International Business Machines Corporation

Agency for Science, Technology and Research

Payel Das

Sufi Zafar

Robert L Bruce

Enara C Vijil

Pin-Yu Chen

Inkit Padhi

Hsinyu Tsai

Kahini Wadhawan

Aditya Vempaty

Igor Melnyk

Samuel Chung Hoffman

Srirupa Chakraborty

James L Hedrick

Yi Yan Yang

Brian Leo Quanz

Sijia Liu

Karthikeyan Shanmugam

Shiyu Chang

Pang Kern Jeremy Tan

Jae-Wook Ahn

Karthikeyan Natesan Ramamurthy

HsinYu Tsai

Cicero Nogueira Dos Santos

Pierre L Dognin

Georgios Kollias

Kevin G Carr

Jesus M Rios Aliaga

Sebastian Gehrmann

Tom D J Sercu

Flaviu Cipcigan

Ryan Hubbard

Pu Zhao

Pin-yu Chen

Minhao Cheng

Subhajit Chaudhury

Yue Cao

Hamid Dadkhahi

Elliot Nelson

A computerized system includes a generative large language model (LLM). The LLM includes an encoder configured to provide a vector input to a content addressable memory system and a decoder configured to receive a vector output from the content addressable memory system and generate a data output. The content addressable memory system is configured to receive the vector input and generate a vector output based on data contents of the content addressable memory system. The content addressable memory system is configured to respond to receiving at least one memory update command by updating the data contents of the content addressable memory system.

Dynamic updating of content addressable associative memories for large language models

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 forces 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 a region of missing residues, and NMR structural ensembles.

Embedding-based generative model for protein design

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. QMO tasks include optimizing drug-likeness and penalized log P scores with similarity constraints and improving the target binding affinity of existing drugs to pathogens such as the SARS-CoV-2 main protease protein while preserving the desired drug properties. QMO tasks further improves optimizing antimicrobial peptides toward lower toxicity.

Query-based molecule optimization and applications to functional molecule discovery

Aspects of the present invention disclose a method, computer program product, and system for optimizing a result for a combinatorial optimization problem. The method includes one or more processors receiving a black-box model. The method further includes one or more processors learning a multilinear polynomial surrogate model employing an exponential weight update rule. The method further includes one or more processors optimizing the learnt multilinear polynomial surrogate model. The method further includes one or more processors applying the black-box model to the optimized solution found by the multilinear polynomial surrogate model. In an additional aspect, the method of learning an optimized multilinear polynomial surrogate model employing an exponential weight update rule further includes one or more processors calculating utilizing data from the black-box model, an update of the coefficients of the multilinear polynomial surrogate model.

Combinatorial black box optimization with expert advice

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.

Transfer learning for molecular structure generation

Systems, computer-implemented methods, and computer program products to facilitate a dual learning bridge between text and a knowledge graph are provided. According to an embodiment, a system can comprise a processor that executes computer executable components stored in memory. The computer executable components comprise a model component that employs a model to learn associations between text data and a knowledge graph. The computer executable components further comprise a translation component that uses the model to bidirectionally translate second text data and one or more knowledge graph paths based on the associations.

Dual learning bridge between text and knowledge graph

Embodiments relate to a system, program product, and method for inducing creativity in an artificial neural network (ANN) having an encoder and decoder. Neurons are automatically selected and manipulated from one or more layers of the encoder. An encoded vector is sampled for an encoded image. Decoder neurons and a corresponding activation pattern are evaluated with respect to the encoded image. The decoder neurons that correspond to the activation pattern are selected, and an activation setting of the selected decoder neurons is changed. One or more novel data instances are automatically generated from an original latent space of the selectively changed decoder neurons.

Inducing creativity in an artificial neural network

Techniques for sanitization of machine learning (ML) models are provided. A first ML model is received, along with clean training data. A path is trained between the first ML model and a second ML model using the clean training data. A sanitized ML model is generated based on at least one point on the trained path. One or more ML functionalities are then facilitated using the sanitized ML model.

Mitigating adversarial effects in machine learning systems

A computer-implemented method according to one aspect includes training a latent variable model (LVM), utilizing labeled data and unlabeled data within a data set; training a classifier, utilizing the labeled data and associated labels within the data set; and generating new data having a predetermined set of labels, utilizing the trained LVM and the trained classifier.

Generating attribute-based samples

Techniques for training robust machine learning models for adversarial input data. Training data for a machine learning (ML) model is received. The training data includes a plurality of labels for data elements. First modified training data is generated by modifying one or more of the plurality of labels in the training data using parameterized label smoothing with a first optimization parameter. The ML model is trained using the first modified training data. The training includes updating a first one or more model weights in the ML model, and generating a second optimization parameter suitable for use in future parameterized label smoothing for future training of the ML model

Training robust machine learning models

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

Artificial intelligence designed antimicrobial peptides

Sensory input data signal is communicated to an AI platform and translated into a corresponding scenario. The translation is directed at an associated force and application of the force to a selected or identified object and environment. Real-time analysis of the force is applied, which includes modeling an expected behavior. An assessment response is received and compared to a corpus. A solution in the corpus proximal to the assessment response is identified, and a reaction of proximity of the response to the identified solution is created. Proximity data is converted to a sensory output signal. Receipt of the sensory output signal by a corresponding sensory output device creates a physical manifestation of generated feedback of the reaction data to a sensory medium. Embodiments are directed at both learning and assessment, and leverage a corpus with the AI platform in support of an interactive learning experience.

Interactive feedback and assessment experience

Embodiments relate to a system, program product, and method for use with an intelligent (AI) computer platform to explore real-world modeling. A sensory input data signal is communicated to the AI platform and translated into a corresponding scenario. The translation is directed at an associated force and application of the force to a selected or identified object and environment. A reaction to the application is created, and reaction data is converted to a sensory output signal. Receipt of the sensory output signal by a corresponding sensory output device creates a physical manifestation of generated feedback of the reaction data to a sensory medium. The embodiments are directed at both learning and assessment, and leverage a corpus with the AI platform in support of an interactive learning experience.

Interactive learning experience

Techniques that facilitate monitoring learning performance using neurofeedback are described. In one embodiment, a system is provided that comprises a memory that stores computer-executable components and a processor that executes computer-executable components stored in the memory. In one implementation, the computer-executable components comprise a feedback component that receives first feedback information regarding mental function of a user in association with participation in a learning experience, wherein the first feedback information is captured via a NIRS spectroscopy sensor worn by the user. The computer-executable components further comprise an assessment component that determines learning performance information for the user based on the first feedback information, wherein the learning performance information reflects the user's learning of content presented to the user in the learning experience, and a notification component that generates a notification based on the learning performance information indicating the user's learning of the content is below a defined learning performance level.

Monitoring learning performance using neurofeedback

The present invention relates generally to the field of microelectronics, and more particularly to a structure and method of forming a biosensor having a nucleotide attracting surface tailored to reduce false detection of nucleotides and enabling optical detection of nucleotides. The biosensor may include an analyte-affinity layer on an upper surface of a dielectric layer. The analyte-affinity layer may include a plurality of cylindrical gold portions with dimensions tailored for a target analyte. A distance between adjacent portions of the plurality of portions may range from approximately 50% of a length of a target analyte to approximately 300% of a length of a target analyte. The plurality of portions of the analyte-affinity layer have an upper surface with a diameter ranging from approximately 3 nm to approximately 20 nm.

Growing community of inventors

Yorktown Heights, NY, United States of America

Payel Das