#mamba #ai #llm
Here’s a super simplified explanation of the Mamba language model with the Selective State Space Model (Selective SSM architecture). In the previous videos, I used the example of sequences of words to show how transformers use the Attention Mechanism to process natural language and predict the next word in a sequence of words, e.g., a sentence. In this video, I show you how Mamba’s AI architecture uses the Selective State Space Model to figure out which parts of the data. e.g., which words in a word sequence, are connected and how they might affect what happens next, e.g., to predict which word comes next.

Don’t forget to subscribe and watch these related videos:

Transformer Language Models Simplified in JUST 3 MINUTES!
   • Transformer Language Models Simplifie...  

This Is How Wxactly Language Models Work in AI – NO background needed!
   • This is how EXACTLY Language Models w...  

Backpropagation Simplified in JUST 2 MINUTES! --Neural Networks
   • The Concept of Backpropagation Simpli...  
https://www.youtube.com/@analyticsCam...

Key terms and concepts in the video:

00:00 Intro
00:31 Why Mamba?
00:52 State Space Models
01:14 Selectivity
01:25 Two stages of Selective SSM
01:48 Parameters
02:01 First stage: Projecting the Input
02:08 Discretization
02:25 Linear Time Invariance (LTI)
02:50 Dynamic data
03:14 B Parameter
03:19 C Parameter
03:39 Selection Mechanism
03:49 Hidden State update
03:58 Delta Parameter resets itself
04:30 Input Selection
04:41 Collocation
05:04 Each state update
05:09 Predicting the next word
05:23 Hardware-aware algorithm for Selective SSM
05:27 GPU with High Bandwidth Memory
05:34 Mamba’s overall architecture (H3 + Multi-layer Perceptron)


Stick around for more videos on LLM, Natural Language Processing (NLP), Generative AI, fun coding and machine learning projects, and follow Analytics Camp on Twitter (X):   / analyticscamp