A neuro-bionic device based on a two-dimensional Ti.sub.3C.sub.2 material is provided. The device includes a Pt/Ti/SiO.sub.2/Si substrate, a neuro-bionic layer formed on a Pt film layer of the Pt/Ti/SiO.sub.2/Si substrate, and an Al electrode layer formed on the neuro-bionic layer. The neuro-bionic layer is made of a two-dimensional Ti.sub.3C.sub.2 material. The neuro-bionic device of the present invention is prepared by an evaporating coating method and a drop-coating method. The preparation process is relatively simple. The prepared device can successfully simulate the characteristics of synapse. More importantly, the resistance of the device can be modulated continuously under a scanning of a pulse sequence with pulse width and interval of 10 ns, which is beneficial to the application of the device in the ultrafast synapse simulation.

The present invention provides a method for detecting events in an input signal. The method uses a volatile resistive switching component to detect the events in the input signal. The method comprising identifying the events based on sampling an output from the resistive switching component.

The disclosed computer-implemented method may include (1) identifying an artificial neural network comprising a set of nodes interconnected via a set of connections, and (2) training the artificial neural network by, for each connection in the set of connections, determining a quantized weight value associated with the connection. Determining the quantized weight value associated with the connection may include (1) associating a loss function with the connection, the loss function including a periodic regularization function that describes a relationship between an input value and a weight value of the connection, (2) determining a minimum of the associated loss function with respect to the weight value in accordance with the periodic regularization function, and (3) generating the quantized weight value associated with the connection based on the determined minimum of the loss function. Various other methods, systems, and computer-readable media are also disclosed.

Systems and methods are provided for reducing power in in-memory computing, matrix-vector computations, and neural networks. An apparatus for in-memory computing using charge-domain circuit operation includes transistors configured as memory bit cells, transistors configured to perform in-memory computing using the memory bit cells, capacitors configured to store a result of in-memory computing from the memory bit cells, and switches, wherein, based on a setting of each of the switches, the charges on at least a portion of the plurality of capacitors are shorted together. Shorting together the plurality of capacitors yields a computation result.

In one aspect, there is provided a method performed by one or more data processing apparatus, the method including: obtaining data defining a connectivity graph that represents synaptic connectivity between multiple biological neuronal elements in a brain of a biological organism, where the connectivity graph includes: multiple nodes, and multiple edges that each connect a respective pair of nodes, determining a partition of the connectivity graph into multiple community sub-graphs by performing an optimization that encourages a higher measure of connectedness between nodes included within each community sub-graph relative to nodes included in different community sub-graphs, and selecting a neural network architecture for performing a machine learning task using multiple community sub-graphs determined by the optimization that encourages the higher measure of connectedness between nodes included within each community sub-graph relative to nodes included in different community sub-graphs.

A system for biologically determined artificial intelligence selection guidance includes a computing device designed and configured to receive at least a biological extraction and an item descriptor from a user, generate, using a classification algorithm and a plurality of past extractions, a user classifier matching user data to user sets, identify, using the classifier and the element user data, a user set identifier matching the user, produce a selection guidance using the user set identifier and the item category identifier, and provide the selection guidance to the user.

A tunable homojunction field effect device-based artificial synapse circuit includes a first tunable homojunction field effect device M1, a second tunable homojunction field effect device M2, a third tunable homojunction field effect device M3, and a capacitor C; the tunable homojunction field effect device can exhibit the electrical properties of NN junction, PP junction, PN junction, and NP junction under the control of gate voltage; in the circuit, whether the device M2 and the device M3 are turned on rely on the combined action of presynaptic pulse and postsynaptic pulse; compared with the circuit structure of a traditional CMOS circuit scheme which exhibits neural synaptic functions of spike-time-dependent plasticity and continuously adjustable pulse-to-synaptic weight, the circuit in the present solution requires a greatly reduced number of devices and shows the feature of reconfigurable function, exhibiting a great advantage in constructing low-power, high-density integrated bionic chips for future neuromorphic applications.

In one aspect, there is provided a method performed by one or more data processing apparatus that includes obtaining a network input and processing the network input using a neural network to generate a network output that defines a prediction for the network input. The method further includes processing the network input using an encoding sub-network of the neural network to generate an embedding of the network input, processing the embedding of the network input using a brain hybridization sub-network of the neural network to generate an alternative embedding of the network input, and processing the alternative embedding of the network input using a decoding sub-network of the neural network to generate the network output that defines the prediction for the network input.

A contactless position and/or distance sensor for determining the distance, the spatial orientation, the material properties, or the like of a target object, and a method for operating the same, uses at least two sensor elements, which form a sensor module, Signals provided by the at least two sensor elements are jointly evaluated using at least one artificial neural network.

A decision support tool for development of drugs targeting central nervous system conditions. The tool receives measurements made on subjects, which are converted to model outputs using neurocircuitry models. The models are used by a computing device to generate neuro-circuitry based signatures. Neuro-circuitry based signatures associated with an investigational compound may be compared to reference neuro-circuitry based signatures to identify parameters of a clinical trial protocol. The neuro-circuitry based signature comparisons, when generated based on measurement data collected in early phases of a clinical trial process, may increase the likelihood that the investigational compound will quickly and cost-effectively emerge from clinical trials with proof that the investigational compound is effective for treating one or more CNS conditions. The decision support tool may also indicate early phase measurements to make based on a condition against which an investigational compound is theorized to be effective against.