Abstract: The present application discloses a method, system, and computer system for detecting malicious SQL or command injection strings. The method includes obtaining an SQL or command injection string and determining whether the command injection string is malicious based at least in part on a machine learning model.

Abstract: A package structure includes a first redistribution layer, a semiconductor die, and through vias. The first redistribution layer includes dielectric layers, first conductive patterns, and second conductive patterns. The dielectric layers are located in a core region and a peripheral region of the first redistribution layer. The first conductive patterns are embedded in the dielectric layers in the core region, wherein the first conductive patterns are arranged in the core region with a pattern density that gradually increases or decreases from a center of the core region to a boundary of the core region. The second conductive patterns are embedded in the dielectric layers in the peripheral region. The semiconductor die is disposed on the core region over the first conductive patterns. The through vias are disposed on the peripheral region and electrically connected to the second conductive patterns.

Abstract: A MIM capacitor structure includes a semiconductor substrate, and a first trench and a second trench in the semiconductor substrate in a capacitance forming region. The second trench is adjacent to the first trench. The second trench is deeper than the first trench. A dielectric liner layer conformally covers a top surface of the semiconductor substrate and interior surfaces of the first trench and the second trench. A bottom electrode layer conformally covers the dielectric liner layer. The bottom electrode layer extends onto a top surface of the semiconductor substrate. A capacitor dielectric layer is disposed on the bottom electrode layer in the first trench and the second trench. A top electrode layer is disposed on the capacitor dielectric layer in the first trench and the second trench. The top surface of the top electrode layer is coplanar with the top surface of the bottom electrode layer.

Abstract: An LLM is adapted to generate database queries that are compatible with a proprietary database of a security provider. Adapting the LLM includes evaluating performance of the LLM after initial prompt engineering/fine-tuning to ensure that generated database queries are valid (i.e., comport to the database schema and can be executed to return results). When the LLM performance is satisfactory, a dialogue system uses the LLM to generate database queries from user queries. The dialogue system determines intent of each user query, which informs whether the query is supported. Supported user queries are converted to database queries using the LLM and submitted to the database. The dialogue system leverages another language model to generate a summarized, natural language representation of the database query results and constructs a response from the summary. The dialogue system also checks for XSS and prompt injection before database queries are ultimately submitted to the database.

Abstract: A wearable device, gesture recognition method, and non-transitory computer readable storage medium thereof are provided. The device determines at a first time point that a single-channel electromyography signal in a first time interval is in a non-neutral state or a neutral state. In response to the single-channel electromyography signal in the first time interval being in the non-neutral state, the device recognizes a gesture of a user at the first time point based on a plurality of inertial measurement parameters.

Abstract: New methods and kits for treating diseases caused or exacerbated by overactivated EphA4 signaling are provided. The method includes administering to a subject in need thereof an effective amount of a small molecule compound inhibitor for EphA4 signaling. Also provided are methods for identifying additional compounds as therapeutic agents useful for treating conditions involving overly active EphA4 signaling.

Abstract: Compositions and methods are provided for the treatment or prophylaxis of amyloid pathologies such as Alzheimer's Disease.

Abstract: An exhaust system for discharging from semiconductor manufacturing equipment a hazardous gas includes: a main exhaust pipe above the semiconductor manufacturing equipment and having a top surface on a first side and a bottom surface on a second side, a first branch pipe connected to a source of a gas mixture containing the hazardous gas on the second side and connected to the main exhaust pipe through the top surface, a second branch pipe connected to a gas box on the second side and connected to the main exhaust pipe through the bottom surface, and a detector on the second branch pipe configured to detect presence of the hazardous gas and downstream to the gas box. The first and the second branch pipes are connected to the main exhaust pipe at a first location and a second location, respectively. The first location is more upstream than the second location.

Abstract: A modular adversarial training data generator (“generator”) generates adversarial training data for a machine learning (ML) model to detect malicious protocol data units (PDUs). The adversarial training data mimics high volume cyberattacks by perturbing PDUs to bypass malicious detection systems. For HyperText Transfer Protocol (HTTP) PDUs, the generator rearranged, replaces, and grid searches values of HTTP header fields to generate the adversarial training data. The generator further biases grid search based on metrics for values of HTTP header fields that quantify impact of replacing the values on malicious verdicts by ML models trained on the adversarial training data.

Abstract: A system stores first and second images generated by first and second cameras; applies a segmentation model to the first image to generate a first segmentation mask identifying object instances; applies the segmentation model to the second image to generate a second segmentation mask identifying the object instances; projects the first segmentation mask to a viewpoint of the second camera to generate a first projected segmentation mask; converts the first projected segmentation mask and the second segmentation mask to first and second semantic masks, respectively; and computes a first similarity value based on the first and second semantic masks. This may be repeated exchanging the first and second images to compute a second similarity value. The system determines a loss value based on the first similarity value and the second similarity value and trains the segmentation model based on the loss value.

Abstract: Disclosed are systems and techniques for image processing. For example, a computing device can generate, using a multi-task model, a segmentation output and a normal output based on image(s) of a scene and a gravity vector for the scene. The computing device can learn semantic prediction(s) based on comparing the segmentation output to at least one ground truth semantic segmentation map. The computing device can also learn normal prediction(s) based on comparing the normal output to at least one ground truth normal map. The computing device can extract a semantics normal from the semantic prediction(s) and the normal prediction(s). The computing device can optimize a regularization loss based on the semantics normal and the gravity vector for the scene by learning gravity-normal regularization(s) for the scene. The computing device can determine final semantic labels for regions of the scene based on the gravity-normal regularization(s).

Abstract: Various embodiments provide a system, method, and device for applying a C2 machine learning-based detection framework. The method incudes (i) generating a fuzzing based on a C2 machine-learning detection model using a large learning model for performing profile-based seed generation; and (ii) detecting C2 traffic using the C2 machine learning detection model.

Abstract: A method of setting a display mode includes installing an application on a device, the device including a first screen; triggering at least one button of a display, the display providing display information to the application installed on the device through a data channel protocol, the display including a second screen; and the application switching the device and the display to the display mode based on the display information.

Abstract: An image displaying device includes a planar display panel and a light penetrating unit. The planar display panel displays a plane image. The planar display panel at least includes a first pixel group, a second pixel group and a third pixel group. The second pixel group is located between the first pixel group and the third pixel group. When vision passes through the light penetrating unit toward the planar display panel, the vision acquires a second distance of a second imaging position within the plane image relevant to the second pixel group relative to the planar display panel being greater than a first distance of a first imaging position within the plane image relevant to the first pixel group relative to the planar display panel and a third distance of a third imaging position within the plane image relevant to the third pixel group relative to the planar display panel.

Abstract: A process for retting biomass includes contacting an enzyme solution with the biomass, degrading at least a portion of the pectin in the biomass using the enzyme to release monomers, proliferating bacteria using the monomers as a feedstock, and separating the cellulosic fibers after degrading at least the portion of the pectin to produce natural fibers. The enzyme solution comprises an enzyme, and the biomass comprises pectin and cellulosic fibers.

Abstract: An electronic device includes a shell element, at least one electronic element, a printed circuit board, and an anti-scalding mask. The shell element has an accommodating space and a metal surface. At least one electronic element is disposed in the accommodating space. The printed circuit board is disposed in the accommodating space and is electrically connected to the at least one electronic element. The anti-scalding mask covers and is disposed on the metal surface. The anti-scalding mask is made of a low thermal conductivity material, and a thermal conductivity of the low thermal conductivity material is less than or equal to 20 W/mK.

Abstract: An exploit detector hooks functions of a WebAssembly engine API that include functions invoked when creating a WebAssembly instance, when permissibly modifying read-write-execute (RWX) memory allocated for a WebAssembly instance, and when calling an exported WebAssembly function. Upon invocation of the hooked instance creation function(s), the detector obtains an address of RWX memory allocated for the instance, reads data stored there, and records an association between the address and the data. If the hooked function(s) for permissibly modifying RWX memory is invoked, the detector updates the data stored in the recorded association with the modified data read from the RWX memory. Based on invocation of the hooked function(s) associated with WebAssembly export function calls, the detector reads the current data in the RWX memory to be executed and compares this data to the data recorded in association with the address. If the actual and expected data differ, the detector detects a browser exploit.

Abstract: An in-kernel virtual machine (“VM”) instantiated in the kernel space of a physical or virtual machine (“machine”) attaches code hooks at a network interface of the machine. The code hooks redirect network traffic from the network interface to a module that generates feature values at the kernel space from protocol data units of network traffic received by the network interface. The machine passes the feature values from kernel space to user space via zero-copy shared memory and a machine learning model in the user space obtains network traffic verdicts as outputs from inputting the feature values. The machine passes the verdicts from user space to kernel space via the zero-copy shared memory and the in-kernel VM performs corrective action based on malicious verdicts.

Abstract: A system, method, and device for detecting Command and Control (C2) traffic is disclosed. The method includes (i) converting, by one or more processors, a header for network traffic to a header representation having a smaller dimensionality than the header, (ii) querying a classifier based at least in part on the header representation to obtain a traffic classification, (iii) automatically detecting C2 traffic based at least in part on the traffic classification, and (iv) handling the network traffic based at least in part on the traffic classification.

Abstract: In a described example, an example no-lead semiconductor device package includes: a die pad in a central portion of a partially etched leadframe and terminals. The terminals include a device side surface formed in an upper layer of the partially etched leadframe, and a first exterior end in the upper layer; a board side surface formed in a lower layer of the partially etched leadframe extending from the upper layer; a second exterior end in the lower layer, the second exterior end inset from the first exterior end, and an inset portion extending from the first exterior end to the second exterior end. A semiconductor die is mounted to the die pad. Mold compound covers the semiconductor die, the second exterior end of the terminals and the inset portion of the terminals are exposed from the mold compound, with the mold compound extending along sides of the terminals.