AI & Machine Learning in Excel 5

What's inside

AI & Machine Learning in Excel 5: 12 courses, 240 hands-on lessons, with a full solutions appendix.

Learn by doing. Every lesson starts with a clear goal and a short explanation, shows the worksheet cell by cell, and gives you tasks to solve on your own computer in Excel, LibreOffice Calc or Google Sheets. When you are done, check your formulas against the solutions and hints in the appendix and correct your sheet.

• 12 complete courses, 240 lessons, 240 exam questions with answers
• Worksheets with row and column labels – easy to type in
• Tasks with exact solutions and hints in the appendix
• All figures are fictional practice data

Every course is also available as an interactive course on tabellab.com, where each cell is checked instantly.

Courses in this book

• AI Key Concepts in Numbers: A Beginner's Glossary Lab – Rules vs Learning, Training vs Inference, Parameters, Tokens & Context Windows, Embeddings, Temperature, Accuracy & Errors
• RAG Pipelines by Hand: Chunk Size & Overlap, Context Packing, Reranking, Groundedness & Citation Checks, Answer Scoring, Index Freshness, Latency and Cost per Answer
• LLM App Pipelines by Hand: Multi-Step Chains, Routing Branches, Conversation Memory Strategies, Summaries, Structured Outputs, Fallback Models & Prefix Caching
• Fine-Tuning by Hand: Full vs Adapter Training Memory, LoRA Parameters & Ranks, 8- and 4-Bit Quantisation, Adapter-on-Quantised-Model Budgets, Training Tokens & Loss Logs
• How Language Models Are Trained & Aligned by Hand: Next-Token Loss, Perplexity, Loss Curves, Instruction Data Mixes, Preference Pairs, Reward Scores, KL Penalties & Win Rates
• Multi-Agent Workflows by the Numbers: Topologies, Parallel Latency, Hand-Off Tokens, Routing Accuracy, Voting Ensembles, Critic Loops, Failure Propagation & Cost per Task
• Tool Calling & Tool Protocols by the Numbers: Catalogue Tokens, Dynamic Tool Loading, Argument Validity, Result Truncation, Server Latency & Timeouts, Parallel Calls & Caching
• AI Coding Assistants by the Numbers: Suggestion Acceptance, Retained Code, Generated Tests & First-Run Pass Rates, Coverage Gains, Attempts to Green, Review Minutes and Defect Escape
• AI Image Production by the Numbers: Asset Lists, Aspect Ratios & Pixels, Keep Rates, Cost per Approved Image, Style Scores, Variants, Upscaling & Campaign Budgets
• AI Filmmaking by the Numbers: Storyboards & Shot Lists, Clip-Length Limits, Generations per Usable Clip, Cost per Second, Queue Times, Continuity Checks & Cost per Finished Minute
• Training Data Curation by Hand: Source Inventories, Exact & Near-Duplicate Removal, Quality Filters, Mixture Weights & Epochs, Test-Set Overlap, Synthetic Shares and a Data Card
• Regularisation by Hand: Bias–Variance & Model Complexity, Learning Curves, Ridge Penalties, Coefficient Paths, L1 Shrinkage to Zero, Elastic Net, Scaling and Choosing λ

Details

Format	PDF, US Letter (8.5 × 11 in)
Language	English
Pages	709
Series	The Tabellab Workbook, volume 5
Author	Babala Özdemir
Delivery	Instant download after payment