Insert New
Slide
​
Roboto
A
Slide Break Defaults…
​
Cell Actions
Window Options
Start Presentation
▼
100%
Slide
1
of
9
​
​

Scholarly Insight through the ArXivExplore Paclet

Daniele Gregori, Soochow University
Institute for Advanced Study
​​
#WolframTechConf
Slide
2
of
9
​
​

ABSTRACT

In[]:=
abstract="ArXivExplore helps the deep data analysis of all 2.6M physics, math, computer science, etc. articles on ArXiv, providing functionality for e.g. title/abstract word statistics, TeX source/formulae/citations dissection, NNs for classification or recommendation and LLM-automated concept explanations and author reports.";
Out[]=
​

Paclet loading


Slide
3
of
9
​
​

Introduction

On ArXiv


A “deep data” problem


Academic “chat” vs scientific “insight”


Some ArXiv questions


Slide
4
of
9
​
​

ArXiv data mining

ArXiv main data


Categories


Full
T
E
X
scraping


Citations API


Slide
5
of
9
​
​

Real insights from “just counting”

Submission trends


Counting total words


Word popularity trends


Word logic combinations


Slide
6
of
9
​
​

Some Machine Learning

Basic classification


Advanced training


Feature extraction


Clustering


Slide
7
of
9
​
​

LLMs to enhance understanding

The importance of introductions


Explaining specific concepts


Creating author reports


Slide
8
of
9
​
​

Conclusions

Unlimited explorations


Planned improvements


New horizons in “deep data”


Slide
9
of
9
​
​

Thank you!

Out[]=