ReLU can pass a shifted input
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.
Prediction rules and layers · Describe or apply prediction rules, layers and parameters.
The project uses a linear rule with two features. The fitted weights produce scores or normalized mass.
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: 6.2 Shortcuts between layers
If x + b is positive, subtracting b after ReLU returns x.
This local identity helps explain the role of skip connections.
Source formula example
exact · The equality under this condition does not establish a general residual-network theorem.
Symbols
- Real input; componentwise in a vector.
- Real shift.
- ReLU function.
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.
This example’s data flow
Code changed. Reset to restore this example’s data flow.
You choose
input.x= 1
Initial values in code
b=3
Calculated here
Result expression: { x, shifted: Math.max(0, x + b) - b, condition: x > -b }
x=input.xUses:input.x
Returns
x- 1
shifted- 1
condition- true
Read numbers from input and return the result. JavaScript basics.
Result
Technical statement
For every component the condition is xᵢ > −bᵢ. The toy chooses b = 3 and restricts x to ensure it.
Source: Source expression