Extracts patterns using weights, biases, and a non-linear activation function.
Article: Building a "No-Code" Neural Network in Modern Excel
For each hidden node, calculate the dot product of inputs and weights, add the bias, and apply an activation function. We will use the ( H1cap H sub 1 Linear Combination (
Once you have the gradient for each parameter, update it using : new_weight = old_weight – (learning_rate × gradient)
Building a neural network in Excel has evolved from complex VBA macros to using modern and LAMBDA functions . With these "new" features, you can now build a fully functional, deep neural network directly in the spreadsheet grid without a single line of code. build neural network with ms excel new
: Select all your weights and biases. Hold Ctrl to select multiple ranges: $F$2:$G$4, $I$2:$I$4 .
dW1=XT⋅δ1d cap W sub 1 equals cap X to the cap T-th power center dot delta sub 1 Excel Formula: =MMULT(TRANSPOSE(Data_Inputs), Delta_1) 6. Step 4: Updating Parameters with Gradient Descent Once the gradients (
If you are looking for the "new" way to use neural networks in Excel, Microsoft and third parties have recently introduced several AI integrations:
A standard neural network consists of three main components you’ll need to map out in your sheets: Your raw data (e.g., petal length, width). Extracts patterns using weights, biases, and a non-linear
In 2026, Andrej Karpathy’s microGPT inspired an Excel implementation that follows the exact same architecture—RMSNorm, ReLU, multi‑head attention, and MLP blocks—but implemented entirely in Excel formulas. The attention heads compute dot products between queries and keys, take weighted sums of values, and recombine them through trained projection matrices. This is not just a toy: it actually generates plausible‑sounding names. It demonstrates how far pure spreadsheet computation has come.
If your outputs never leave 0.5, your learning rate ( Alpha ) is too high or too low. The "new" Excel allows you to hook Alpha to a slider control (Developer Tab > Spin Button).
This public link is valid for 7 days and shares a thread, including any personal information you added. This link or copies made by others cannot be deleted. If you share with third parties, their policies apply. Can’t copy the link right now. Try again later.
This public link is valid for 7 days and shares a thread, including any personal information you added. This link or copies made by others cannot be deleted. If you share with third parties, their policies apply. Can’t copy the link right now. Try again later. With these "new" features, you can now build
In Python, this happens in an automated loop. In Excel, we can handle this update mechanism using two distinct approaches: Method A: Excel Solver (The Low-Code Way)
): =\delta^[2] * W_1^[2] * A_1^[1] * (1 - A_1^[1]) Step 5: Training the Network
Building a neural network with MS Excel using the functions ( RANDARRAY , MMULT , LAMBDA , spill ranges) democratizes deep learning.
