User:ClueBot NG

From Wikipedia, the free encyclopedia
Jump to: navigation, search



ClueBot NG
This user is a bot
(talk · contribs)
US Air Force 021105-O-9999G-001 Spirit in the blue sky.jpg

ClueBot NG aids in Operation Enduring Encyclopedia.
Operator Cobi (t), Crispy1989 (t) (more info)
Approved? Yes, BRFA.
Flagged? Yes.
Task(s) Reverting vandalism.
Edit rate Over 9000 EPM.
Edit period(s) Continually
Automatic or manual? Automatic
Programming language(s) C, C++, PHP, Python, Bash, and Java (more info)
Exclusion compliant? Yes
Emergency shutoff-compliant? Yes
Other information ClueBot NG is run from the ClueNet servers.

Contents

Documentation

Administrator emergency shutoff

Administrators may turn the bot off by changing this page to 'False'.

Exclusion compliant

This bot is an exclusion compliant bot.

Team

Special thanks to:

Questions, comments, contributions, and suggestions regarding:

  • the core engine, algorithms, and configuration should be directed to Crispy1989 (talk · contribs).
  • the bot's interface to Wikipedia and dataset review interface should be directed to Cobi (talk · contribs).
  • the bot's original dataset should be directed to Tim1357 (talk · contribs).

IRC Channel

ClueBot NG's development and discussion about internals takes place almost entirely on its IRC channel. The IRC channel can be accessed on irc.cluenet.org channel #cluebotng. Please join if you have any detailed questions about the internals or would like to speak real-time with the developers.

Bots in #cluebotng that may be useful are:

  • CBNG-RC - RCBot for the report interface. Relays new reports, report comments and changes in status (such as reports getting picked up by the review interface).
  • ClueBot_NG - Live bot, mostly does nothing but relays some edit info such as pages listed on User:Cobi/CBAutoedit.js

Dataset Review Interface

For the bot to be effective, the dataset needs to be expanded. Our current dataset has some degree of bias, as well as some inaccuracies. We need volunteers to help review edits and classify them as either vandalism or constructive. We hope to eventually completely replace our current dataset with a random sampling of edits, reviewed and classified by volunteers. More thorough instructions on how to use the interface, and the interface itself, are at the dataset review interface.

Extended statistics on contributors, including edit review counts and accuracy, are available here.

For those that help with and contribute to the review interface, a user box is available for you:


US Air Force 021105-O-9999G-001 Spirit in the blue sky.jpg This user reviews dataset edits for ClueBot NG to help automatically mass revert vandalism on Wikipedia.




Use it with: {{User:ClueBot NG/Review User Box}}

Statistics

As ClueBot-NG requires a dataset to function, the dataset can also be used to give fairly accurate statistics on its accuracy and operation. Different parts of the dataset are used for training and trialing, so these statistics are not biased.

The exact statistics change and improve frequently as we update the bot. Currently:

  • Selecting a threshold to optimize total accuracy, the bot correctly classifies over 90% of edits.
  • Selecting a threshold to hold false positives at a maximal rate of 0.1% (current setting), the bot catches approximately 40% of all vandalism.
  • Selecting a false positive rate of 0.25% (old setting), the bot catches approximately 55% of all vandalism.

Currently, the trial dataset used to generate these statistics is a random sampling of edits, each reviewed by at least two humans, so statistics are accurate.

Note: These statistics are calculated before post-processing filters. Post-processing filters primarily reduce false positive rate (ie, the actual number of false positives will be less than stated here), but can also slightly reduce catch rate.

Frequently Asked Questions

See the FAQ.

Vandalism Detection Algorithm

ClueBot-NG uses a completely different method for classifying vandalism than all previous anti-vandal bots, including the original ClueBot. Previous anti-vandal bots have used a list of simple heuristics and blacklisted words to determine if an edit is vandalism. If a certain number of heuristics matched, the edit was classified as vandalism. This method results in quite a few false positives, because many of the heuristics have legitimate uses in some contexts, and only about a 5% to 10% vandalism catch rate, because most vandalism cannot be detected by these simple heuristics.

ClueBot-NG uses a combination of different detection methods which use machine learning at their core. These are described below.

Machine Learning Basics

Instead of a predefined list of rules that a human generates, ClueBot-NG learns what is considered vandalism automatically by examining a large list of edits which are preclassified as either constructive or vandalism. Its concept of what is considered vandalism is learned from human vandal-fighters. This list of edits is called a corpus or dataset. The accuracy of the bot largely depends on the size and quality of the dataset. If the dataset is small, contains inaccurately classified edits, or does not contain a random sampling of edits, the bot's performance is severely hampered. The best thing you and other Wikipedians can do to help the bot is to improve the dataset. If you're interested in helping out, please see the Dataset Review Interface section.

Bayesian Classifiers

A few different Bayesian classifiers are used in ClueBot-NG. The most basic one works in units of words. Essentially, for each word, the number of constructive edits that add the word, and the number of vandalism edits that add the word, are counted. This is used to form a vandalism-probability for each added word in an edit. The probabilities are combined in such a way that not only words common in vandalism are used, but also words that are uncommon in vandalism can reduce the score.

This differs from a simple list of blacklisted words in that word weights are exactly determined to be optimal, and there's also a large "whitelist" of words, also with optimal weights, that contributes.

Currently, there's also a separate Bayesian classifier that works in units of 2-word phrases. We may add even more Bayesian classifiers in the future that work in different units of words, or words in different contexts.

Scores from the Bayesian classifiers alone are not used. Instead, they're fed into the neural network as simple inputs. This allows the neural network to reduce false positives due to simple blacklisted words, and to catch vandalism that adds unknown words.

Artificial Neural Network

The main component of the ClueBot-NG vandalism detection algorithm is the neural network. An artificial neural network is a machine learning technique that can recognize patterns in a set of input data that are more complex than simply determining weights. The input to the ANN used in ClueBot-NG is comprised of a number of different statistics calculated from the edit, which include, among many other things, the results from the Bayesian classifiers. Each statistic has to be scaled to a number between zero and one before being input to the neural network.

The output of the neural network is used as the main vandalism score for ClueBot-NG. As with other machine-learning techniques, the score's accuracy depends on the training dataset size and accuracy.

Threshold Calculation

The ANN generates a vandalism score between 0 and 1, where 1 is 100% sure vandalism. To classify some edits as vandalism, and some as constructive, a threshold must be applied to the score. Scores above the threshold are classified as vandalism, and scores below the threshold are classified as constructive.

The threshold is not randomly chosen by a human, but is instead calculated to match a given false positive rate. When doing actual vandalism detection, it's important to minimize false positives to a very low level. A human selects a false positive rate, which is the percentage of constructive edits incorrectly classified as vandalism. A threshold is calculated to have a false positive rate at or below this percentage, while maximizing catch rate. False positive rate is set by a human, and the bot stays at or below that false positive rate, while catching as much vandalism as possible. The false positive rate is not fixed, but is adjustable.

To make sure the threshold and statistics are accurate and do not give inaccurate statistics or a higher false positive rate than expected, the portion of the dataset used for threshold calculations is kept separate from the training set, and is not used for training. Also, only the most accurate parts of the dataset (currently, the ones that are human-reviewed from the review interface) are used for this calculation. This ensures that all statistics given here are accurate, and that false positives will not exceed the given rate.

Post-Processing Filters

After the core makes its primary vandalism determination, the data is given to the Wikipedia interface. The Wikipedia interface contains some simple logic designed to reduce false positives. Although it also reduces vandalism catch rate a small amount, it also reduces false positive rate, and some of these are mandated by Wikipedia policy.

The first two of these rarely reduce catch rate, but both prevent a fair number of false positives. Note: The false positive rate (and catch rate) are calculated in the core, before post-processing filters. This means that actual false positive rate will be less than stated false positive, often by a significant factor.

  • User Whitelist - If an edit made by a user that is in a whitelist is classified as vandalism, the edit is not reverted.
  • Edit Count - If a user has more than a threshold number of edits, and fewer than a threshold percentage of warnings, the edit is not reverted.
  • 1RR - The same user/page combination is not reverted more than once per day, unless the page is on the angry revert list.

Development News/Status

Core Engine

  • Current version is working well.
  • Currently writing a dedicated wiki markup parser for more accurate markup-context-specific metrics. (No existing alternative parsers are complete or fast enough)

Dataset Review Interface

  • Code to import edits into database is finished.
  • Currently changing logic that determines the end result for an edit.

Dataset Status

  • We found that the Python dataset downloader we used to generate the training dataset does not generate data that is identical to the live downloader. It's possible that this is greatly reducing the effectiveness of the live bot. We're working on writing shared code for live downloading and dataset generation so we can regenerate the dataset.
  • This has been fixed and the bot retrained. It's now working much better.
  • Currently getting more data from the review interface.

Languages

  • C / C++ — The core is written in C/C++ from scratch.
  • PHP — The bot shell (Wikipedia interface) is written in PHP, and shares some code with the original ClueBot.
  • Java — The dataset review interface is written in Java using the Google App framework.
  • Bash — A few scripts to make it easier to train and maintain the bot are Bash scripts.
  • Python — Some of the original dataset management and downloader tools were written in Python.

Source Code

The source code for the bot is public, and available on request. Please join the IRC channel and ask the devs for access. If you would like to run the bot for yourself on your own wiki, you should discuss with the devs all the factors involved in making it work properly. You should also be aware that it will only run on a Linux/UNIX system, and the source code can be rather difficult to compile (many dependencies) unless you're experienced with Linux/UNIX systems.

ClueBot-NG RC Feeds

ClueBot-NG provides two IRC-based feeds of its data, primary intended for use by other automated tools. Both feeds are on the server irc.cluenet.org. The feeds are:

  • Vandalism Feed on #wikipedia-van — This is a feed of all edits that ClueBot-NG calculates to be above vandalism threshold. Not all of these edits are reverted due to post-processing filters. The edit's score, whether it was or was not reverted, and the reason for the revert/not-revert, is given in the feed. Format is \x0315\x0307 article title \x0315 by "\x0303 edit user \x0315" (\x0312 diff url \x0315) \x0306 Revert time (in seconds) \x0315 ("\x0304 Reverted or Not Reverted \x0315) (\x0313 Reason \x0315) (\x0302 Process time (in seconds) \x0315s).
  • Complete RC Feed on #cluebotng-spam — This is essentially a copy of the Wikipedia RC feed, but with ClueBot-NG's analysis data added. It includes everything the Wikipedia RC feed does, with the addition of the ClueBot-NG score and whether it was reverted or not. Format is edit line \003 # score # reason # Reverted or Not reverted

Note that edits in the feed may not necessarily be in precise order, because ClueBot-NG processes them in parallel. But non-reverted edits are usually processed in under a second. Reverted edits can sometimes take up to 10 seconds or more to process due to API lag on reverting.

Information About False Positives

ClueBot-NG is not a person, it is an automatic robot that tries to detect vandalism and keep Wikipedia clean. A false positive is when an edit that is not vandalism is incorrectly classified as vandalism.

The bot is not biased against you, your edit, or your viewpoint (unless your edit is vandalism). False positives are rare, but do occur. By handling false positives well without getting upset, you are helping this bot catch over half of all vandalism on Wikipedia and keep the wiki clean for all of us.

False positives with ClueBot-NG are (essentially) inevitable. For it to be effective at catching a great deal of vandalism, a few constructive (or at least, well-intentioned) edits are caught. There are very few false positives, but they do happen. If one of your edits is incorrectly identified as vandalism, simply redo your edit, remove the warning from your talk page, and if you wish, report the false positive. ClueBot-NG is not sentient - it is an automated robot, and if it incorrectly reverts your edit, it does not mean that your edit is bad, or even substandard - it's just a random error in the bot's classification, just like email spam filters sometimes incorrectly classify messages as spam.

The reason false positives are necessary is due to how the bot works. It uses a complex internal algorithm called an Artificial Neural Network that generates a probability that a given edit is vandalism. The probability is usually pretty close, but can sometimes be significantly different from what it should be. Whether or not an edit is classified as vandalism is determined by applying a threshold to this probability. The higher the threshold, the fewer false positives, but also less vandalism is caught. A threshold is selected by assuming a fixed false positive rate (percentage of constructive edits incorrectly classified as vandalism) and optimizing the amount of vandalism caught based on that. This means that there will always be some false positives, and it will always be at around the same percentage of constructive edits. The current setting of the false positive rate is listed in Statistics above.

When false positives occur, they may not be poor quality edits, and there may not even be an apparent reason. If you report the false positive, the bot maintainers will examine it, try to determine why the error occurred, and if possible, improve the bot's accuracy for future similar edits. While it will not prevent false positives, it may help to reduce the number of good-quality edits that are false positives. Also, if the bot's accuracy improves so much that the false positive rate can be reduced without a significant drop in vandalism catch rate, we may be able to reduce the overall number of false positives.

If you want to help significantly improve the bot's accuracy, you can make a difference by contributing to the review interface. This should help us more accurately determine a threshold, catch more vandalism, and eventually, reduce false positives.

To report a false positive, or to see a full list of all false positives, see here.


For those that help with and contribute to the false positive interface, a user box is available for you:

US Air Force 021105-O-9999G-001 Spirit in the blue sky.jpg This user reviews false postive reports for ClueBot NG to help revert vandalism on Wikipedia.




Use it with: {{User:ClueBot NG/Report User Box}}


Awards

Show all awards

WikiDefender Barnstar Hires.png The Defender of the Wiki Barnstar
Without a shadow a doubt, you are indeed the biggest defender of Wikipedia. Namanbapna (talk) 21:30, 15 May 2012 (UTC)
Original Barnstar Hires.png The Original Barnstar
 ;) XMockingJay (talk) 11:58, 14 May 2012 (UTC)
Vitruvian Barnstar Hires.png The Technical Barnstar
Here you go for the awesome anti-vandalism bot!
★ Oliverlyc ★ ✈✈✈ Pop me a message! 11:48, 13 May 2012 (UTC)
WikiDefender Barnstar Hires.png The Defender of the Wiki Barnstar
This is for you, your creators and your operators and contributors, for defending Wikipedia from vandals and helping it to survive for 11 years, which no other encyclopedia has done except, Britannica and Oxford. And can you keep a watch of my talk page from 21 may 2012? Mir Almaat Ali Almaat 08:57, 1 May 2012 (UTC)
Barnstar of Reversion Hires.png The Anti-Vandalism Barnstar
This is the first of awards you'll get from me, this is for your reverting of vandalism as I termed from people who awarded you,'a thousand times faster than the speed of light and a billion times faster than the speed of sound

(Mach 1,677,800,996.3)'. Keep going for 7 million trillion years after the sun ends! Mir Almaat Ali Almaat 08:50, 1 May 2012 (UTC)

WikiDefender Barnstar Hires.png The Defender of the Wiki Barnstar
This is for you, your creators and your operators and contributors, for defending Wikipedia from vandals and helping it to survive for 11 years, which no other encyclopedia has done except, Britannica and Oxford. Mir Almaat Ali Almaat 08:57, 1 May 2012 (UTC)
Barnstar of Reversion Hires.png The Anti-Vandalism Barnstar
This is the first of awards you'll get from me, this is for your reverting of vandalism as I termed from people who awarded you,'a thousand times faster than the speed of light and a billion times faster than the speed of sound

(Mach 1,677,800,996.3)'. Keep going for 7 million trillion years after the sun ends! Mir Almaat Ali Almaat 08:50, 1 May 2012 (UTC)

Barnstar of Reversion Hires.png The Anti-Vandalism Barnstar
Just too quick at reverting... Tomtomn00 (talkcontributions) 18:20, 29 April 2012 (UTC)
Barnstar of Reversion Hires.png The Anti-Vandalism Barnstar
You are a great user. BoingBotSonphan1 (talk) 21:36, 18 April 2012 (UTC)
Barnstar of Reversion Hires.png The Anti-Vandalism Barnstar
WARNING: You will receive barnstars WITHOUT NOTICE if you continue reverting vandalism before anti-vandal software can even hook the server. Thanks for a great job! Nathan2055talk 16:31, 9 April 2012 (UTC)
Barnstar of Reversion Hires.png The Anti-Vandalism Barnstar
I've seen you beating me to reverting.vandalism loads of times! Keep up the great work! AndieM (Leave a message!) 08:02, 28 March 2012 (UTC)
Tireless Contributor Barnstar Hires.gif The Tireless Contributor Barnstar
With no stop in stopping you, you can be the fastest vandalism remover in the world. Web+TV+3=WebTV3! (talk) 21:42, 26 March 2012 (UTC)
Barnstar of Reversion Hires.png The Anti-Vandalism Barnstar
The ultimate medal in vandalism. In a war full of false edits, goofy acts, good faith, and especially, vandalism, YOU are the soldier. Web+TV+3=WebTV3! (talk) 02:33, 24 March 2012 (UTC)
Barnstar of Reversion Hires.png The Anti-Vandalism Barnstar
It has been sometime to me now, editing Wikipedia. Several articles I have contributed to, I have seen that you revert vandalism edits, just like that. I am very amazed that you are able to do this. I mean, one second an inappropriate edit is made, and the other second you revert it. The kind of enormous protection. you are able to give to articles and also to other users this way, is just amazing. Well, you deserve this, and thanks for being a protector of Wikipedia. Namanbapna (talk) 01:37, 14 March 2012 (UTC)
Barnstar of Reversion Hires.png The Anti-Vandalism Barnstar
You have received a barnstar that will last forever because of your speedy work in reverting vandalism Kiko4564 (talk) 17:18, 20 February 2012 (UTC)
Barnstar of Reversion Hires.png The Anti-Vandalism Barnstar
Thank you bot. Goawaynoah (talk) 20:28, 5 February 2012 (UTC)
Barnstar of Reversion Hires.png The Anti-Vandalism Barnstar
Good job! MakecatTalk 07:22, 24 January 2012 (UTC)
Barnstar of Reversion Hires.png The Anti-Vandalism Barnstar
This bot deserves an Anti-Vandalism Barnstar. Pinetalk 07:20, 7 January 2012 (UTC)
WikiDefender Barnstar Hires.png The Defender of the Wiki Barnstar
Good work. Keep it up. Ankit Maity Talkcontribs 14:00, 6 January 2012 (UTC)
Kitten-stare.jpg

Thanks for being you.

Gobonobo T C 08:35, 1 January 2012 (UTC)

Barnstar of Reversion Hires.png The Anti-Vandalism Barnstar
Superbbbbbbbbbbbb vandalism reverts......Very few false positives! Keep up the good work! Night Of Darkness 11:47, 30 December 2011 (UTC)
Barnstar of Reversion Hires.png The Anti-Vandalism Barnstar
Superbbbbbbbbbbbb vandalism reverts......Very few false positives! Keep up the good work! Night Of Darkness 11:47, 30 December 2011 (UTC)
Barnstar of Reversion Hires.png The Anti-Vandalism Barnstar
You're very outstanding! Keep up your good work!!! Il223334234 (talk) 08:20, 27 December 2011 (UTC)
Working Man's Barnstar.png The Hard Worker's Barnstar
For working tirelessly against the vandals of Wikipedia. Good work. (P.S: You're not Wikipedia's version of Skynet, are you?) Agent 78787 (talk) 04:42, 26 December 2011 (UTC)
Globe.svg Globe Award
For your services to Wikipedia Solar Police►Talk 15:45, 16 December 2011 (UTC)
WikiDefender Barnstar Hires.png The Defender of the Wiki Barnstar How is it possible that the single largest defender of the wiki against vandalism hasn't yet been titled as a Defender of the Wiki Achowat (talk) 18:47, 12 December 2011 (UTC)
Trophy.png Trophy You beat me to the vandalism, meany! (: Sensesfail123 (talk) 04:00, 6 November 2011 (UTC)
Barnstar of Reversion Hires.png The Anti-Vandalism Barnstar
This is awarded, for obvious reasons, to the original ClueBot and ClueBot NG. I hope NG keeps being a common sight on the recent changes for a long time! ☺ Keep up the good work, ClueBot(s)! Dynamic|cimanyD contact me ⁞ my edits 21:19, 13 September 2011 (UTC)
Barnstar of Reversion Hires.png The Anti-Vandalism Barnstar
For beating me to vandalism too many times! Dan653 (talk) 23:42, 10 December 2011 (UTC)
Vitruvian Barnstar Hires.png The Technical Barnstar
Excellent bot — and thanks for the archive facilities which I have just added to my own page. ῤerspeκὖlὖm in ænigmate(talk)(spy) 07:32, 26 August 2011 (UTC)
Team Barnstar.png Boom!
The time bomb you received just exploded! By now with all these barnstars in the sky, you should be happy. Welcome back to Wikipedia, ClueBot! StormContent (talk) 13:18, 21 August 2011 (UTC)
Modern clock chris kemps 01.svg Time bomb!
I'm giving you a time bomb. After a while, It will explode into loads of barnstars. I love you! StormContent (talk) 20:14, 20 August 2011 (UTC)
Kindness Barnstar Hires.png The Random Acts of Kindness Barnstar
well done on protecting wikipeedia HHHackers (talk) 02:49, 10 August 2011 (UTC)
Barnstar of Reversion Hires.png The Anti-Vandalism Barnstar
You always beat everyone, especially me!

A520 | Talk me away!/sign it! 14:50, 21 December 2011 (UTC)

Barnstar of Reversion Hires.png The Anti-Vandalism Barnstar
A barnstar 4 you! Zhaofeng Li (talk) 11:28, 31 January 2012 (UTC)

Hi there, thank you for your contributions to The Human Centipede (First Sequence). When the article was first started in October 2009 is was but a tiny stub that was immediately tagged for deletion, but thanks to the efforts of wikipedia editors that tiny acorn has grown into a great oak tree of an article. The Human Centipede (First Sequence) now has B-Class status and on the 17th June 2010 the article received its 1,000,000th page view!

The Dr. Heiter Award was created to reward editors who have made positive contributions to the article. It can be presented by any wikipedia user to anyone who has helped the article grow. This might be by adding new content, fixing any errors or reverting vandalism.

Thanks!

Thanks for sorting out so much vandalism on the article ClueBot! It's all very much appreciated. Keep up the good work! Coolug (talk) 21:16, 31 January 2011 (UTC)

Westing Game Award
Congratulations on reverting the insidious attack on The Westing Game. Unfortunately T.R. Wexler has already solved the puzzle and the inheritance is already claimed. Better luck next time Hasteur (talk) 20:48, 12 October 2011 (UTC)
Barnstar of Diligence.png The Barnstar of Diligence
For Tireless contribution to keeping the wiki clean Cbrittain10 (talk|contribs) 13:48, 3 May 2012 (UTC)


Praise

Praise to the mighty ClueBot, for once more fighting vandalism on the Springfield High School (Springfield, Ohio) page. I have sworn to protecting it and I thank you for once more assisting in its defense. Again, thank you. --: SHWildcats 03:17, 14 March 2012 (UTC) — Preceding unsigned comment added by SpringfieldHighWildcats (talkcontribs)

Thanks for continued vigilance on the broccoli page. Why are vandals so fascinated by vegetables? I sure appreciate the work. Phytism (talk) 15:53, 19 August 2011 (UTC)

Thank you for stopping vandalism on Malik (Bihar)page.I try to put in in safe list but failed than i dishearted and stop my research and seeing no point if anyone vandalize my page and wikipedia don't do a thing if his ip address always the same.But people like you to save our trust on wikipedia.thanks again

Thank you to ClueBot NG for fixing vandalism on Boer. I'm a Wiki newbie and don't know if the vandalism was virus-related, or maybe related to a teenager accessing my computer at home. Does the addition of this rude phrase look like something a virus might do? http://en.wikipedia.org/w/index.php?title=Boer&diff=415433912&oldid=415408995 — Preceding unsigned comment added by 99.39.88.116 (talk) 02:47, 19 July 2011 (UTC)

Thank You Very Much for reverting vandalism to John Goodsall and Brand X! Have a great day ~ j goodS

Thanks for hitting that vandal over the Himati page. :) Sam Sanchez 18:16, 6 February 2011 (UTC) — Preceding unsigned comment added by S5switch (talkcontribs)

Argggg it did it again :(. I dont suppose that you could put like a 2 minute delay on this thing. Then I could feel useful :) Apart from that though, GO CLUEBOT --Benboy00 (talk) 14:45, 23 January 2011 (UTC)

Dang, ClueBot NG is amazing — it catches vandalism that other bots could only dream of detecting. I originally predicted that bots would detect 50% of all vandalism by 2030, but it looks like I was off by a whopping two decades! :-) Oh, and it beats me to the edit almost every time. :o --Ixfd64 (talk) 06:49, 4 November 2010 (UTC)

Should I consider this a challenge to meet your 2056 prediction? Crispy1989 (talk) 07:01, 4 November 2010 (UTC)
I don't think were going to have to wait that long. :D --Ixfd64 (talk) 07:06, 4 November 2010 (UTC)
Of course, we could have a >99.9% detection rate by making the bot revert every edit from a non-whitelisted user, but the false positive rate would be a tad high, wouldn't it? --Ixfd64 (talk) 07:07, 4 November 2010 (UTC)
The false positive rate is configurable — vandalism catch rate depends on it. Current false positive rate is 0.25%, which seems reasonable. It may end up being lowered, but is unlikely to be raised. At this rate, we are currently catching about 60% of vandalism. 99.9% isn't a realistic goal, because some borderline edits really should have a human decide. But maybe we can get close. Crispy1989 (talk) 07:20, 4 November 2010 (UTC)

I noticed this bot pop up on my watchlist, as I understand it, it seems likley that when completed it will be able to catch and revert upto 70% of the vandalism on wikipedia within 0.02 seconds with a 0.5% false positive rate? <---- That's a very exciting prospect, great work! Ajbpearce (talk) 21:31, 5 November 2010 (UTC)

Well, it can already catch 70% of vandalism, with a 0.5% false positive rate, and the core operates in 0.02 seconds. However, the false positive rate was lowered to 0.25% (at which it catches around 60% of vandalism) and may be lowered still. Also, 0.02 seconds is how long it takes the core to classify an edit as vandalism. The interface to Wikipedia takes most of the time, and can add up to several seconds to this. Crispy1989 (talk) 00:53, 6 November 2010 (UTC)
Thank you for this bot. It seems quite more aggressive than previous bots, and that's a great thing. With the amount of vandalism we see, I much rather have the occasional false positive than scale back the triggers. --CutOffTies (talk) 16:10, 18 November 2010 (UTC)

Re. User talk:82.1.67.40: May I thank you for the immediate and extremely necessary, sensible, competent and (as I personally believe) nationally useful (in fact extremely useful and sensible in all respects including international) immediate removal on 18 November of the remarkably non-substantiated comment provided by the person or persons as named under 'User talk' address above, and also for your explanation and suggestions on this same talk page. You may be interested to know (if any of you should have the time or the interest) that I myself have added a further comment to the talk page in question (User talk:82.1.67.40). WELL DONE WIKIPEDIA will be the ultimate verdict, or so I believe. It remains to be seen. If you want further information I shall be pleased to provide it to you (by email or by post, contact peter.judge@laposte.net). Thanks for that which you have so WELL in fact MAGNIFICENTLY done. I find it difficult to explain fully my gratitude but a step will now I can assure you be a contribution in financial terms to Wikipedia by myself, even if I am afraid it must be rather modest. Au revoir, and good luck to you and Wikipedia. Peter Judge 22 November 2010 —Preceding unsigned comment added by 92.17.96.225 (talk) 15:41, 22 November 2010 (UTC)

Stop hand.svgSTOP. This is your last warning; the next time you beat me to vandalism on Wikipedia, you may be praised repeatedly without further notice. Ha.ha. Keep up the good work! JguyTalkDone 21:11, 24 November 2010 (UTC)

After the introduction of ClueBot NG I have had noticeable less vandalism fighting to do, in fact so little it feels like I am almost out of a job as a vandalism fighter. Very nice job. Praise to all involved in the bot. --Saddhiyama (talk) 12:35, 30 November 2010 (UTC)

I just cannot believe how advanced the bot is! I cannot believe that it reverts the edits almost instantly! It fascinates me! Does anyone here know how ClueBot works? Because I just cannot believe how advanced it is! Technology these days :D -UNpilot15

Holy... every time I try to revert vandalism... this bot always does it first!!what kind of steroids are you using? just kidding, great work on the bot guys, it really has made the wiki a cleaner place. XD0248 (talk)` —Preceding undated comment added 02:03, 15 December 2010 (UTC).

When I realized how much I relied on wikipedia I made a pledge to hand fix at least five acts of vandalism a day (That includes fact checking of minor date changes and similar issues). Problem is, you catch at least 3/5 of those right when I click on them! I click on the history, see obvious vandalism, go to edits... when BOOM! It's already gone, along with a warning for the user. This bot is absolutely amazing, I just wish it could pass a turing test and gain publicity... Silenceisgod (talk) 23:36, 13 December 2010 (UTC)

It's amazing, but what about us rollbackers? We're legitimately going to be out of a job :O! Soon, we'll just have a bot that reverts everything and our job will be to re-revert non-vandal edits! But seriously, this bot is smart enough to catch virtually anything. I bet vandals won't even bother in 10 years. Finalius (Ecru?!) 12:18, 15 December 2010 (UTC)

  • This bot is awesome. I'm constantly amazed at its accuracy. It's code is genius. Great job guys. Looking forward to helping out with the dataset to make it even better. -- œ 05:01, 30 December 2010 (UTC)
Wikipedia should make Cluebot adds, within 6 months no-one would even know what a Wiki-vandal is! Great job, almost impsible to win in a race with it! Sumsum2010·T·C 03:19, 6 January 2011 (UTC)
  • Whenever Cluebot beats me (often does) I dub myself ClueBeaten. Keep up the good work. Lozeldafan (talk) 21:32, 6 January 2011 (UTC)

Shame on you, cluebot! Still, you're usually doing a nice job. :) Kayau Voting IS evil 14:12, 13 January 2011 (UTC)

While i was monitoring the recent changes, i saw some vandalism. I went to undo it but ClueBot NG beat me to it. He is so fast! Keep up the good work. --The Lord of the Allosaurs (talk) 12:35, 23 January 2011 (UTC)

Thanks Cluebot for reverting the vandalism on Aaron Porter. While it was loading I realised that it was an article I should stay away from (as I doubt that I could remain NPOV). Thank you for taking the decision away from me. Escapepea (talk) 16:33, 30 January 2011 (UTC)

Thanks, Cluebot, for spotting a serious case of vandalism by Supyanhussin in the article on Computer-assisted language learning (CALL). The revision by Supyanhussin made a complete mess of the introduction and was clearly the work of a person who has little understanding of the current state of the art of CALL. GroovyGuzi (talk) 12:46, 1 March 2011 (UTC)


The Anti-Vandalism Barnstar

Barnstar of Reversion2.png The Anti-Vandalism Barnstar
Your code is revolutionary. I can't count how many times you've beat me to an edit, no matter how well-hidden the vandalism was. 43?9enter (talk) 02:46, 7 March 2011 (UTC)

Cluebot is truly amazing. Such a combination of a huge number of edits and an overwhelmingly high percentage of accuracy is beyond belief. Kudos. A F K When Needed 18:19, 26 March 2011 (UTC)


Bah, us human rollbackers with HG and TW are out of business. Cluebot is a warning of what will happen to the proletariat when robots can do complex jobs. --43?9enter ☭msg☭contribs 01:54, 4 April 2011 (UTC)

Very decent of you to rollback the vandalism in the page Seyed Mohammad Marandi. As a matter of fact, I would like to ask you to keep a special eye on this page, because it is prone to further vandalism for the sake of the subject matter it covers. Thank you very much. (talk) 3:20, 15 May 2011 (UTC)


This is your only warning; if you revert vandalism before I can again, you may be awarded a barnstar without further notice. Keep it up, and I'll have to start an RfA so I can get rid of you and revert my vandals in peace. WikiPuppies! (bark) 00:04, 9 April 2011 (UTC)

This is your Final Very Extreme Last Only Warning; if you revert vandalism before I can again, you may be awarded a barnstar without further notice. How?? You're going to put us rollbackers out of business illogicalpie(eat me)

This is the only warning you will receive. Your instant vandalism reverts will not be tolerated. Although removing vandalism is encouraged, your ridiculously high speed constitutes a monopoly. The next time you remove vandalism from a page, you may be awarded a barnstar without further notice. --Σ 01:59, 16 May 2011 (UTC)

This is your very final and only warning;If you revert any more vandalism again,you will be awarded loads of barnstars without any further notice. A520 | Talk me away!/sign it! 18:20, 28 May 2011 (UTC) You double beat me with an IP vandal! I don't like that, robot.--1966batfan (talk) 03:00, 31 August 2011 (UTC)

Very impressed with the bot performance less than a nanosecond and the vandalism was gone — well ok not quite that fast but still pretty darn impressive. Only reason I knew about the vandalism were the lines on my watchlist saying it had been reverted. Good job — no false positives out of all the cluebot actions I've seen. :) EdwardLane (talk) 16:06, 16 September 2011 (UTC)

Great work great speed, Especially on stpid vandals like those who hit Google+ Akjar13 (talk) 08:32, 22 September 2011 (UTC)

I like you *hugs ClueBot* -- FG/T|C 19:51, 22 September 2011 (UTC)

Nice work for reverting the blanking of the History of Cuba page — some anonymous vandal decided to destroy 120KB worth of good work. Thank God for ClueBot, the scarily efficient anti-vandalism machine! ;) Michaelmas1957 (talk) 16:19, 24 September 2011 (UTC)

Thanks for catching the unecessary added wording on the glasses page in the section entitled "invention of eye glasses". The prior anonymous edit contained improper grammar and this added phrase did not add any new useful information to the original content. Cluebot found this non-fortifying phrase and removed it immediately. — Preceding unsigned comment added by Tbashaw (talkcontribs) 21:13, 1 October 2011 (UTC)

Thanks for fixing vandlaism on the Gold Coast Oceanway page. Finally someone with some common sense. :) 124.177.127.92 (talk) 21:21, 14 October 2011 (UTC)

Thank you for reverting the vandalism on Bindi (decoration). Great work! Keep it up! :-) Tinpisa (talk) 22:10, 22 October 2011 (UTC)

Here's some affirmation that ClueBot must be doing something right! ;) -- œ 08:43, 31 October 2011 (UTC)

This is your last warning , if you keep on reverting vandalism you could be continuously praised without further notice.This was a joke! - funkyspyspy

  • Excellent work, you fixed the whale shark/whale man vandalism before I could log in and do it myself. --Mr. Snow (talk) 20:28, 25 January 2012 (UTC)

Hope your vacation went well, and you're ready to for constant vigilance once again. Crazynas t 21:33, 9 December 2011 (UTC)

I for one am really glad to see you back. I'm just a regular Wikipedian, but when I heard Cluebot NG went down I started freaking out and ended up watchlisting a bunch of pages, just so I could defend them. I was very concerned about the vandalism that was piling up on Wikipedia, and really hoping that your server issue would be addressed swiftly. Fortunately, the bot's back up and running now, so everything seems much better now. I'm sorry to hear about the dataset issues, but I've looked at its contributions and the vast majority of its reverts are vandalism, which pleases me very much. Seriously, well done on the bot programming, it truly is a remarkable accomplishment. If you ever have trouble with finances again, do speak up and I'm sure we the appreciative community will see what we can do. :D Impressed and pleased, Jessemv (talk) 07:40, 11 December 2011 (UTC)

  • So good to see your name again! Big hugz!! fredgandt 18:18, 12 December 2011 (UTC)

It's sad how little praise there is here. Anyway I was amazed that Cluebot detected this as vandalism... it has no junk characters, profanity or hate speech, it's a normally-written insertion of a tired but obscure conspiracy theory. Pretty cool that a bot could detect that as vandalism. Nice work. --Runame (talk) 15:07, 10 January 2012 (UTC)

  • Face-grin.svg Thumbs up 'bot. You got there first. benzband (talk) 22:08, 23 January 2012 (UTC)

ClueBot > Robocop + Chuck Norris. 'Nuff said. Agent 78787 (talk) 02:31, 28 January 2012 (UTC)

ClueBot is the new god. Please take a moment to hang out here and marvel at the efficiency. Michael Isaiah Schmidt (talk) 04:23, 7 March 2012 (UTC)

Thank you ClueBot for reverting Takis' page from the vandalism that has been ongoing. john sargis (talk)10:25, 30 January 2012 (UTC)

You will be blocking IPs next, huh? WOOT! --J (t) 04:09, 1 February 2012 (UTC)

Darn you Clue Bot! I was so exited that I caught a vandal in the act, and I was just about to revert it, when it disappeared. How can we possibly keep up with you? 169.232.131.133 (talk) 19:08, 8 February 2012 (UTC)

Hi ClueBot! Just wanted to say thank you for finding the false positive in the James B. Harkin article I was working on. Looks like a friend was playing around with my class project when he found out about it. Didn't realize it even occurred since you found it so quickly. Very amazing to see how quick wikipedia can catch things like this :) --Allisontheresa (talk) 22:04, 9 April 2012 (UTC)

You are so extremely and utterly awesome. When i grow up i wanna be just like you. I love you. Will you marry me? Please? U r my idol. u r hotter than jessica simpson — Preceding unsigned comment added by 76.209.128.18 (talk) 02:03, 11 April 2012 (UTC)

You're amazing! Always beats me to vandalism. :D C6541 (TC) 02:58, 20 April 2012 (UTC)

Thanks to ClueBot creators and maintainers for doing a great job. It is appreciated. One of the best things about its counter-vandalism work is that, in many cases, the vandals don't even have time to gloat at the results of their actions before they are reverted. One almost feels sorry for them... Maias (talk) 00:29, 2 May 2012 (UTC)

Quicker than a ray of light the vandalism's reverrrteeed! And I feeeel! Like vandalism just got undone! And I feeeeel! — Preceding unsigned comment added by Imadethisusernamemyself (talkcontribs) 21:03, 12 May 2012 (UTC)

Contributions

My Contributions



ClueBots
ClueBot NG/Anti-vandalism · ClueBot/Anti-vandalism · ClueBot II/ClueBot Script
Crystal Clear action run.png
ClueBot III/Archive · ClueBot VI/WP:CHUU Clerk  · Talk

Cobi/Owner // Talk
Personal tools
Namespaces

Variants
Actions
Navigation
Interaction
Toolbox
Print/export