s.

Mindful observation is just a first step. It’s a means to a far larger, far more practical and practically gratifying choice-sets and reward. Human decision making routinely involves choice among temporally extended courses of action, response and reward, as pay-off, over a broad range of time scales depending on cognitive state. Consider a traveler deciding to undertake a journey to a distant city for work. To decide – go-no-go – the end-benefits in terms of reward, as pay-off, of the trip must be weighed against the cost. Having decided to go, choices must be made at each fragmented “smaller” decision e.g., whether the work is worth paying or not, whether to fly or to drive, whether arrange a local accommodation or stay with friends or relatives.

One of the fundamental principles in precision engineering is that of determinism where systemic behavior is fully predictable, even to an individual’s, or atomic-scale, activities. To do the job efficiently and correctly, one needs models and algorithms, where the basic idea is that machine follows a set of rules, cause and effect relationships, that are within human ability to understand and control and that there is nothing random or probabilistic about their behavior. Further, the causalities are not esoteric and uncontrollable, but can be explained in terms of familiar and precise engineering principles. Intelligence, on the other hand, as opposed to fact, is stochastic in nature. It finds optimal solutions, derives reasons, infers actions, recognizes patterns, comprehends ideas, solves problems and uses language to communicate, from (im)perfect and (in)complete information conditions.

In its broadest application, it is a means for constructing objective-driven models, improving overall decision making and judgment ability, starting from the most basic building block of your own mind. These large-scale models have led to systems that understand when we talk or write, such as the natural-language processing and understanding programs we use every day, from digital assistants to speech-to-text programs.

With a pre-trained RIME – as the cognitive system and foundation models (ensembles) – we reduce labeled data requirements dramatically. First, we fine-tune it domain-specific reasoning and inferences to create a domain-specific foundation model. Then, using a much smaller amount of labeled data, potentially just in millions examples, we train a model for optimal choice-sets. The domain-specific RIME – as  the cognitive system and foundation model – has the potential to use for many tasks as opposed to the previous technologies that required building models from scratch in each use case.

Building on RIME – as the cognitive system and foundation models (ensembles) – platform has several benefits and here highlights some of the major benefits:

Forecast demand flow and likelihood of density: Designed to manage network demand flows in consumer journeys that organize decisions such as pricing, promotion, part assortment, resource assignment, etc.

Optimize costs and risk exposure with better reasoning, causality and inferences: a better interpretability of causalities, infer-ability of reasoning, proactivity in choice-setting and explainability of choice-set.

Objective-driven multi-layered multi-dimensional latent variables for choice-set (options): “Give me ordered recommendations I wouldn’t have thought of myself” cues optimization of value, speed and scale through augmented approach to maximize yields, throughput, resource capacity and dynamic resource assignments

Improve Quality of Decisions on Setting Choice-sets For End-Customers: “Trigger me when I’m in shopping mode” – Timing is just as important as content cues. This improves determining the optimal options for a set of decisions in a given business environment and business target, thus recommend consistent feasible choices for end customers