Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an action selection neural network used to select actions to be performed by an agent interacting with an environment. In one aspect, a system comprises a plurality of actor computing units and a plurality of learner computing units. The actor computing units generate experience tuple trajectories that are used by the learner computing units to update learner action selection neural network parameters using a reinforcement learning technique. The reinforcement learning technique may be an off-policy actor critic reinforcement learning technique.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an action selection neural network used to select actions to be performed by an agent interacting with an environment. In one aspect, a system comprises a plurality of actor computing units and a plurality of learner computing units. The actor computing units generate experience tuple trajectories that are used by the learner computing units to update learner action selection neural network parameters using a reinforcement learning technique. The reinforcement learning technique may be an off-policy actor critic reinforcement learning technique.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for reinforcement learning. One of the methods includes selecting an action to be performed by the agent using both a slow updating recurrent neural network and a fast updating recurrent neural network that receives a fast updating input that includes the hidden state of the slow updating recurrent neural network.

Abstract: A polynucleotide, comprising the following base sequences: (a) a base sequence encoding a fusion protein of a nuclease-deficient CRISPR effector protein and a transcription repressor, and (b) a base sequence encoding a guide RNA targeting a continuous region of 18 to 24 nucleotides in length in a region set forth in SEQ ID NO: 54, 55, 56, 57, 68, 153 or 97 in the expression regulatory region of human MAPT gene. are expected to be useful for treating or preventing tauopathy including Alzheimer's disease.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an action selection neural network used to select actions to be performed by an agent interacting with an environment. In one aspect, a system comprises a plurality of actor computing units and a plurality of learner computing units. The actor computing units generate experience tuple trajectories that are used by the learner computing units to update learner action selection neural network parameters using a reinforcement learning technique. The reinforcement learning technique may be an off-policy actor critic reinforcement learning technique.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an action selection neural network used to select actions to be performed by an agent interacting with an environment. In one aspect, a system comprises a plurality of actor computing units and a plurality of learner computing units. The actor computing units generate experience tuple trajectories that are used by the learner computing units to update learner action selection neural network parameters using a reinforcement learning technique. The reinforcement learning technique may be an off-policy actor critic reinforcement learning technique.

Abstract: The present invention provides a method of treating a disease associated with elevated KRAS activity or expression in a subject, comprising suppressing KRAS expression in the subject by targeting an expression regulatory region of KRAS gene using a CRISPR-Guide Nucleotide Directed Modulation (GNDM). Also, provided is a CRISPR-GNDM system for suppressing KRAS expression comprising (a) a protein selected from the group consisting of dCas9 or dCpf1, a fusion protein of dCas9 or dCpf1 and Kruppel associated box (KRAB), and (b) a guide nucleotide targeting an expression regulatory region of KRAS gene.

Abstract: Polynucleotides comprising the following base sequences: (a) a base sequence encoding a fusion protein of a nuclease-deficient CRISPR effector protein and a transcription activator, and (b) a base sequence encoding a guide RNA targeting a continuous region of 18 to 24 nucleotides in length in a region set forth in SEQ ID NO: 104, 105, 135, 141, 153, 167, or 172 in the expression regulatory region of human Utrophin gene are expected to be useful for treating or preventing DUCHENNE muscular dystrophy or BECKER muscular dystrophy.

Abstract: The present invention provides compounds of formula (I): Formula (I) compositions comprising such compounds; the use of such compounds in therapy; and methods of treating patients with such compounds; wherein A, Y, n, R1, R2A, R2B, R3 and *1 are as defined herein.

Abstract: The present invention provides a method of treating a disease associated with activated MYD88 signaling in a subject, including suppressing MYD88 expression in a subject by targeting an expression regulatory region of MYD88 gene by using a CRISPR-Guide Nucleotide Directed Modulation (GNDM), without affecting the expression of an adjacent ACAA1 gene. Also, provided is a CRISPR-GNDM system for suppressing MYD88, including expression comprising (a) a protein selected from the group consisting of dCas9 or dCpf1, a fusion protein of dCas9 or dCpf1 and Kruppel associated box (KRAB) and (b) a guide RNA (gRNA) targeting an expression regulatory region of MYD88 gene.

Abstract: The present application pertains to methods of using indole-oxadiazole compounds of Formula I to modulate cannabinoid receptor activity: I In particular diseases, disorders or conditions that benefit from modulating cannabinoid receptor activity, such as non-alcoholic steatohepatitis (NASH), non-alcoholic fatty liver disease (NAFLD), schizophrenia, bipolar disorder, psychosis, metabolic syndrome, type-2 diabetes, dyslipidaemia, obesity, eating disorders, cardiovascular diseases and disorders, and other conditions as described herein, may be treated. Also included in the present application are certain novel compounds of Formula Ia and pharmaceutical compositions comprising these compounds.

Abstract: Polynucleotides comprising the following base sequences: (a) a base sequence encoding a fusion protein of a nuclease-deficient CRISPR effector protein and a transcription activator, and (b) a base sequence encoding a guide RNA targeting a continuous region of 18 to 24 nucleotides in length in a region set forth in SEQ ID NO: 104, 105, 135, 141, 153, 167, or 172 in the expression regulatory region of human Utrophin gene are expected to be useful for treating or preventing DUCHENNE muscular dystrophy or BECKER muscular dystrophy.

Abstract: A Squalene Hopene Cyclase (SHC) isolated from Gluconobacter morbifer is provided as are variants and a method for using the G. morbifer SHC to biocatalytically convert homofarnesol to ambroxan.

Abstract: Systems and methods associated for providing haptic feedback are provided. For instance, data indicative of a feedback scheme associated with an input mechanism can be accessed. The feedback scheme includes one or more feedback points associated with a position of a moveable portion of the input mechanism. The input mechanism is configured to change positions responsive to an actuation of the input mechanism by a user. One or more positions of the moveable portion of the input mechanism can be determined as the input mechanism is actuated by the user. Operation of one or more feedback motors can be controlled to provide feedback to the user in accordance with the feedback scheme based at least in part on the determined one or more positions of the input mechanism.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for reinforcement learning. One of the methods includes selecting an action to be performed by the agent using both a slow updating recurrent neural network and a fast updating recurrent neural network that receives a fast updating input that includes the hidden state of the slow updating recurrent neural network.

Abstract: A device (10) for simulating the dispensing of a medication into the nasal passage. The device includes a device body (12), the device body defining an extending portion (14) for inserting into the nasal passage of a user. An activation mechanism (16) movable from a start position to a dispensed position can be returned to the start position by manipulation of part of the device. Manipulation may be by rotation of a main part of the device body (12) relative to the extending portion (14).

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an action selection neural network used to select actions to be performed by an agent interacting with an environment. In one aspect, a system comprises a plurality of actor computing units and a plurality of learner computing units. The actor computing units generate experience tuple trajectories that are used by the learner computing units to update learner action selection neural network parameters using a reinforcement learning technique. The reinforcement learning technique may be an off-policy actor critic reinforcement learning technique.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for reinforcement learning. One of the methods includes selecting an action to be performed by the agent using both a slow updating recurrent neural network and a fast updating recurrent neural network that receives a fast updating input that includes the hidden state of the slow updating recurrent neural network.