Trend Micro Incorporated

A target binary file is clustered by reducing the target binary file to its architecture-agnostic functions, which are converted into an input string. The target digest of the input string is calculated and compared to digests of malicious binary files. A cluster having digests of malicious binary files that are similar to the target digest is identified. In response to identifying the cluster, the target binary file is detected to be malicious and of the same malware family as the malicious binary files of the cluster.

A file is stored in a public cloud storage. A serverless computing platform receives an event notification that the file has been stored and, in response, creates an instance of an ephemeral environment wherein a security module is executed. The security module creates a memory-mapped space with memory locations that are mapped to the entire content of the file but does not allocate memory for all of the memory locations. Instead, the security module retrieves sections of the file from the public cloud storage as these sections are accessed in their designated memory locations in accordance with the memory mapping, allocates memory for the retrieved sections, stores the retrieved sections in their designated memory locations, and scans the retrieved sections in their designated memory locations for malicious code. The security module continues scanning the file in sections until relevant sections of the file have been scanned.

An endpoint system receives a target file for evaluation for malicious scripts. The original content of the target file is normalized and stored in a normalized buffer. Tokens in the normalized buffer are translated to symbols, which are stored in a tokenized buffer. Strings in the normalized buffer are stored in a string buffer. Tokens that are indicative of syntactical structure of the normalized content are extracted from the normalized buffer and stored in a structure buffer. The content of the tokenized buffer and counts of tokens represented as symbols in the tokenized buffer are compared against heuristic rules indicative of malicious scripts. The contents of the tokenized buffer and string buffer are compared against signatures of malicious scripts. The contents of the tokenized buffer, string buffer, and structure buffer are input to a machine learning model that has been trained to detect malicious scripts.

A machine learning system includes multiple machine learning models. A target object, such as a file, is scanned for machine learning features. Context information of the target object, such as the type of the object and how the object was received in a computer, is employed to select a machine learning model among the multiple machine learning models. The machine learning model is also selected based on threat intelligence, such as census information of the target object. The selected machine learning model makes a prediction using machine learning features extracted from the target object. The target object is allowed or blocked depending on whether or not the prediction indicates that the target object is malicious.

A virtual keyboard rendered on a separate computing device is independent of the user's computer. A virtual keyboard displayed on the user's computer screen is blank without any alphanumeric characters. Another virtual keyboard displayed on the user's independent computing device has a randomly generated layout of alphanumeric characters on a keypad. The user enters a password by pressing the blank keys of the blank keyboard on his computer screen with reference to the other virtual keyboard. The position sequence of these entered keys is sent to an application on a remote server computer. The remote server computer shares a virtual keyboard having the randomly generated layout of characters with the independent computing device via an online or off-line technique. When online, an encoded image of the encrypted layout is sent to the client computer and displayed for scanning by the device. When off-line, both the application and the device generate the same random key sequence by using the same pseudo random number generator and the same seed value.

An inline network traffic monitor is deployed inline between two endpoints of a computer network. A particular endpoint of the two endpoints works in conjunction with the inline network traffic monitor to decrypt encrypted network traffic transmitted between the two endpoints. A series of Change Cipher Spec (CCS) messages is exchanged between the inline network traffic monitor and the particular endpoint during a Transport Layer Security (TLS) handshake between the two endpoints. The series of CCS messages allows the particular endpoint and the inline network traffic monitor to detect each other on the computer network. After detecting each other's presence, the particular endpoint sends the inline network traffic monitor a session key that is used by the two endpoints to encrypt their network traffic. The inline network traffic monitor uses the session key to decrypt encrypted data of the network traffic transmitted between the two endpoints.

An attachment to an e-mail message received at an e-mail gateway is scanned by a scan server and then is converted into an HTML file. The HTML file includes preview data of the attachment (minus any macro scripts), the entire original data of the attachment, scan functionality enabling a user to send the attachment back to a scan server for a second scan, or extract functionality enabling a user to extract the original attachment data for saving or opening in an application. The recipient is able to open or save the attachment directly if he or she believes it comes from a trusted sender. If the attachment seems suspicious, the recipient previews the attachment first before performing a scan, opening the attachment or deleting it. The recipient performs a scan of the attachment by clicking a “scan” button to send the attachment to a backend server for a second scan where an updated virus pattern file may be available to detect any zero-day malware.

A pause command is sent to a Subscriber Identity Module (SIM) card of a cellular device in response to detecting a cyberattack against the cellular device on the cellular network. To mitigate the cyberattack, the SIM card temporarily disconnects the cellular device from the cellular network for a pause time. The SIM card prohibits the cellular device from connecting to the cellular network during the pause time and automatically allows the cellular device to reconnect to the cellular network after the pause time.

Disclosed are a method and system for static behavior-predictive malware detection. The method and system use a transfer learning model from behavior prediction to malware detection based on static features. In accordance with an embodiment, machine learning is used to capture the relations between static features, behavior features, and other context information. For example, the machine learning may be implemented with a deep learning network model with multiple embedded layers pre-trained with metadata gathered from various resources, including sandbox logs, simulator logs and context information. Synthesized behavior-related static features are generated by projecting the original static features to the behavior features. A final static model may then be trained using the combination of the original static features and the synthesized features as the training data. The detection stage may be performed in real time with static analysis because only static features are needed. Other embodiments and features are disclosed.

Language-based machine learning approach for automatically detecting universal charset and the language of a received document is disclosed. The language-based machine learning approach employs a plurality of text document samples in different languages, after converting them to a selected Unicode style (if their original encoding schemes are not the selected Unicode), to generate a plurality of language-based machine learning models during the training stage. During the application stage, vector representations of the received document for different combinations of charsets and their respective applicable languages are tested against the plurality of machine learning models to ascertain the charset and language combination that is most similar to its associated machine learning model, thereby identifying the charset and language of the received document.