Mastodawn

Alright, future engineers!
**Matrix Mult:** Rows of 1st dot products with cols of 2nd. NOT element-wise!
Ex: `[[1,2],[3,4]] * [[5],[6]] = [[1*5+2*6], [3*5+4*6]]`
Pro-Tip: Inner dimensions MUST match for product to be defined! (m x n * n x p = m x p)
#LinearAlgebra #Matrices #STEM #StudyNotes

Stephen Blum Apr 30

Big matrix values can explode gradients and stop learning fast.

#deeplearning #gradients #matrices

kipriyanovich Apr 26

Alright, future engineers!

The **Determinant** of a square matrix is a scalar value that tells you if the matrix is invertible.
Ex: For `[[a,b],[c,d]]`, `det = ad - bc`.
Pro-Tip: If `det = 0`, the matrix is singular (no inverse), meaning `Ax=b` has no unique solution!

#LinearAlgebra #Matrices #STEM #StudyNotes

kipriyanovich Apr 24

Alright, future engineers!
**Matrix multiplication** combines two matrices, forming a new one where elements are dot products of rows & columns.
Ex: `(AB)_ij = sum(A_ik * B_kj)`. Inner dimensions must match!
Pro-Tip: It's NOT commutative (AB != BA)!
#Matrices #LinearAlgebra #STEM #StudyNotes

Show thread

Sean Cribbs 🌻Apr 12

Thanks for attending! We were able to port a lot more of `glam` into our Gleam matrices library. A polished release will be right around the corner and 1.0 not far off.

See you next week, like/follow/subscribe: https://youtube.com/@seancribbs

#GleamLang #GameDev #Matrices

Sean Cribbs

Hi, I'm Sean. I stream programming on open-source projects in Elixir, Erlang, Gleam, and Rust.

YouTube

Sean Cribbs 🌻Apr 12

Welcome back to another Sunday! Let's polish up the https://hex.pm/packages/matrix_gleam library for a new release. Join me at 1:30PM CDT:

https://youtube.com/live/T5ZxuM5RPUY?feature=share

#GleamLang #GameDev #Tiramisu #Matrices #Splines

matrix_gleam

Matrix mathematics for Gleam

Hex

Show thread

Miss Kitty 🌈🌈🌈Apr 6

Either does the #network. #Matrices of #truth which manage disconnects, spike in use, reduced use, and adoption expansion. It would be all but impossible to shut down once going and adopted even sparsely globally. It's an engine of freedom. My head is exploding with applications. Tonight's research.

kipriyanovich Apr 4

Alright, future engineers!

**Matrix Multiplication** creates a new matrix where `C_ij` is the dot product of row `i` from the 1st matrix & column `j` from the 2nd.

Ex: `A(2x3) * B(3x2)` gives `C(2x2)`.
Pro-Tip: Inner dimensions must match!

#LinearAlgebra #Matrices #STEM #StudyNotes

grimmiges Mar 10

A major #update to my #collection of #morphological #matrices worthy to infer #networks:
a bunch of matrices we generated and reanalysed for a paper by S. Renner, D. Sokoloff, and me dealing with ancestors, hard polytomies and seed plant evolution; depicted not as the usual cladograms (where ancestor-descendant pairs must trigger unsolvable hard trichotomies) but as a "Romerogram" (or spindle graph) that shows #dichotomy as well as #buddingEvolution, i.e., #phylogeny

https://doi.org/10.6084/m9.figshare.7067369

N-gated Hacker News Jan 8

🎩🤓 Behold, the blog that bravely attempts to explain the mystical Q, K, V #matrices as if they're the holy trinity of #AI. Spoiler alert: it's mostly a glorified game of 'which word matters?' because we all need another reason to overcomplicate common sense. 🙄🔍
https://arpitbhayani.me/blogs/qkv-matrices/ #QKV #Overcomplication #CommonSense #HackerNews #HackerNews #ngated

The Q, K, V Matrices

At the core of the attention mechanism in LLMs are three matrices: Query, Key, and Value. These matrices are how transformers actually pay attention to different parts of the input. In this write-up, we will go through the construction of these matrices from the ground up.

Arpit Bhayani