Akhil Singh RAWAT
@alpharoot04
Followers
49
Following
511
Media
31
Statuses
148
Joined October 2022
Day 0 .Starting my #100DaysOfAIML challenge! Over the next 100 days, I'll be diving deep into AI, ML, and everything in between—sharing my learnings, projects, and insights. #AI #MachineLearning #100DaysOfCode.
3
0
10
Day 32 . ▪️Entropy . ▪️Information Gain . ▪️Decision Tree Learning
1
0
7
Day 31. ▪️ Decision Tree introduction . ▪️ Decision Tree working
1
0
9
Day 30. ▪️Evaluation of Linear algorithm . ▪️Bias and Variance . ▪️Evaluating a Learning algorithm (Neural Network ). ▪️ Regularization.
colab.research.google.com
Colab notebook
0
0
7
Day 28. ▪️ Iterative loop of ML development . ▪️ Error analysis. ▪️ Adding data
1
1
11
A large neural network will usually do well or better than a smaller one so long as regularisation is chosen appropriately.
0
0
2
RT @alpharoot04: Day 27. ▪️Bias and Variance . ▪️ Debugging a learning algorithm . ▪️Neural Network regularis….
0
1
0
Day 27. ▪️Bias and Variance . ▪️ Debugging a learning algorithm . ▪️Neural Network regularisation
1
1
9
RT @Avinashabroy: Common Machine Learning Algorithms!. 1️⃣ Linear Regression .->Used for predicting continuous values. ->Models the relatio….
0
95
0
RT @alpharoot04: Day 26 . ▪️Diagnosing bias and variance . ▪️ Regularisation and bias/variance. ▪️ Establishing….
0
1
0
Day 26 . ▪️Diagnosing bias and variance . ▪️ Regularisation and bias/variance. ▪️ Establishing a baseline level of performance
1
1
11
RT @alpharoot04: Day 25. ▪️Choosing the model. ▪️Evaluating the model. ▪️Cross Validation Method https://t.co….
0
1
0
If you choose the model degree .d based on which has the lowest test set error, that error will likely be too optimistic as an estimate of the model's ability to generalize to unseen data. So the solution is
0
0
2
Day 25. ▪️Choosing the model. ▪️Evaluating the model. ▪️Cross Validation Method
2
1
9
RT @alpharoot04: Day 24. ▪️ Derivative . ▪️ Forward and Back Propogation . ▪️ Computation GraphA .
0
1
0
Back propagation is an efficient way to compute derivatives as N nodes and P parameters compute derivatives in roughly N+P steps rather than N*P steps. Which is much more efficient in larger calculations.
0
0
3
Day 24. ▪️ Derivative . ▪️ Forward and Back Propogation . ▪️ Computation GraphA
3
1
11
RT @alpharoot04: Day 23. ▪️Adam Algorithm. ▪️Convolutional Layer. ▪️ Implementation of Softmax and Neural network….
0
1
0