ConnectOS

ReadOS / Technology

AI and Machine Learning in Practice — Part 1

Understanding artificial intelligence, language models and practical automation decisions.

24 chapters · About 255 minutes · Free to read

Start reading →
KEVOS / READOSTECHNOLOGYAI and MachineLearning inPractice — Part 1READOS COLLECTION

Contents

  1. 01Acting when you can't see everything: beliefs, filters and partial observabilityMany decisions must be made without knowing the true situation. How beliefs, Bayesian filters, Kalman and particle filters and belief-state planning handle partial observability.
  2. 02Adopting AI in a small business: start with the work, choose bets deliberately, redesign tasksAI creates value when it changes a workflow or decision, not when a tool is installed. How to pick use cases as bets, redesign tasks before changing roles and keep people accountable.
  3. 03AI agents and automated workflows: what they can safely do and how to keep them under controlAI agents can plan and act across systems, but each step adds risk. When to use a fixed workflow instead, and how to bound agents with permissions, approvals, limits and audit trails.
  4. 04Bayesian networks: mapping how causes and evidence connectBayesian networks represent uncertain relationships as a map of influences. How they work, how they support diagnosis and prediction, and what they can and cannot tell you about cause and effect.
  5. 05Be careful what you reward: designing objectives for decisionsAutomated systems and people both optimise what they are measured on. How objectives go wrong, how to balance multiple goals with trade-off analysis, and how to design rewards and KPIs that work.
  6. 06Data readiness is a business habit, not an IT projectMessy data is usually a symptom of how work is done. How to define good-enough data for each use, give information an owner, control access and fix errors where they start.
  7. 07Decision making under uncertainty: agents, observations and actionsA plain-English introduction to how people and machines make decisions when outcomes, models, situations and other actors are uncertain, and the main ways to design decision systems.
  8. 08Evaluating generative AI: gold datasets, rubrics, grounding checks and regression testsGenerative AI changes silently when prompts, models or documents change. How to build gold datasets and rubrics, use judge models carefully, test safety and gate every release.
  9. 09Fair use of AI in hiring and people decisions: bias, testing and human reviewAI tools can screen applications and rate staff quickly, and can repeat past unfairness at scale. How to choose, test and govern them so people decisions stay fair and explainable.
  10. 10Finding and designing AI use cases that pay: from business problem to a scored, specified opportunityLong lists of AI ideas rarely become value. A disciplined discovery method: frame the problem, map the work, set requirements, check data readiness and score value, feasibility and risk.
  11. 11Governing AI decisions: what you have let a machine decide, and oversight that actually worksAutomated decisions often slip from advice to action without anyone authorising it. How to record what machines may decide, set limits, make human review real and plan how to switch it off.
  12. 12How large language models work, for business decision-makers: tokens, training, context and limitsYou do not need to build an AI model to use one well, but you do need to know how it works. Tokens, training, context windows, settings and why models invent things, in plain terms.
  13. 13Is the AI good enough to rely on? Benchmarks, test scenarios and the decision to deploy, support or shelveA strong demo is not evidence. How to compare AI with your current process, test it on unseen and difficult cases, set acceptance rules first and decide to deploy, support or shelve.
  14. 14Is your AI a feature you sell or a tool you use? Different owners, economics and risksAI inside what you sell and AI inside how you work are different assets. How to give each the right owner, cost model and measures, and what changes when a tool becomes a product.
  15. 15Learning from experts: imitation learning and its limitsSystems can learn by copying expert demonstrations. How behavioural cloning, cascading errors, expert correction and inverse reinforcement learning work, and what they teach about know-how.
  16. 16Looking ahead: tree search, rollouts and planning by simulationWhen problems are too large to solve completely, decision makers plan from the current situation by looking a few steps ahead. How lookahead, rollouts, Monte Carlo tree search and re-planning work.
  17. 17Machine learning in a small factory: inspection, maintenance and forecasting, and what each needs to workMachine learning can inspect parts, flag failing machines and forecast demand, given the right data and checks. How to match the method to the problem, test it honestly and keep it working.
  18. 18Prompting, retrieval or fine-tuning: choosing how to adapt an AI model to your businessWhen an AI model gets things wrong, the fix is often not fine-tuning. How to choose between better prompts, retrieval, tools and fine-tuning, and what each costs to build and maintain.
  19. 19Putting company knowledge to work with retrieval-augmented generation: how it works and how to build it wellAn AI assistant that answers from your own documents can save hours or confidently quote the wrong version. How retrieval-augmented generation works, where it fails and how to build it.
  20. 20Reinforcement learning explained: learning good decisions from experienceReinforcement learning lets systems learn decision strategies by trial and error, guided by rewards. A plain-English guide to how it works, its main methods, its risks and where it fits in business.
  21. 21Securing AI systems: prompt injection, data leakage, poisoned documents and least privilegeAI systems can be talked into misbehaving by the content they read. How prompt injection, poisoned documents, data leakage and excessive agency work, and the controls that contain them.
  22. 22Sequential decisions: planning when today's choice shapes tomorrowMany decisions are steps in a sequence. A plain-English guide to Markov decision processes, policies, value functions, discounting and dynamic programming, with business examples.
  23. 23Starting small with AutoLISP automationTurn a repeated drafting task into a small AutoLISP routine with explicit inputs, predictable selections and checks for cancellation, units and drawing state.
  24. 24Taking AI from pilot to production: integration, operations, adoption and benefitsMany AI pilots impress, then stall. How to deliver in thin slices, meet production standards for security, monitoring and support, drive real adoption and show benefits against a baseline.

All chapters preserve the original KEVOS Insights content. These topic-based reading collections are organised alphabetically; chapters may be read in any order.

← All books