Microsoft Technology Licensing, LLC

Arunkumar Gururajan

Jack Wilson Stokes, Iii

Diana Anca Carutasu

Edir Vinicio Garcia Lazo

Farid Tajaddodianfar

Pranav Ravindra Maneriker

Michelle Elena Keslin

Daniel Robert Snyder

Eugene Chudin

Si-Qing Chen

Tony Y Tzeng

John Hoegger

Jeff D Jarrard

Mihai Aldea

Sanjeev Bagaria

Theodor J Scott

Cynthia Guo

Kamal Choudhary

Jin Hee Lim

Jack Wilson Iii Stokes

The technology described herein can identify phishing URLs using transformers. The technology tokenizes useful features from the subject URL. The useful features can include the text of the URL and other data associated with the URL, such as certificate data for the subject URL, a referrer URL, an IP address, etc. The technology may build a joint Byte Pair Encoding for the features. The token encoding may be processed through a transformer, resulting in a transformer output. The transformer output, which may be described as a token embedding, may be input to a classifier to determine whether the URL is a phishing URL. Additional or improved URL training data may be generated by permuting token order, by simulating a homoglyph attack, and by simulating a compound word attack.

Phishing URL detection using transformers

The technology described herein identifies malicious URLs using a classifier that is both accurate and fast. Aspects of the technology are particularly well adapted for use as a real-time URL security analysis tool because the technology is able to quickly process a URL and produce a warning when a malicious URL is identified. The rapid processing speed of the technology described herein is produced, in part, by use of only a single input signal, which is the URL itself. The high accuracy produced by the technology described herein is achieved by analyzing the unstructured text on both a character-by-character level and a word-by-word level. The technology described herein uses both character-level and word-level information from the incoming URL.

Unstructured text classification

The technology described herein can identify phishing URLs using transformers. The technology tokenizes useful features from the subject URL. The useful features can include the text of the URL and other data associated with the URL, such as certificate data for the subject URL, a referrer URL, an IP address, etc. The technology may build a joint Byte Pair Encoding for the features. The token encoding may be processed through a transformer, resulting in a transformer output. The transformer output, which may be described as a token embedding, may be input to a classifier to determine whether the URL is a phishing URL. Additional or improved URL training data may be generated by permuting token order, by simulating a homoglyph attack, and by simulating an a compound word attack.

Automatic generation of document templates based on recognized composition element patterns in a group of clustered documents is provided. Composition elements used in documents are typically unique to a particular user or to a group of users. An automated template generation system detects composition element patterns in documents associated with a given user. Sequences of composition elements from one document are aligned with sequences of composition elements of one or more other documents. The aligned sequences are scored to generate a document distance matrix. The documents are clustered together based on the alignment scores and a document template is generated for each corresponding cluster of documents. In one or more aspects, selecting a document template and updating it results in a modified document template or, in certain cases, a new document template. The generated document templates are displayed in a user interface for selection by a user.

Generating templates from user's past documents

Growing community of inventors

Redmond, WA, United States of America

Arunkumar Gururajan