Abstract: Embodiments of the present disclosure include a system for optimizing an artificial neural network by configuring a model, based on a plurality of training parameters, to execute a training process, monitoring a plurality of statistics produced upon execution of the training process, and adjusting one or more of the training parameters, based on one or more of the statistics, to maintain at least one of the statistics within a predetermined range. In some embodiments, artificial intelligence (AI) processors may execute a training process on a model, the training process having an associated set of training parameters. Execution of the training process may produce a plurality of statistics. Control processor(s) coupled to the AI processor(s) may receive the statistics, and in accordance therewith, adjust one or more of the training parameters to maintain at least one of the statistics within a predetermined range during execution of the training process.

Abstract: A cocktail making device that combines one or more recipes for a respective cocktail as well as the measurements into one product that is easy to clean. The cocktail making device includes a base cube, a top cavity, one or more first side cavities, and one or more second side cavities. The ingredient poured into the top cavity is printed on a surface of the cocktail making device for added convenience. The top cavity includes a pour spout disposed on each corner of the top cavity to facilitate ease of pouring the ingredient out of the top cavity. There is also a corresponding method for making a cocktail with the cocktail making device.

Abstract: An electric double layer capacitor (EDLC) is disclosed including: a first electrode including a first current collector and first plurality of carbon nanotubes (CNTs) disposed substantially directly upon the first current collector; a second electrode comprising a second current collector and second plurality of CNTs disposed substantially directly upon the second current collector; and an electrolyte disposed between and in contact with (e.g., wetting) the first and second electrodes. In some embodiments, the EDLC is configured to have a capacitive frequency window comprising about 1 Hz to about 50 Hz.

Abstract: A MEMS device is formed by a body of semiconductor material which defines a support structure. A pass-through cavity in the body is surrounded by the support structure. A movable structure is suspended in the pass-through cavity. An elastic structure extends in the pass-through cavity between the support structure and the movable structure. The elastic structure has a first and second portions and is subject, in use, to mechanical stress. The MEMS device is further formed by a metal region, which extends on the first portion of the elastic structure, and by a buried cavity in the elastic structure. The buried cavity extends between the first and the second portions of the elastic structure and communicates laterally with the pass-through cavity.

Abstract: The present application provides methods and systems for detecting and quantifying chromosomal instability from histology images with machine-learning. Also described herein are methods for selecting treatments for a medical disease, by determining a chromosomal instability metric from histology images. The disclosed methods and systems may also be used to investigate disease progression and prognosis.

Abstract: The present invention relates to compounds of formula (I) inhibiting Discoidin Domain Receptors (DDR inhibitors), methods of preparing such compounds, intermediate compounds useful in such preparations, pharmaceutical compositions containing them and therapeutic use thereof. The compounds of the invention may be useful for instance in the treatment of many disorders associated with DDR mechanisms.

Abstract: Disclosed herein are apparatuses and methods for treating neurological disorders by electro-stimulation. The apparatus (10) for treating neurological disorders includes at least one electrode (12) implantable in the brain of a patient, and a processing and stimulation device (14) connected to the at least one electrode (12). The processing and stimulation device (14) may include a stimulation module (16) configured to generate a stimulation signal to be sent to the at least one electrode (12), and an acquisition module (20) that measures cerebral activity coming from the brain of the patient. The acquisition module (20) may have a front-end block (27) configured to amplify the potential difference of its input signals (V1a, V2a) and to filter a stimulus artifact and may include a multi-stage fully-differential switched capacitor circuit (e.g., an integrated circuit) configured for discrete-time signal processing.

Abstract: Systems and methods described herein relate to determining whether to incorporate recognized text, that corresponds to a spoken utterance of a user of a client device, into a transcription displayed at the client device, or to cause an assistant command, that is associated with the transcription and that is based on the recognized text, to be performed by an automated assistant implemented by the client device. The spoken utterance is received during a dictation session between the user and the automated assistant. Implementations can process, using automatic speech recognition model(s), audio data that captures the spoken utterance to generate the recognized text. Further, implementations can determine whether to incorporate the recognized text into the transcription or cause the assistant command to be performed based on touch input being directed to the transcription, a state of the transcription, and/or audio-based characteristic(s) of the spoken utterance.

Abstract: An energy storage apparatus suitable for mounting on a printed circuit board using a solder reflow process is disclosed. In some embodiments, the apparatus includes: a sealed housing body (e.g., a lower body with a lid attached thereto) including a positive internal contact and a negative internal contact (e.g., metallic contact pads) disposed within the body and each respectively in electrical communication with a positive external contact and a negative external contact. Each of the external contacts provide electrical communication to the exterior of the body, and may be disposed on an external surface of the body. An electric double layer capacitor (EDLC) (also referred to herein as an “ultracapacitor” or “supercapacitor”) energy storage cell is disposed within a cavity in the body including a stack of alternating electrode layers and electrically insulating separator layers. An electrolyte is disposed within the cavity and wets the electrode layers.

Abstract: An electrical storage system comprises a first energy storage system and a second energy storage system having a lower electrical energy density and a higher rated electrical power output capability than the first energy storage system, at least one electrical power sensor configured to sense over a plurality of time intervals, electrical power usage information for a load electrically coupled to the first energy storage system and the second energy storage system, and at least one computer processor programmed to determine based, at least in part, on the sensed electrical power usage information and a power requirement of the load in a current time interval, charging/discharging parameters for each of the first energy storage system and the second energy storage system, and control charging/discharging of each of the first and second energy storage systems in accordance with the determined charging/discharging parameters during the current time interval.

Abstract: The present invention relates to a method for the diagnosis of tumoural conditions and/or of the corresponding state of advance, wherein a sample from a patient comprising at least one tumour cell is obtained. According to the invention, a purified specimen of the at least one tumour cell is obtained by individually selecting and isolating single cells in a microfluidic device the purified specimen having a purity of at least 90%. On the purified specimen thus obtained there is subsequently performed a molecular analysis such as to highlight a characteristic thereof suited to enabling diagnosis.

Abstract: A method of monitoring and protecting access to an online service from Account Take Over may include: providing a Traffic Inspector in communication with a client device and web server; providing a Traffic Analyzer in communication with the Inspector; identifying each browsing session of the client device; extracting and identifying one or more usernames when a user performs authentication to the online service by analyzing traffic exchanged between the client device and web server; collecting first data concerning unique and/or non-unique technical parameters and associating the first data with respective identified one or more usernames and with the client device; collecting second data concerning unique and/or non-unique technical parameters and associating the second data with an anonymous application browsing session or web beacon; and taking security measures when a number of identified usernames associated with the anonymous application browsing session or web beacon exceeds a predetermined threshold v

Abstract: An electrode active layer is disclosed that includes a network of high aspect ratio carbon elements (e.g., carbon nanotubes, carbon nanotube bundles, graphene flakes, or the like) that provides a highly electrically conductive scaffold that entangles or enmeshes the active material, thereby supporting the layer. A surface treatment can be applied to the high aspect ratio carbon elements to promote adhesion to the active material and any underlying electrode layers improving the overall cohesion and mechanical stability of the active layer. This surface treatment forms only a thin (in some cases even monomolecular) layer on the network, leaving the large void spaces that are free of any bulk binder material and so may instead be filled with active material. The resulting active layer may be formed with excellent mechanical stability even at large thickness and high active material mass loading.

Abstract: The present disclosure relates to a method for analyzing the degree of similarity of at least two samples in a plurality of samples comprising genomic DNA. The method comprises the following steps. a) Providing a plurality of samples comprising genomic DNA. b) Carrying out, separately on each sample, a deterministic restriction-site whole genome amplification (DRS-WGA) of said genomic DNA, c) Preparing a massively parallel sequencing library using a fragmentation-free, sequencing-adaptor/WGA fusion-primer PCR reaction from each product of said DRS-WGA. d) Carrying out low-pass whole genome sequencing at a mean coverage depth of <1× on said massively parallel sequencing library. e) Aligning for each sample the reads obtained in step d) on a reference genome. f) Extracting for each sample the allelic content at a plurality of polymorphic loci. g) Calculating a pair-wise similarity score for the at least two samples as a function of the allelic content measured at said plurality of loci.

Abstract: The present invention pertains to a method for making a reinforced shaped article comprising the steps of: a) providing one or more fluoropolymers having a melting point Tm, said one or more fluoropolymers being characterized in that they: (i) comprise more than 30% by moles of recurring units derived from VDF, with respect to the total number of recurring units of the polymer; (ii) have iodine-containing chain ends —CH2I in an amount of 0.1 to 0.9 per chain; b) dissolving said fluoropolymer in a suitable solvent and forming a precursor solution c) precipitating said fluoropolymer as a solid from said precursor solution thereby obtaining a shaped article comprising said fluoropolymer e) thermally treating said shaped article for at least 15 minutes at a temperature comprised between 100° C. and the lowest melting point Tm of said one or more fluoropolymers thereby obtaining said reinforced shaped article.

Abstract: Examples of the present disclosure describe systems and methods for probabilistic neural network architecture generation. In an example, an underlying distribution over neural network architectures based on various parameters is sampled using probabilistic modeling. Training data is evaluated in order to iteratively update the underlying distribution, thereby generating a probability distribution over the neural network architectures. The distribution is iteratively trained until the parameters associated with the neural network architecture converge. Once it is determined that the parameters have converged, the resulting probability distribution may be used to generate a resulting neural network architecture. As a result, intermediate architectures need not be fully trained, which dramatically reduces memory usage and/or processing time.

Abstract: A purification method for a biological fluid comprising at least a filtration step through a membrane obtained from a sulfone polymer (PSI) derived from bio-based feed-stocks. In particular the polymer (PSI) comprises more than 50% moles recurring units (RPSI) comprising sugar moieties selected from the group consisting of those of formulae (E?-I) to (E?-III): A membrane free from pore-forming agents comprising at least one sulfone polymer (PSI) and a method for preparing such a membrane.

Abstract: Disclosed herein is a thermal interface material comprising a sheet extending between a first major surface and a second major surface, the sheet comprising a base material; and a filler material embedded in the base material comprising anisotropically oriented thermally conductive elements; wherein the thermally conductive elements are preferentially oriented along a primary direction from the first major surface towards the second major surface to promote thermal conduction though the sheet along the primary direction; and wherein the base material is substantially free of silicone.

Abstract: It is provided a joint prosthesis (1) configured to reciprocally rotate a first prosthesis (1a) and a second prosthesis (1b) and comprising: a first attachment (2) to the first prosthesis (1a); a second attachment (3) to the second prosthesis (1b): a hinge (4) defining an axis of rotation (4a) between the attachments (2, 3); a mover (5) configured to apply to the hinge (4) a driving torque defining a reciprocal rotation between the attachments (2, 3); and a compensator (6) configured to apply to the hinge (4) a variable locking torque depending on the angular opening between the attachments (2, 3) and defining the minimum torque to be applied to the hinge (4) to have said reciprocal rotation.

Abstract: Some implementations process, using warm word model(s), a stream of audio data to determine a portion of the audio data that corresponds to particular word(s) and/or phrase(s) (e.g., a warm word) associated with an assistant command, process, using an automatic speech recognition (ASR) model, a preamble portion of the audio data (e.g., that precedes the warm word) and/or a postamble portion of the audio data (e.g., that follows the warm word) to generate ASR output, and determine, based on processing the ASR output, whether a user intended the assistant command to be performed. Additional or alternative implementations can process the stream of audio data using a speaker identification (SID) model to determine whether the audio data is sufficient to identify the user that provided a spoken utterance captured in the stream of audio data, and determine if that user is authorized to cause performance of the assistant command.