The Modern Mathematics of Deep Learning

Embedding into a feature network

Penguin learning project

The project uses 342 complete Palmer penguin records: 240 training-pool and 102 held-out rows. The selected training rows determine feature scaling and model parameters. Source CSV

This project trains weights with softmax, unlike the homepage lab, which uses nearest species averages. Same penguin measurements; different classification methods. Predictions can differ.

Quantum models · Describe wavefunctions, particle coordinates and energy relations.

PDE and quantum models are not operations on these penguin measurements.

Consider this concept alongside the project; its source assumptions may differ. These observations do not prove a theorem.

Input: 12 × 2; Held-out evaluation: 102 · #58 · 40.6 mm, 193 mm. Known answer: Adélie; Prediction: Adélie · [≈0.7383, ≈0.2444, ≈0.01737].

Used in these source sections: 8.2 Equations as guides

The source feeds the particle embedding and the other spin block’s positions into a function Γ to obtain a particle feature.

These expressions specify the inverse or quantum model.

Source formula example

\Psi_{k}\left(r,\theta_{k}\right)=\Psi_{k}\left(r,(\theta_{k,\mathrm{fc}},\theta^{+}_{\mathrm{emb}})\right)=\Gamma_{k}\left(\left(\mathrm{Emb}_{k}(r,\theta^{+}_{\mathrm{emb}}),(r_{n_{+}+1},\dots,r_{n})\right),\theta_{k,\mathrm{fc}}\right),

illustration · The invented finite inputs illustrate this expression. They do not prove a general identity, bound, or existence claim.

Symbols

$Ψₖ$: Particle-indexed feature function.
$Γₖ$: Function applied to an embedding and other-spin positions.
$Embₖ$: Embedding for particle k.
$θₖ,fc, θemb⁺$: Feature-network and shared embedding parameters.
$r$: Full particle configuration.

Source formula example

This source example uses its own inputs and assumptions. Editing it does not change the project.

Edit a value or the code, then run the example. This is a finite demonstration, not a proof of the paper’s theorem.

Embedding

This example’s data flow

You choose

input.embedding = 2

Initial values in code

otherPosition = 3

Calculated here

Result expression: { embedding, otherPosition, feature }

embedding = input.embedding Uses: input.embedding
feature = embedding + 2 * otherPosition Uses: embedding, otherPosition

Returns

Example result

embedding: 2
otherPosition: 3
feature: 8

JavaScript

Read numbers from input and return the result. JavaScript basics.

Result

Enable JavaScript to run this example.

Technical statement

The invented Γ is a linear function of two scalar inputs. The source specifies a learned function and permutation constraints; this example does not establish those constraints.

Source: Source expression