RAx News
Published in

RAx News

Making reading of research papers easy with RAx Key Insights

Making research paper reading easier with Key Insight

With the increase in the volume of valuable information available to us, quality reading time is shrinking fast.

Most researchers rely upon skim reading to quickly get the idea about the paper and decide whether to invest more time in it or not. It is quite reliable, but only when you already have a certain expertise in that area. You must ensure that you are not missing any critical point while skimming the paper. BUT you can only realize you are missing out on critical information if you know there is some in the first place. Additionally when you are new to the field, even skim-reading becomes time-consuming. Here’s how I use RAx to tackle this issue!

Where to focus on when reading a new paper?

I can skim through my ever-increasing reading list if I have a guideline to get to the most relevant aspect of the paper. If I can find and focus on that aspect quickly, I can decide whether to invest more time right now or not. For this, I upload all the relevant papers on my RAx project and scan them using Key Insights.

When I upload a paper on RAx, the RAxBot (in-house AI engine) extracts key information from the paper and collate it as key insights.

Fun-fact: RAx research team is using RAx for doing the research to build Key Insight feature.

Below is the example of the paper published in 2020 that I need to review quickly to know if it’s really important or not. Open the paper and click on the Key Insight tab.

Compilation of important aspects of paper inside Key Insight
A compilation of key points of the paper as Key Insights

Notice that Outline on the left-hand side? It’s not simply a list of sections in the paper. This is where Key Insight differs from other summary generating tools.

Bucketing of research paper in Key Insight
Bucketing of a research paper in Key Insight

RAxBot is designed to collate information from different sections of the paper into logical groups such as Research Goal, Approach, Originality, Methodology, Speculations, etc. These groups act as a guideline to dissect and approach a new paper on my list.

So Key Insight is more about having all the critical information pieces into structured buckets rather than simply compressing information in bullet points. That’s why we have not yet kept the word limit on Key Insight. Reading a few more words vs missing an important point? We decided to choose the latter.

How Key Insight help me to read papers quickly?

I and my research team were in the late stage of the development of the key insight feature. Thus, while tackling all engineering work we need to keep up with new papers coming up about summarisation. So instead of diving into the paper directly, I opened the Key Insight of the new paper to see if their approach was unique or original.

Going through Approach to see

I skim through 4 points in the approach and I get the gist of their originality. This paper is worth investing more time in!

Yet, before going into the paper, I quickly wanted to check if there is anything new in their evaluation setup. So I click on Evaluation setup & goal from the outline.

Narrow down to relevant sections of paper summarised in Key Insights
Quickly narrow down to the Evaluation setup

Turns out the S2ORC pipeline was quite important to check out. Similarly, I quickly checked the dataset they are using.

Control the details you like to see

There are two views of each section under Key Insight — short and detailed. Although there isn’t a strict word limit, I do have control over the depth of details I would like to see in a generated Key Insight. You can toggle details by clicking on the show-hide details button on the right of every section.

Quick and Detailed view of summary of paper
Quick and Detailed view of each section

So if I have to quickly skim through multiple papers in my list, I click on the Key Insight button on the list item and I get a shorter version of each section (image on left) with the option to see details. Whenever I open Key Insight in the reading room, all the sections are expanded in a detailed view (image on right). The assumption here is I am in a focused mode while exploring key insight in the reading room.

Compilation of important aspects of paper inside Key Insight
Left: Quickly scanning papers from the list Right: Getting into more details inside the reading room


This is just one of the examples of paper that I was able to do a quick read without losing into critical details. Also, with the logical bucketing, I get an opportunity to focus on the aspect that matters to me at a given stage of my research and my current need.

AI-generated compilation of information is neither going to make us smarter nor dumber. But we at RAx believe that modern problems require modern solutions. Key Insight is our attempt at solving the issue of information overload and less time. Do try it out and let us know how can we improve it!!




Official Newsletter of RAx. Stay updated with how your personal AI-powered literature analysis tool is getting smarter.

Recommended from Medium

A visual exploration of Germany’s media landscape

Hoteliers shouldn’t sleep on this!

Finding Value in Unlabeled Data

A person holding a magnifying glass up to a laptop, examining the keys.

The Data Science Research-Production Chasm

Modeling the NFL Field Goal

From image to information — 10 questions for QMENTA

4 — The Z-test: An Illustrative Hypothesis Test


Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Sourish Dasgupta

Sourish Dasgupta

Founder, RAx & Faculty, DA-IICT (Ph.D. Computer Science, University of Missouri - Kansas City)

More from Medium

Challenge for the C-suite as machine learning (AI) becomes the new “normal”​

What are we talking about when we say AI interpretability?

My CHI 2022 Reading List

Process diagram: Card-sorting workshop, formative evaluation, mindset quantification, summative evaluation, and design guidelines.

Life After CDS: A Conversation with CDS Faculty Fellow Angela Radulescu