menu
{ "item_title" : "The Terrain of Learning", "item_author" : [" Zixi Li "], "item_description" : "The Terrain of Learning is a concise introduction to reasoning geometry and intelligent dynamics.Instead of treating deep learning as a list of formulas, this book asks a different question: what if learning is motion through a terrain? Loss becomes an energy landscape, gradients become directions, optimizers become ways of walking, and reasoning becomes a trajectory through hidden states and belief space.Across twelve chapters, Zixi Li develops a geometric language for modern AI: parameter space, representation space, Bregman divergence, KL divergence, fixed points, ResNet, GPT, DEQ, chain-of-thought, LoRA, and diffusion models. The goal is not to replace algebra, but to make its structure visible.For readers who want to understand why models learn, why reasoning stabilizes, and how intelligence moves through space, this book offers a map.", "item_img_path" : "https://covers1.booksamillion.com/covers/bam/9/79/818/107/9798181079496_b.jpg", "price_data" : { "retail_price" : "18.99", "online_price" : "18.99", "our_price" : "18.99", "club_price" : "18.99", "savings_pct" : "0", "savings_amt" : "0.00", "club_savings_pct" : "0", "club_savings_amt" : "0.00", "discount_pct" : "10", "store_price" : "" } }
The Terrain of Learning|Zixi Li

The Terrain of Learning : An Introduction to Reasoning Geometry and Intelligent Dynamics

local_shippingShip to Me
In Stock.
FREE Shipping for Club Members help

Overview

The Terrain of Learning is a concise introduction to reasoning geometry and intelligent dynamics.

Instead of treating deep learning as a list of formulas, this book asks a different question: what if learning is motion through a terrain? Loss becomes an energy landscape, gradients become directions, optimizers become ways of walking, and reasoning becomes a trajectory through hidden states and belief space.

Across twelve chapters, Zixi Li develops a geometric language for modern AI: parameter space, representation space, Bregman divergence, KL divergence, fixed points, ResNet, GPT, DEQ, chain-of-thought, LoRA, and diffusion models. The goal is not to replace algebra, but to make its structure visible.

For readers who want to understand why models learn, why reasoning stabilizes, and how intelligence moves through space, this book offers a map.

This item is Non-Returnable

Details

  • ISBN-13: 9798181079496
  • ISBN-10: 9798181079496
  • Publisher: Independently Published
  • Publish Date: June 2026
  • Dimensions: 11 x 8.5 x 0.46 inches
  • Shipping Weight: 1.15 pounds
  • Page Count: 220

Related Categories

You May Also Like...

    1

BAM Customer Reviews