AI/ML Modules for Engineering Students
Budget: $50 – $90 CAD
Please read the full project before bidding. I consider and read all serious bids and bidders. Don't waste my time with a fake bid to get me to message you (I will ignore your bid). Based on your bid including response, description text, budget and time as well as after reviewing your profile, I contact you. This is a seed project with the potential of related ongoing projects and rehire. This bid is only for 1 topic i.e. the total projects are worth >1200
I am developing a series of **AI/ML modules** tailored for engineering students (non-CS background). The focus is on **mathematical clarity, algorithmic reasoning, and illustrative MATLAB examples. I am not looking for production-level code or numerical optimization tricks to save memory, computation time...etc.
Each topic must follow the **same structured format**:
---
## 1) Engineering Mathematics (no skipped steps)
- **Setup**: define the data, feature map, and parameters.
- **Model**: write down the hypothesis (or classification rule).
- **Objective**: derive the loss function from first principles (e.g., MLE ⇒ cross-entropy, Gaussian noise ⇒ least squares) and give a **meaningful interpretation**.
- **Optimization**: provide closed-form solution if available, otherwise derive gradient-based update steps. Always explain the **meaning of optimization**.
- **Assumptions**: state clearly (noise model, independence, convexity, convergence notes).
- **Tuning & Hyperparameters**: explain their role in a meaningful way (not just “set α = 0.01”).
---
## 2) Numerical Method / Algorithm (no missing steps)
- **Flow (textual flowchart)**:
1. Input data and build matrices.
2. Initialize / pick solver.
3. Iterate: compute predictions, loss, gradient, update/solve.
4. Check stopping criteria.
5. Return final parameters and diagnostics.
- **Pseudocode**: language-agnostic, step-by-step, with **nothing skipped **.
---
## 3) MATLAB Illustrative Example
- **Synthetic Dataset**: generate small, meaningful data (2D, 3D, or toy examples).
- **Training**: apply either closed-form or iterative solver.
- **Visualization**: plots of fit line, decision boundaries, clusters, reconstruction error, etc.
- **Short Interpretation**: 1–3 bullets explaining results.
MATLAB 2025 has many AI tools — prefer built-in functionality **only for visualization and diagnostics**; the **core algorithm must be derived and coded manually** for educational clarity.
## Topics to Cover
1. Regression
2. Classification
3. kNN & Decision Tree
4. Clustering: K-means
5. Dimension Reduction
6. Artificial Neural Networks (ANN)
7. Autoencoder
8. Convolutional Neural Networks (CNN)
9. Explainable AI (XAI)
10. Recurrent Neural Networks (RNN)
11. Transfer Learning
---
## Deliverables
For **each topic**:
- A **step-by-step mathematical explanation** (clear, no skipped steps).
- A **numerical method/algorithm section** (flow + pseudocode).
- A **MATLAB example** (synthetic dataset, training, visualization, short interpretation).
Consistency across topics is critical.
This is bid is for only one topic which topic?
I am developing a series of **AI/ML modules** tailored for engineering students (non-CS background). The focus is on **mathematical clarity, algorithmic reasoning, and illustrative MATLAB examples. I am not looking for production-level code or numerical optimization tricks to save memory, computation time...etc.
Each topic must follow the **same structured format**:
---
## 1) Engineering Mathematics (no skipped steps)
- **Setup**: define the data, feature map, and parameters.
- **Model**: write down the hypothesis (or classification rule).
- **Objective**: derive the loss function from first principles (e.g., MLE ⇒ cross-entropy, Gaussian noise ⇒ least squares) and give a **meaningful interpretation**.
- **Optimization**: provide closed-form solution if available, otherwise derive gradient-based update steps. Always explain the **meaning of optimization**.
- **Assumptions**: state clearly (noise model, independence, convexity, convergence notes).
- **Tuning & Hyperparameters**: explain their role in a meaningful way (not just “set α = 0.01”).
---
## 2) Numerical Method / Algorithm (no missing steps)
- **Flow (textual flowchart)**:
1. Input data and build matrices.
2. Initialize / pick solver.
3. Iterate: compute predictions, loss, gradient, update/solve.
4. Check stopping criteria.
5. Return final parameters and diagnostics.
- **Pseudocode**: language-agnostic, step-by-step, with **nothing skipped **.
---
## 3) MATLAB Illustrative Example
- **Synthetic Dataset**: generate small, meaningful data (2D, 3D, or toy examples).
- **Training**: apply either closed-form or iterative solver.
- **Visualization**: plots of fit line, decision boundaries, clusters, reconstruction error, etc.
- **Short Interpretation**: 1–3 bullets explaining results.
MATLAB 2025 has many AI tools — prefer built-in functionality **only for visualization and diagnostics**; the **core algorithm must be derived and coded manually** for educational clarity.
## Topics to Cover
1. Regression
2. Classification
3. kNN & Decision Tree
4. Clustering: K-means
5. Dimension Reduction
6. Artificial Neural Networks (ANN)
7. Autoencoder
8. Convolutional Neural Networks (CNN)
9. Explainable AI (XAI)
10. Recurrent Neural Networks (RNN)
11. Transfer Learning
---
## Deliverables
For **each topic**:
- A **step-by-step mathematical explanation** (clear, no skipped steps).
- A **numerical method/algorithm section** (flow + pseudocode).
- A **MATLAB example** (synthetic dataset, training, visualization, short interpretation).
Consistency across topics is critical.
This is bid is for only one topic which topic?
Related categories:
Engineering
Matlab and Mathematica
Algorithm
Machine Learning (ML)
Mathematics
MATLAB