Lesson 01: Finding Papers¶

Before you can read a paper, you need to find it. This lesson covers the main tools physicists use to locate research literature, how ArXiv is structured, and how to set up alerts so new relevant work lands in your inbox automatically.

---

What is ArXiv?¶

ArXiv is a free, open-access preprint server run by Cornell University. A preprint is a version of a paper that the authors have made publicly available before (or alongside) formal peer review.

In physics, mathematics, and related fields, ArXiv is the central hub for research. Papers usually appear on ArXiv on the same day they are submitted to a journal, which means you can read new results weeks or months before they appear in a published journal.

---

How ArXiv is organised¶

ArXiv divides papers into subject categories. Each category has a short identifier. The ones most common in physics research are:

Identifier	Area
hep-ph	High-energy physics: phenomenology
hep-th	High-energy physics: theory
hep-ex	High-energy physics: experiment
astro-ph	Astrophysics (with sub-categories like astro-ph.HE, astro-ph.CO)
cond-mat	Condensed matter physics
gr-qc	General relativity and quantum cosmology
nucl-th / nucl-ex	Nuclear theory / experiment
physics	Applied, atomic, plasma, and other physics
math-ph	Mathematical physics

Ask your supervisor which categories are most relevant to your project. It is common to watch two or three related categories at once.

---

Where to start: review articles¶

Before diving into primary research papers, look for a review article on your topic. Review articles are written specifically to introduce a field, summarise its current state, and point to the key primary literature, exactly what you need when you are new to an area.

Good places to find review articles:

• Search ArXiv for your topic with the word "review" or "introduction" in the title: ti:review AND ti:neutrino AND cat:hep-ph
• Physics Reports: in-depth reviews across all physics.
• Annual Review of Nuclear and Particle Science: HEP and nuclear physics.
• Living Reviews in Relativity: continuously updated reviews in GR and cosmology.
• Ask your supervisor directly. They will know which review papers people in your subfield actually read.

---

Finding papers on ArXiv¶

Browsing new submissions¶

The ArXiv listing page for a category (e.g. arxiv.org/list/hep-ph/new) shows every paper submitted in the last working day. This is useful once you know roughly what you are looking for, but can be overwhelming at first.

Searching ArXiv¶

Use ArXiv search to find papers by keyword, author, or title.

Tips for effective searching:

• Use AND, OR, NOT to combine terms (e.g. dark matter AND direct detection).
• Search by author name to find all papers from a specific researcher.
• Restrict to a category using the "Subject area" dropdown.
• Use the "Abstract" field to search within abstracts rather than just titles.

Inspire-HEP¶

Inspire-HEP is a database specifically for high-energy physics. It indexes ArXiv papers and journal articles together, tracks citation counts, and makes it easy to find which papers cite a given paper (useful for tracing how a field has developed).

Inspire-HEP is also the standard source for BibTeX citation entries in HEP. Each paper has a texkey, a short identifier like ATLAS:2012yve, which is what you use in \cite{} commands in a LaTeX document. You can copy a ready-to-use BibTeX entry directly from any paper's Inspire page.

If your project is in particle physics or related areas, Inspire-HEP is worth bookmarking alongside ArXiv.

Following citations forward: once you find a key paper, use the Cited by tab on its Inspire page to see every paper that has cited it. This is one of the most effective ways to trace how a subfield has developed and to find the current state of the art. Start from a foundational paper and follow the thread forward to the most recent work.

NASA ADS¶

NASA ADS (Astrophysics Data System) indexes astrophysics, physics, and space science literature. Like Inspire-HEP, it tracks citations and links ArXiv preprints to their published journal versions. It covers a broader range of astrophysics subfields than Inspire-HEP, and its citation tools are widely used in that community.

If you are unsure which database your subfield uses, ask your supervisor.

Google Scholar¶

Google Scholar indexes papers across all disciplines, including ArXiv preprints and published journal articles. It is useful for finding papers outside physics, or when you want citation counts and links to the published version.

---

Following a specific paper's identifier¶

Every ArXiv paper has a unique identifier, for example 2312.01234. You can reach any paper directly at arxiv.org/abs/2312.01234.

When a colleague or supervisor mentions a paper, they will often give you just this number.

---

Setting up email alerts¶

Checking ArXiv daily by hand is impractical once your project is running. Email alerts let new relevant papers come to you.

ArXiv email alerts¶

1. Go to your category listing page (e.g. arxiv.org/list/hep-ph/new).
2. Click "Subscribe" at the top of the page.
3. Enter your email address. You will receive a daily digest of new submissions.

This gives you everything in the category, which is useful for staying broad, but noisy if the category is large.

Inspire-HEP author alerts¶

On Inspire-HEP, you can follow a specific author. When they upload a new paper, you receive a notification. This is a good way to track your supervisor's collaborators or key figures in your field.

Google Scholar alerts¶

1. Search for a keyword or author on Google Scholar.
2. Scroll to the bottom of the results and click "Create alert".
3. Google will email you when new matching papers appear.

This works across disciplines and catches papers that may not be on ArXiv.

---

What to do with a paper once you find it¶

Finding a paper is just the first step. Once you have identified something relevant, save it to your reference manager before you forget about it. You will thank yourself later. The next lesson, Lesson 02, covers how to actually read a paper efficiently. Lesson 03 covers how to manage and organise what you find.

---

Code examples¶

The examples/arxiv/ folder contains runnable Python scripts and Jupyter notebooks that demonstrate everything covered in this lesson. You do not need to run these to follow the track. They are for when you want to automate or explore programmatically. Each topic has a plain script for quick reference and a notebook with step-by-step commentary.

ArXiv API: arxiv_api.py · arxiv_api.ipynb¶

Uses the arxiv Python package to:

• Search ArXiv by keyword with boolean filters and category restrictions.
• Fetch a single paper directly by its ArXiv identifier.

import arxiv

client = arxiv.Client()
search = arxiv.Search(
    query="ti:neutrino mass AND cat:hep-ph",
    max_results=5,
    sort_by=arxiv.SortCriterion.SubmittedDate,
)
for result in client.results(search):
    print(result.title, result.entry_id)

Inspire-HEP API: inspire_api.py · inspire_api.ipynb¶

Uses the Inspire-HEP REST API (no authentication needed) to:

• Search by keyword, author, or any Inspire query syntax.
• Fetch a paper's full metadata, including its texkey and citation count, by ArXiv ID.
• Retrieve a ready-to-use BibTeX entry from Inspire, which is the standard citation format in HEP.
• Verify that a paper actually exists: useful when working with LLMs, which sometimes hallucinate plausible-sounding citations. Querying Inspire with the ArXiv ID or texkey is a fast sanity check: a clean hit with a matching title is strong evidence the paper is real; no hit means proceed with caution.

import requests

# Check whether a paper with a given ArXiv ID exists in Inspire-HEP
response = requests.get(
    "https://inspirehep.net/api/literature",
    params={"q": "eprint 1207.7214", "size": 1},
    timeout=10,
)
hits = response.json()["hits"]["hits"]
print("Found" if hits else "Not found")