Academic Papers

Markdown Viewer is ideal for writing academic papers with mathematical formulas, citations, and structured content. Export to Word for submission or further editing.

Why Markdown for Academia?

Advantage	Description
Focus on Content	No formatting distractions
Version Control	Track all revisions in Git
Math Support	Native LaTeX formulas
Collaboration	Easy diff/merge
Export Options	Word, PDF (via Word)

Paper Structure

Standard Academic Paper

# Paper Title: A Study of Something Important

**Authors:** John Smith¹, Jane Doe²

¹ Department of Computer Science, University A
² Department of Mathematics, University B

## Abstract

This paper presents a novel approach to... 
We demonstrate that... Our results show...

**Keywords:** machine learning, optimization, neural networks

---

## 1. Introduction

The problem of X has been widely studied [1, 2]. 
Previous approaches have limitations...

## 2. Related Work

Smith et al. [3] proposed... However, their method...

## 3. Methodology

### 3.1 Problem Formulation

Given a dataset $D = \{(x_i, y_i)\}_{i=1}^n$...

### 3.2 Proposed Approach

We propose a new algorithm that...

## 4. Experiments

### 4.1 Experimental Setup

### 4.2 Results

## 5. Conclusion

## References

[1] Author, "Title," Journal, 2023.
[2] Author, "Title," Conference, 2024.

Mathematical Formulas

Inline Math

The relationship is given by $E = mc^2$ where $m$ is mass.

Display Math

The optimization objective is:

$$
\min_{\theta} \frac{1}{n} \sum_{i=1}^{n} \mathcal{L}(f_\theta(x_i), y_i) + \lambda \|\theta\|_2^2
$$

Complex Equations

Systems of Equations

$$
\begin{cases}
\frac{\partial u}{\partial t} = D_u \nabla^2 u - uv^2 + F(1-u) \\
\frac{\partial v}{\partial t} = D_v \nabla^2 v + uv^2 - (F+k)v
\end{cases}
$$

Matrices

The transformation matrix is:

$$
\mathbf{A} = \begin{pmatrix}
\cos\theta & -\sin\theta & 0 \\
\sin\theta & \cos\theta & 0 \\
0 & 0 & 1
\end{pmatrix}
$$

Aligned Equations

$$
\begin{aligned}
\nabla \cdot \mathbf{E} &= \frac{\rho}{\epsilon_0} \\
\nabla \cdot \mathbf{B} &= 0 \\
\nabla \times \mathbf{E} &= -\frac{\partial \mathbf{B}}{\partial t} \\
\nabla \times \mathbf{B} &= \mu_0 \mathbf{J} + \mu_0 \epsilon_0 \frac{\partial \mathbf{E}}{\partial t}
\end{aligned}
$$

Figures and Diagrams

Experimental Setup

```mermaid
graph LR
    subgraph "Data Pipeline"
        Raw[Raw Data] --> Clean[Preprocessing]
        Clean --> Split[Train/Val/Test Split]
    end
    
    subgraph "Model"
        Split --> Train[Training]
        Train --> Val[Validation]
        Val --> |Tune| Train
    end
    
    subgraph "Evaluation"
        Val --> Test[Test Set]
        Test --> Metrics[Compute Metrics]
    end
```

Algorithm Flowchart

```mermaid
graph TD
    Start([Start]) --> Init[Initialize θ randomly]
    Init --> Loop{Converged?}
    Loop -->|No| Grad[Compute gradient ∇L]
    Grad --> Update[θ ← θ - α∇L]
    Update --> Loop
    Loop -->|Yes| End([Return θ*])
```

Network Architecture

```mermaid
graph LR
    Input[Input Layer\n784 neurons] --> H1[Hidden Layer 1\n256 neurons]
    H1 --> H2[Hidden Layer 2\n128 neurons]
    H2 --> H3[Hidden Layer 3\n64 neurons]
    H3 --> Output[Output Layer\n10 neurons]
    
    style Input fill:#e1f5fe
    style Output fill:#c8e6c9
```

Tables

Results Table

## Results

| Method | Accuracy | Precision | Recall | F1 Score |
|--------|----------|-----------|--------|----------|
| Baseline | 0.823 | 0.801 | 0.798 | 0.799 |
| Method A | 0.867 | 0.854 | 0.861 | 0.857 |
| **Ours** | **0.912** | **0.903** | **0.908** | **0.905** |

Table 1: Performance comparison on the test dataset.

Hyperparameter Table

| Parameter | Value | Description |
|-----------|-------|-------------|
| Learning rate | 0.001 | Adam optimizer |
| Batch size | 32 | Mini-batch SGD |
| Epochs | 100 | Early stopping at 10 |
| Dropout | 0.3 | Applied to hidden layers |
| Weight decay | 1e-5 | L2 regularization |

Table 2: Hyperparameter settings for all experiments.

Data Visualization

Performance Comparison Chart

```vega-lite
{
  "data": {
    "values": [
      {"method": "Baseline", "metric": "Accuracy", "value": 0.823},
      {"method": "Method A", "metric": "Accuracy", "value": 0.867},
      {"method": "Ours", "metric": "Accuracy", "value": 0.912},
      {"method": "Baseline", "metric": "F1 Score", "value": 0.799},
      {"method": "Method A", "metric": "F1 Score", "value": 0.857},
      {"method": "Ours", "metric": "F1 Score", "value": 0.905}
    ]
  },
  "mark": "bar",
  "encoding": {
    "x": {"field": "method", "type": "nominal"},
    "y": {"field": "value", "type": "quantitative"},
    "color": {"field": "metric", "type": "nominal"},
    "xOffset": {"field": "metric"}
  }
}
```

Training Curves

```vega-lite
{
  "data": {
    "values": [
      {"epoch": 0, "loss": 2.5, "type": "train"},
      {"epoch": 20, "loss": 1.2, "type": "train"},
      {"epoch": 40, "loss": 0.6, "type": "train"},
      {"epoch": 60, "loss": 0.3, "type": "train"},
      {"epoch": 0, "loss": 2.6, "type": "val"},
      {"epoch": 20, "loss": 1.4, "type": "val"},
      {"epoch": 40, "loss": 0.9, "type": "val"},
      {"epoch": 60, "loss": 0.7, "type": "val"}
    ]
  },
  "mark": {"type": "line", "point": true},
  "encoding": {
    "x": {"field": "epoch", "type": "quantitative", "title": "Epoch"},
    "y": {"field": "loss", "type": "quantitative", "title": "Loss"},
    "color": {"field": "type", "type": "nominal", "title": "Dataset"}
  }
}
```

Common Academic Formulas

Statistics

**Mean:** $\bar{x} = \frac{1}{n}\sum_{i=1}^{n} x_i$

**Variance:** $\sigma^2 = \frac{1}{n}\sum_{i=1}^{n}(x_i - \bar{x})^2$

**Standard Error:** $SE = \frac{s}{\sqrt{n}}$

**t-statistic:** $t = \frac{\bar{x} - \mu_0}{s/\sqrt{n}}$

Machine Learning

**Cross-Entropy Loss:**
$$H(p,q) = -\sum_{x} p(x) \log q(x)$$

**Softmax:**
$$\sigma(z_i) = \frac{e^{z_i}}{\sum_{j=1}^{K} e^{z_j}}$$

**Gradient Descent:**
$$\theta_{t+1} = \theta_t - \alpha \nabla_\theta L(\theta_t)$$

Physics

**Schrödinger Equation:**
$$i\hbar\frac{\partial}{\partial t}\Psi = \hat{H}\Psi$$

**Maxwell's Equations:**
$$\nabla \times \mathbf{E} = -\frac{\partial \mathbf{B}}{\partial t}$$

Recommended Theme

For academic papers, use the Academic theme:

Chinese body text: SimSun (宋体)
Headings: SimHei (黑体)
Line height: 1.75 (academic standard)
Proper spacing for Chinese documents

For English papers, Default or Palatino work well.

Export for Submission

Export to Word

Finalize your paper in Markdown
Press Ctrl/Cmd + S to export
Open in Word for final touches:
- Add page numbers
- Adjust margins if needed
- Final proofreading

Key Benefits

✅ Formulas export as editable equations
✅ Diagrams become high-resolution images
✅ Tables properly formatted
✅ Code blocks syntax highlighted

Tips for Academic Writing

Math Formatting

Use \text{} for words in formulas: $\text{accuracy} = \frac{TP + TN}{Total}$
Use \boldsymbol{} for vectors: $\boldsymbol{x}$
Use \mathbf{} for matrices: $\mathbf{A}$

References

While Markdown doesn’t have built-in citation support, you can:

Use numbered references [1], [2]
Keep a References section at the end
Export to Word, then use Zotero/Mendeley for final formatting

Version Control

# Commit regularly
git add paper.md
git commit -m "Add experimental results section"

# Branch for major revisions
git checkout -b revision-round-1