Eltron Research, Inc.

Amazon Technologies, Inc.

Google Inc.

Itn Energy Systems, Inc.

Michael F Schwartz

Anthony F Sammells

James H White

Vikram Kumar Gundeti

Rongzhou Shen

Bo Li

Leopold Bushkin

Yaser Khan

Munir Mahmood

Alexander Thomas Loeb

Shemyla Anwar

Mohammed Arif

Edward Page Foyle

Joe Pemberton

Brian Spencer Berland

Alexander Say Go

Richard Mackay

Archit Jain

Joseph H Armstrong

Paul E Newson

David Erb

Erkki Ville Juhani Aikas

Lin Jay Simpson

Michael T Carter

Vijitha Raji

Steven Mack Saunders

Ronald William Schaller

Sabina Gade

William G Barker

Anthony F Sammels

Tara Raj

Techniques for a natural language processing (NLP) system to implement more than one assistant are described. The NLP system may receive a natural language input corresponding to more than one user command. The NLP system may respond to a first command, of the natural language input, using a TTS voice of a first NLP system assistant. The NLP system may respond to a second command, of the natural language input, using a TTS voice of a second NLP system assistant.

Multi-assistant natural language input processing

Techniques for selecting a skill component to process a natural language input are described. When a natural language input is received, natural language understanding (NLU) output data representing the natural language input is generated, and skill components (capable of processing the NLU output data) are determined. Thereafter, rules (for preventing the invocation of skill components) are implemented in a tiered manner, resulting in the determination of a subset of the skill components. The subset of skill components is ranked using a machine learning model(s), and the top-ranked skill component is called to process the NLU output data.

Natural language input routing

Techniques for performing runtime ranking of skill components are described. A skill developer may generate a rule indicating a skill component is to be invoked at runtime when a natural language input corresponds to a specific context. At runtime, a virtual assistant system may implement a machine learned model to generate an initial ranking of skill components. Thereafter, the virtual assistant system may use skill component-specific rules to adjust the initial ranking, and this second ranking is used to determine which skill component to invoke to respond to the natural language input. Overtime, if a rule results in beneficial user experiences, the virtual assistant system may incorporate the rule into the machine learned model.

Techniques for a natural language processing (NLP) system to implement more than one assistant are described. The NLP system may receive a natural language input from a device. The NLP system may also receive one or more signals representing one or more assistants to be implemented with respect to the natural language input. The NLP system may intelligently select an assistant to be invoked with respect to the natural language input. Once the assistant is selected, the NLP system may cause content, output to a user, to have characteristics specific to the assistant.

Multi-assistant natural language input processing to determine a voice model for synthesized speech

A hosted storage system includes a namespace that contains object identifiers. A first object containing a first data component is stored. A first object identifier in the namespace uniquely identifies the first object. A second object containing a second data component is stored. A second object identifier in the namespace uniquely identifies the second object. A compose object request that designating the first object identifier, the second object identifier, and a third object identifier in the namespace is received. A third object that includes the first data component and the second data component without duplicating the first data component and the second data component is created. The third object identifier is associated with the third object such that the third object identifier uniquely identifies the third object.

Composing objects in hosted storage

In one general aspect, this document describes a computer-implemented method for creating buckets in a hosted storage platform. The method may include receiving, at a hosted storage computer system that utilizes a single namespace across data stores of the hosted storage computer system, a request from a user of the hosted storage computer system to create a bucket having a name that includes a domain name. The method may also include verifying that the user is an authorized manager of a domain that corresponds to the domain name. The method may also include creating the bucket on the hosted storage computer system upon verification that the user is an authorized manager of the domain.

Domain name buckets in a hosted storage system

An initial portion of an object is received at a hosted storage system. The initial portion of the object is stored at the hosted storage system. After receiving the initial portion of the object, a timer is started. Prior to the expiration of the timer, whether an additional portion of the object is received at the hosted storage system is determined. The initial portion is deleted upon the expiration of the timer if the additional portion is not received. The timer is reset if the additional portion is received.

Resumable upload for hosted storage systems

Materials for use in proton transport characterized by several formulas are disclosed. Mixed ion and electron conductors may include metals and/or ceramic electron conductors and a proton conducting material. Hydrogen separation membranes may include porous layers and an electrolyte layer including a proton conducting material and an electron conductor. Hydrogen separation membranes may be formed by thermal spray techniques. Hydrogen separation membranes may include a catalyst layer. A method of separating hydrogen from a mixed gas stream includes passing the mixed gas through a first porous layer to an electrolyte layer, dissociating protons and electrons, diffusing the protons and electrons through the electrolyte layer, recombining them, and passing molecular hydrogen through a second porous layer.

Proton conducting materials and devices incorporating them

This invention relates to gas-impermeable, solid state materials fabricated into membranes for use in catalytic membrane reactors. This invention particularly relates to solid state oxygen anion- and electron-mediating membranes for use in catalytic membrane reactors for promoting partial or full oxidation of different chemical species, for decomposition of oxygen-containing species, and for separation of oxygen from other gases. Solid state materials for use in the membranes of this invention include mixed metal oxide compounds having the brownmillerite crystal structure.

Solid state oxygen anion and electron mediating membrane and catalytic membrane reactors containing them

The present invention relates to the design and manufacture of single cell units for planar, thin-film, ceramic electrochemical devices such as solid oxide fuel cells, electrochemical oxygen generators, gas separation membranes, and membrane modules and stacks and the fabrication of multi-cell stacks and modules of the single cell units. The design is based upon a single cell wherein manufacturing all layers of the device into an integral unit produces a monolithic structure. The design produces a gas-tight single cell that is easily assembled into multi-cell stacks and modules without external seals or sealing mechanisms. The design may use standard ceramic and metallurgical production techniques. The design of the present invention enhances device performance since the single cell units are inherently sealed for gas tightness and have reduced interfacial electrical resistances. All these features of the novel monolithically integrated unit cell design result in lower manufacturing costs for ceramic electrochemical devices.

Apparatus and method for the design and manufacture of thin-film electrochemical devices

Materials and methods for the separation of oxygen from air

This invention relates to catalytic reactor membranes having a gas-impermeable membrane for transport of oxygen anions. The membrane has an oxidation surface and a reduction surface. The membrane is coated on its oxidation surface with an adherent catalyst layer and is optionally coated on its reduction surface with a catalyst that promotes reduction of an oxygen-containing species (e.g., O , NO , SO , etc.) to generate oxygen anions on the membrane. The reactor has an oxidation zone and a reduction zone separated by the membrane. A component of an oxygen containing gas in the reduction zone is reduced at the membrane and a reduced species in a reactant gas in the oxidation zone of the reactor is oxidized. The reactor optionally contains a three-dimensional catalyst in the oxidation zone. The adherent catalyst layer and the three-dimensional catalyst are selected to promote a desired oxidation reaction, particularly a partial oxidation of a hydrocarbon.

Two component-three dimensional catalysis

This invention provides a fiber optic sensor and sensor devices for the reversible detection of basic gas analytes, including hydrazine, alkyl hydrazines, amines, ammonia, and related chemical species. The optical fiber sensor is formed in an optical fiber which conventionally comprises an optical fiber core and a cladding layer. Transducer molecules are immobilized in contact with an exposed surface of the fiber core to allow interaction of the immobilized species with analytes in the environment around the sensor. Preferred transducer molecules for detection of basic gases, particularly hydrazine and hydrazine derivatives, are xanthene dyes and triphenylmethane dyes, including malachite green and crystal violet. The sensors of this invention can be employed in a variety of device configurations including single-site sensors and multiple-site sensor networks.

Growing community of inventors

Boulder, CO, United States of America

Michael F Schwartz