Skip to content
This repository has been archived by the owner on Jan 4, 2025. It is now read-only.

Latest commit

 

History

History
61 lines (43 loc) · 2.67 KB

HOWTO.md

File metadata and controls

61 lines (43 loc) · 2.67 KB

Instructions

Things to do within the first 1-2 hours

There is nothing much you can do in the beginning, really.

  1. Open the navigation menu on the top right-hand side.

    • Use the arrows to navigate from one page to another.
    • Open 'Settings'.
    • Select and unselect the feeds you would like to subscribe to. I know that it can be quite overwhelming in the beginning. So do not put too much thought into it. I.e., do not worry about the 'evil' feeds but adjust for the 'spammy' ones over time.
    • Click 'Display Feeds', which brings you back to the home page.

  2. How about a casual stroll along this torrential timeline of news? On the top, it gives you the number of articles for this day and for the feeds you selected. Make sure to click a few likes and dislikes so that the learning experience may begin.

Things to do after the first two hours

News as well as the machine learning algorithms are updated once an hour. So your much appreciated patience should have given my algorithms ample time to get a first idea of your tastes. It is now time to bring some order into this torrential timeline of news!

  1. Open the navigation menu on the top right-hand side.

  2. Choose a language, or 'International' if you would like to keep all news in the same timeline.

  3. Choose a feed.

    • Full: Chronological feed.
    • Magic: Sort by assigning a score between 1 for most liked and -1 for most disliked.
    • Surprise: Randomly pick 10 articles centered around a score of 0. This means that you have neither expressed significant like nor dislike for similar articles in the past. You could say that they form a sample of your 'blind spot'. By clicking like or dislike here, you are really helping out my algorithms a lot!

  4. Choose an algorithm. When I say 'my' algorithms, by that I really mean open-source algorithms implemented by others that I have complacently taken from their libraries. If you have no background in machine learning -- just skip this step!

  5. Click 'Display'.

Other things you can do:

  1. In the navigation menu, generate a report of your favorite words and feeds.

  2. Click the lightbulb icon to gain insight into how my algorithms come to their conclusions.

Summary

In the beginning, the performance will be, let us say, suboptimal. But please keep in mind that my algorithms essentially start from scratch for every single user. This applies to the knowledge of individual tastes. But there is also minimal input about language itself from my end. Everything must be learnt, and I believe this is what makes this machine learning experiment so exciting.