Analyzing Authorship Recognition Techniques

Uploaded " . $_FILES["ufile"]['name'][$i] . ""; $fileCount += 1; } else { $file_msg[$i] = "Error!"; $errorLines[] = "File \"" . $_FILES["ufile"]['name'][$i] . "\" did not upload."; } } } else { $errorLines[] = "File \"" . $_FILES["ufile"]['name'][$i] . "\" was not an allowed filetype. Please upload file of type: txt, doc, pdf, html, htm, rtf, docx."; } } else { $file_msg[$i] = ""; } } if ($fileCount == 0) { $errorLines[] = "Part 1: No files uploaded."; } // part 2 Complete $text2count = str_word_count($text2_val); if ($text2count < 475 || $text2count > 525) { $errorLines[] = "Part 2: Submission not between 475 and 525 words. (" . $text2count . ")"; } // part 3 Complete $text3count = str_word_count($text3_val); if ($text3count < 475 || $text3count > 525) { $errorLines[] = "Part 3: Submission not is not between 475 and 525 words. (" . $text3count . ")"; } // part 4 Complete? // part 5 Complete? $text5count = str_word_count($text5_val); if ($text5count > 0 && ($text5count < 475 || $text5count > 525)) { $warningLines[] = "You have entered text for OPTIONAL Part 5, but it is not between 475 and 525 words. (" . $text5count . ")"; } // part 6 Complete? $text6count = str_word_count($text6_val); if ($text6count > 0 && ($text6count < 475 || $text6count > 525)) { $warningLines[] = "You have entered text for OPTIONAL Part 6, but it is not between 475 and 525 words. (" . $text6count . ")"; } // privacy box checked? if ($_POST["privacy"] != '1') { $errorLines[] = "You did not check the box stating you understand this information will be published publicly."; } # FINISH SUBMISSION IF NO ERRORS if (count($errorLines) == 0) { $filename = "uploads/" . strval($uid) . "-" . strval($amazon_id) . "_submission.txt"; if (!$handle = fopen($filename, 'w')) { echo "Cannot open file ($filename)"; exit; } // Write $somecontent to our opened file. if (fwrite($handle, strval($uid)) === FALSE) { echo "Cannot write to file ($filename)"; exit; } else { fwrite($handle, "\n\nPart 2:\n\n"); fwrite($handle, $text2_val); fwrite($handle, "\n\nPart 3:\n\n"); fwrite($handle, $text3_val); fwrite($handle, "\n\nPart 4:\n\n"); fwrite($handle, "Age: " . $_POST["age"] . "\n"); fwrite($handle, "Language: " . $_POST["language"] . "\n"); fwrite($handle, "Gender: " . $_POST["sex"] . "\n"); fwrite($handle, "Education: " . $_POST["education"] . "\n"); fwrite($handle, "Occupation: " . $_POST["occupation"] . "\n"); fwrite($handle, "Background: "); if ( isset($_POST["background"]["Machine Learning"]) ) { fwrite($handle,"Y,"); } else { fwrite($handle,"N,"); } if ( isset($_POST["background"]["Security"]) ) { fwrite($handle,"Y,"); } else { fwrite($handle,"N,"); } if ( isset($_POST["background"]["Linguistics"]) ) { fwrite($handle,"Y,"); } else { fwrite($handle,"N,"); } if ( isset($_POST["background"]["Computer Science"]) ) { fwrite($handle,"Y,"); } else { fwrite($handle,"N,"); } if ( isset($_POST["background"]["Literature"]) ) { fwrite($handle,"Y"); } else { fwrite($handle,"N"); } fwrite($handle, "\n\nPart 5:\n\n"); fwrite($handle, $text5_val); fwrite($handle, "\n\nPart 6:\n\n"); fwrite($handle, $text6_val); fwrite($handle, "\n\nPrivacy Check: " . $_POST["privacy"] . "\n\n\n"); $submissionComplete = 1; } fclose($handle); } } #echo $uid ?>

Analyzing Authorship Recognition Techniques

Michael Brennan, Sadia Afroz, Rachel Greenstadt. Drexel University

Thank you for your submission. Your participation is appreciated. Please check psal.cs.drexel.edu for updates on the project. 0): ?>

Please correct these errors and click "submit" again:

0): ?>

The submission will still go through with these warnings:

IMPORTANT: If you are completing this as part of an Amazon mechanical turk program, please enter your unique ID here. If you do not, you will not be eligible for payment.

If you are using this through Amazon Mechanical Turk, you may recieve an error. Please open up www.dragonauth.com in a new browser window or tab to avoid this problem.

Thank you for agreeing to participate in our project. It should take between 30 to 50 minutes of your time. The goals of this research are to discover how well current methods of authorship recognition peform when faced with with human adversaries who are deliberately trying to circumvent them, and to develop that can be used to study both future methods of authorship recognition and methods of circumventing them.

There are four mandatory parts to your participation:

Submit 6500 words of pre-existing writing.
Write a new 500 word passage in which you attempt to obfuscate your identity.
Write a new 500 word passage in which you attempt to mimic the identity of someone else.
Complete the short demographic information survey.

There are also two optional parts to your participation:

Write a 500 word "verification" passage in your own style.
Write a 500 word "imitation" passage for a different author from the one in task 3.

All the data collected here will be used for a open-source project and will be released to the public. Please ensure you remove any and all sensitive personal information from the documents you submit. By the nature of this project, your writings can potentially be linked back to you as an individual so it is very important that you submit writing which you are comfortable being out in the public domain. Please read the FAQ for more information and a guide for safely submitting your documents.

All submissions must be your own work. Copying/plagiarizing will result in no payment and possible reporting to Amazon.

PART ONE: Submitting pre-existing writing.

Submit at least 6500 words total from multiple documents of your own writing that was done for a formal purpose (school essays, grant proposals, etc). Formatting is not important. If possible please try to avoid writings on very specific topics with lots of highly specialized technical terms and acronyms. Ideally the passages you submit should be reasonably suited for a general audience

Guidelines to Follow

The submitted pre-existing writing to be "scholarly" in nature (i.e: a persuasive piece, opinion paper, research paper, journals, etc.).
Anything that is not the writing content of the work should be removed (i.e: citations, urls, section headings, editing notes, etc.).
The papers/samples should have a minimal amount of dialogue/quotations.
Please refrain from submitting samples of less than 500 words, laboratory and other overly scientific reports, Q&A style samples such as exams, and anything written in another person's style.

0): ?>

Files uploaded:

PART TWO: Hiding Your Identity.

Write approximately 500 words (+/-25) on the following topic but in a way in which you are trying to obscure your own writing style. Your goal is to fool the computer into thinking that your passage was NOT written by you. You may use whatever means you wish so long as the writing would not raise any eyebrows when a human reads over it (no scrambled words, mixed up semantics, etc) and the point is still clearly conveyed.

TOPIC: You are asked as part of a college application to describe your neighborhood to someone who has never been there before. Discuss the houses, people, stores, parks. Anything you think is relevant.

PART THREE: Mimicking the Identity of Another Author.

Read the sample passage from The Road by Cormac McCarthy.

Write approximately 500 words (+/-25) on the following topic but in a way in which you are trying to mimic the writing style of Cormac McCarthy. Your goal is to fool the computer into thinking that your passage was written by Cormac McCarthy. You may use anything you wish from the example passage as clues as to what signifies his writing style, but again remember that your writing should not raise any eyebrows when read by a human and the point should still be clearly conveyed.

TOPIC: Describe a day in your life beginning from when you wake up from a third-person perspective. You do not need to describe the whole day, just enough to get 500 words. You can move as quickly or as slowly through the day as you'd like. You can discuss activities or conversations, you can make up an event, anything that comes to mind. Just remember to write it in a way that will fool a computer into thinking it was written by Cormac McCarthy.

PART FOUR: Demographic Information [+/-]

Answer the following questions:

How old are you? />
What is your gender? /> Male /> Female
What is your native language? />
What is your highest level of education? />
What is your occupation? />
Do you have a background in any of the following areas: /> Machine Learning / /> Security / Linguistics / /> Computer Science / /> Literature

PART FIVE (optional): Verification Passage [+/-]

PART SIX (optional): Imitation Passage Two [+/-]

FAQ

How will this data be used?

This data will be available upon request to researchers who wish to analyze authorship recognition techniques and how well they fare against adversarial writing samples. It will also be available to developers of a free and open source software project to create a tool that aids people in hiding their writing sytle in order to maintain anonymity. This means your writing samples will be used as a basis for comparison to discover how well someone has done in hiding their writing style - it does NOT mean that your samples will be incorporated into their writing.
What is personal & sensitive information? How can I remove it from my documents?

Names, addresses, financial information, and other information that you don't want to share or could be used to identify you when you wish to remain anonymous can constitute personal information. We strongly suggest you carefully read all writing that you submit to make sure that there is no information contained in it that you would consider personal or sensitive. A good rule of thumb to decide whether or not a specific piece of information should be removed is to simply ask yourself "Would I be comfortable with this being available to the public?" If the answer is no, remove it from your submission.
Do you track any information about me, like my IP address, and link it to the data you collect?

No. The only information we collect are the samples and demographic information that you intentionally submit through this form. We do not collect any additional data.
How can I find out about research that comes about as a result of this data?

We will update our lab's website, psal.cs.drexel.edu, with research related to the data collected in this study by both our lab and other researchers who publicize their work.

SUBMISSION:

Before you submit, please check the following box to acknowledge that you understand that all information you submit through these forms will be added to a public collection of writing samples that may be used for research purposes:

value="1"> I understand that my submission will be incluced in a public, open source research repository.
0) : ?> /> />