Abstract: Embodiments of the invention provide a computer-implemented method that includes applying input representations of a three-dimensional (3D) domain to a generative neural network (GNN); and using the GNN to form a generative model of the 3D domain based at least in part on the input representations. The input representations include a global-shape input representation of the 3D domain.

Abstract: One embodiment of the invention provides a method of training a chatbot. The method comprises identifying one or more chat logs that exhibit a trait. The method further comprises identifying one or more labels associated with the trait based on the one or more chat logs. The method further comprises training the chatbot to generate a response that models the trait based on the one or more chat logs. The method further comprises labeling the chatbot with the one or more labels.

Abstract: Provided are systems, methods, and media for handling dialogs based on user behavior data. An example method includes receiving an input paragraph having one or more factual sentences, in which each of the one or more factual sentences includes one or more words. Receiving an input question comprising one or more words. Performing word-level gaze prediction on the input paragraph to identify one or more predicted gaze attributes for the input paragraph. Extracting an answer to the input question based, at least in part, on the input paragraph, the input question, and the one or more predicted gaze attributes of the input paragraph. Transmitting the extracted answer.

Abstract: A system and method for designing protein sequences conditioned on a specific target fold. The system is a transformer-based generative framework for modeling a complex sequence-structure relationship. To mitigate the heterogeneity between the sequence domain and the fold domain, a Fold-to-Sequence model jointly learns a sequence embedding using a transformer and a fold embedding from the density of secondary structural elements in 3D voxels. The joint sequence-fold representation through novel intra-domain and cross-domain losses with an intra-domain loss forcing two semantically similar (where the proteins should have the same fold(s)) samples from the same domain to be close to each other in a latent space, while a cross-domain loss forces two semantically similar samples in different domains to be closer. In an embodiment, the Fold-to-Sequence model performs design tasks that include low resolution structures, structures with region of missing residues, and NMR structural ensembles.

Abstract: Generating a molecule design by training a binding affinity model using a first molecular database and an embedding of a second molecular database and generating a molecule design according to the embedding and the binding affinity model.

Abstract: Generating a drug molecule design by training an attribute predictor model using an embedding of a first molecular data base, training a first machine learning model using the attribute predictor, yielding a second embedding of the first molecular data base, training a binding affinity model using a second molecular database and the second embedding of the first molecular database, and generating a molecule design according to the second embedding and the binding affinity model.

Abstract: Various embodiments select advertisements for delivery to at least one mass media streaming device at a given location and time slot. Content and corresponding time and physical location information being posted in an online portal from at least one electronic device are obtained. The content is filtered based on time slot and physical location. A set of typical topics for a location and time slot are determined. A subset of advertisements is selected from a set of available advertisements based on the set of typical topics. The subset of advertisements is delivered to the at least one mass media streaming device. The mass media streaming device may be a television. A set of current topics for the location and time slot may be determined and aggregated with the set of typical topics. The subset of advertisements may be selected from the set of available advertisements based on the aggregated set.

Abstract: Metadata about a movie is retrieved. The metadata includes a plurality of associated viewer responses from at least one previous audience viewing of the video. The plurality of associated viewer responses from the at least one previous audience viewing are associated with one or more segments of the video. A segment of the video associated with a type of viewer reaction based on emotion and sentiment recognition is identified. Additional media content based on the identified video segment is retrieved. A segment of the additional media content that exceeds a threshold of similarity with the segment of the video is determined. A video clip that includes the segment of the additional media content is created.

Abstract: A query-based generic end-to-end molecular optimization (“QMO”) system framework, method and computer program product for optimizing molecules, such as for accelerating drug discovery. The QMO framework decouples representation learning and guided search and applies to any plug-in encoder-decoder with continuous latent representations. QMO framework directly incorporates evaluations based on chemical modeling, analysis packages, and pre-trained machine-learned prediction models for efficient molecule optimization using a query-based guided search method based on zeroth order optimization. The QMO features efficient guided search with molecular property evaluations and constraints obtained using the predictive models and chemical modeling and analysis packages.

Abstract: De novo, artificial intelligence (AI) designed antimicrobial peptides (AMPs), antibacterial products comprising the AMPs and methods for treating bacterial infections using the products are provided. In one or more embodiments, the AMPs were designed using conditional latent attribute space sampling (CLaSS). The AMPs comprise up to twenty natural amino acids in length, including one with twelve and another with thirteen natural amino acids in length. The AMPs demonstrate low-toxicity and show high antimicrobial potency against diverse pathogens including multi-medication-resistant Gram negative Klebsiella pneumoniae.

Abstract: Techniques regarding generating molecular structures with attributes of interest are provided. For example, one or more embodiments described herein can comprise a system, which can comprise a memory that can store computer executable components. The system can also comprise a processor, operably coupled to the memory, and that can execute the computer executable components stored in the memory. The computer executable components can comprise a transfer learning component that determines a molecular structure of a compound by employing a transfer learning process that utilizes lessons learned from an unconditional generative machine learning model to train a conditional machine learning model that regards a target attribute profile.

Abstract: Techniques for filtering artificial intelligence (AI)-designed molecules for laboratory testing provided. According to an embodiment, computer implemented method can comprise selecting, by a system operatively coupled to a processor, a first subset of AI-designed molecules from a set of AI-designed molecules as candidate pharmaceutical agents based on classification of the AI-designed molecules using one or more classifiers. The method further comprises selecting, by the system, a second subset of the candidate pharmaceutical agents for wet laboratory testing based on evaluation of molecular interactions between the candidate pharmaceutical agents and one or more biological targets using one or more computer simulations.

Abstract: De novo, artificial intelligence (AI) designed antimicrobial peptides (AMPs), antibacterial products comprising the AMPs and methods for treating bacterial infections using the products are provided. In one or more embodiments, the AMPs were designed using conditional latent attribute space sampling (CLaSS). The AMPs comprise up to twenty natural amino acids in length, including one with twelve and another with thirteen natural amino acids in length. The AMPs demonstrate low-toxicity and show high antimicrobial potency against diverse pathogens including multi-medication-resistant Gram negative Klebsiella pneumoniae.

Abstract: De novo, artificial intelligence (AI) designed antimicrobial peptides (AMPs), antibacterial products comprising the AMPs and methods for treating bacterial infections using the products are provided. In one or more embodiments, the AMPs were designed using conditional latent attribute space sampling (CLaSS). The AMPs comprise up to twenty natural amino acids in length, including one with twelve and another with thirteen natural amino acids in length. The AMPs demonstrate low-toxicity and show high antimicrobial potency against diverse pathogens including multi-medication-resistant Gram negative Klebsiella pneumoniae.

Abstract: A contextual zoom control method, system, and computer program product, includes in a single individual capturing mode, performing a camera function to assist the user in performing an image capturing action on an identity that is paired with a highest ranked individual on a ranked list within a set of key individuals attending an event, where the camera function includes at least one of a zoom function and a focus function and where the set of key individuals attending the event are determined by any of data mining a context of the event, a search engine, a social media source, and using natural language processing on a published program of the event.

Abstract: A method, computer system, and computer program product for sport training on an augmented reality device or a virtual reality device is provided. The embodiment may include capturing a plurality of user movement data using one or more sensors. The embodiment may also include measuring a user body and eye gaze position based on the plurality of captured user movement data. The embodiment may further include calculating a body position difference by comparing the measured user body and eye gaze position with an expert-specified body position sequence. The embodiment may also include determining a body position quality threshold is not satisfied based on the calculated body position difference. The embodiment may further include generating an instruction based on the compared calculated body position difference.

Abstract: According to an embodiment of the present invention, a video is generated from a communication session pertaining to support for use of a user interface. A scenario is identified within a document including content of the communication session. One or more items corresponding to the identified scenario and associated with the user interface are extracted from the document. The extracted items are mapped to corresponding aspects of the user interface, wherein at least one of the extracted items remains unmapped to the user interface. At least one question is determined for a user to receive information to map an unmapped item to a corresponding aspect of the user interface. A video is generated based on the mapped aspects of the user interface to reproduce one or more activities performed during the use of the user interface.

Abstract: In an approach to generating HTML based on a plurality of content and design controls, one or more computer processors crawl one or more conforming websites. The one or more computer processors create a training set of crawled conforming webpages, wherein each crawled conforming webpage in the training set of crawled conforming webpages includes associated web code labeled a combination of respective one or more content controls and respective one or more design controls; encode the combination of the respective one or more design controls and the respective one or more content controls based on one or more user preferences utilizing a created design encoder and a created content encoder; create a decoder with the training set of crawled conforming webpages and associated encoded content and design vectors; generate web code based on the encoded design and content controls utilizing the created decoder; implement generated web code.

Abstract: One embodiment of the invention provides a method of training a chatbot. The method comprises identifying one or more chat logs that exhibit a trait. The method further comprises identifying one or more labels associated with the trait based on the one or more chat logs. The method further comprises training the chatbot to generate a response that models the trait based on the one or more chat logs. The method further comprises labeling the chatbot with the one or more labels.

Abstract: In an approach to generating HTML based on a plurality of content and design controls, one or more computer processors crawl one or more conforming websites. The one or more computer processors create a training set of crawled conforming webpages, wherein each crawled conforming webpage in the training set of crawled conforming webpages includes associated web code labeled a combination of respective one or more content controls and respective one or more design controls; encode the combination of the respective one or more design controls and the respective one or more content controls based on one or more user preferences utilizing a created design encoder and a created content encoder; create a decoder with the training set of crawled conforming webpages and associated encoded content and design vectors; generate web code based on the encoded design and content controls utilizing the created decoder; implement generated web code.