Skip to content
←

Example Research Project

A sample page demonstrating math, figures, tables and code

Ziren Jin, Collaborator Name

Venue or Workshop 2026 · Sep 2026 · In progress

  • #machine-learning
  • #ai-for-science
Abstract line plot used as a placeholder cover image

This leading # Heading is optional — the page title comes from the frontmatter, so a first-level heading at the very top of the body is dropped automatically.

Motivation

Write normal Markdown. Bold, italic, inline code and links work, as do footnotes.1

Blockquotes are useful for highlighting a key claim or a quoted result.

Method

Inline math uses single dollar signs, like E=mc2E = mc^2 or L(θ)\mathcal{L}(\theta). Display math uses double dollar signs:

Gθ(x)=E(x)+PV+N rθ(z,T,P,c)G_\theta(x) = E(x) + PV + N\, r_\theta(z, T, P, c) L(θ)=1N∑i=1N∥yi−fθ(xi)∥22\mathcal{L}(\theta) = \frac{1}{N} \sum_{i=1}^{N} \left\lVert y_i - f_\theta(x_i) \right\rVert_2^2

An image on its own line becomes a figure. The optional quoted title becomes the caption:

Diagram of a three-stage pipeline: input, encoder, prediction head
Figure 1. Images are referenced relative to this Markdown file — this one is ./architecture.svg.

Results

Method In-distribution MAE OOD MAE
Baseline A 0.0520 0.1326
Baseline B 0.0310 0.0871
Ours 0.0021 0.0043

The numbers above are placeholders.

Code

import torch

def mae(pred: torch.Tensor, target: torch.Tensor) -> float:
    """Mean absolute error."""
    return (pred - target).abs().mean().item()
  1. Ordered lists
  2. work as expected
    • and can nest

  • GitHub-style task lists
  • render as checkboxes

Footnotes

  1. Footnotes are collected at the bottom of the page. ↩