: Adjust the weights based on the error. You can do this manually by adding "Weight Delta" columns or automate it using the Excel Solver Add-in . New AI Features in Excel (2025-2026)
┌─────────────────────────────────────────────┐ │ Neural Network Builder [X] [?] │ ├─────────────────────────────────────────────┤ │ Layers: │ │ [Layer 1: Input ] size: 5 │ │ [Layer 2: Hidden] size: 12 Act: ReLU [X]│ │ [Layer 3: Hidden] size: 6 Act: ReLU [X]│ │ [Layer 4: Output] size: 1 Act: Sigmoid │ │ [+ Add Layer] │ ├─────────────────────────────────────────────┤ │ Training: │ │ Learning rate: [0.01 ▼] Epochs: [2000] │ │ Batch size: [32 ▼] Optimizer: [Adam ▼]│ │ Loss function: [Binary Cross-Entropy ▼] │ │ [ Initialize ] [ Train ] [ Predict ] │ ├─────────────────────────────────────────────┤ │ Current Loss: 0.237 │ Loss chart (live) │ │ Best Loss: 0.191 │ \_/‾‾‾‾‾\_ │ └─────────────────────────────────────────────┘
Sigmoid(z)=11+e−zSigmoid open paren z close paren equals the fraction with numerator 1 and denominator 1 plus e raised to the negative z power end-fraction build neural network with ms excel new
Next, apply the Sigmoid function in an adjacent cell to get the actual activation ( AH1cap A sub cap H 1 end-sub ): =1 / (1 + EXP(-Z_H1)) Repeat this process for H2cap H sub 2 3. Calculating the Output Layer Now, use the hidden layer activations ( ) as inputs for the final output node ( O1cap O sub 1 ): Z_O1 = (A_H1 * Wo1) + (A_H2 * Wo2) + B2
For organizations, data scientists can deploy deep neural network classifiers as custom functions. Microsoft Azure =AZUREML() function to access a catalog of pre-built AI models. : Adjust the weights based on the error
Sub TrainNeuralNetwork() Dim i As Long Dim learningRate As Double Dim epochs As Long learningRate = 0.1 epochs = 500 ' Number of training loops Application.ScreenUpdating = False For i = 1 To epochs ' Example: Updating a single weight cell using its calculated gradient cell ' Repeat this pattern for all weight and bias cells Range("Weight_W11").Value = Range("Weight_W11").Value - (learningRate * Range("Grad_W11").Value) Range("Weight_W12").Value = Range("Weight_W12").Value - (learningRate * Range("Grad_W12").Value) ' ... include updates for all variables ' Force sheet recalculation to refresh gradients for the next loop Application.Calculate Next i Application.ScreenUpdating = True MsgBox "Training Complete! Optimization finished.", vbInformation End Sub Use code with caution.
Click , go to the GRG Nonlinear tab, and ensure Use Multistart is checked. This helps the network escape local minima to find global weights. Click Solve . Calculating the Output Layer Now, use the hidden
Happy spreading the signal!
Error=0.5×(Target−Prediction)2Error equals 0.5 cross open paren Target minus Prediction close paren squared
This is an excellent for a hypothetical version of Microsoft Excel (or an add-in like “Excel Labs” or “Analyze Data”).