SmartTensors AI is a suit of new Machine Learning (ML) methods and software that utilize artificial intelegence (AI) solutions for various problems. Our platform is a winner in IT category and bronze medalist in the Market Disruptor category in 2021 R&D100 World Competition. Our framework delivers accurate and precise analysis of extra-large (exabyte) datasets using high-performance computing (HPC) and emerging GPU/TPUs technologies, as well as analysis of datasets with limited number of samples. Our AI solutions encompass exploratory data analysis where we can identify essential robust hidden patterns and model data for downstream classification and characterization, mining, and mining large-scale text data in various formats, as well as deliver innovative solutions and novel information of the uncertainty and sensitivity of computer-generated data.
Our platform build the foundation to develop advance AI solutions: SLIC AI, and ALERTs AI.
SLIC AI
SLIC AI is an exascale ML tool ready to analyze and extract unique information from various text documents, such as, scientific publications, e-mails, reports, and their metadata. SLIC AI has also an interactive regime, allowing human (domain expert) on the loop, for a better accuracy and specificity.
- SLIC AI performs robust unsupervised learning (it does not require training or labeled data) of arbitrary text corpus and extract the topics and subtopics in a hierarchical manner, while considering the semantic of the text.
- SLIC AI uses a LANL patent to determine the number of topics, which is vital for explainability.
- SLIC AI is an HPC tool that can analyze exascale data (sparse or dense) with a unique scaling on heterogenous CPU/GPUs clusters.
- SLIC AI can build unique and specific corpora and knowledge graphs through SMEs interactions human on the loop).
- SLIC AI can rank authors/institutions (based on their research on a specific topic), using their network interactions - e.g., co-authoring, co-citations data, etc.
- SLIC AI can determines the roles of the authors, such as, brain, working bee, mediator, and others, based on graph centrality.
- SLIC AI can build and analyze scientific ecosystems of a) country, or b) institution, or c) group of authors.
- SLIC AI can build a temporal, topic specific, authors profile, which includes their social scientific interactions (such as, citation and co-authors networks), as well as the evolution of their professional affiliations.
- SLIC AI can determine changes/evolution of a specific technology trend of interest, related to a country, or institution, or a group of authors.
ALERTs AI
ALERTs AI is a tool for Fingerprinting and Detecting of Various Anomalies.
- ALERTs AI classifies different types of multimodal data via tensor dimension reduction based on the extraction and identification of the hidden/features in the data.
- For each specific subject, ALERTs AI builds an archive of latent multi-modal signatures/fingerprints, whose combinations describe and characterize the data. The archive then is ready to be used for a rapid/real-time identification, characterization, and classification of new incoming data.
- ALERTs AI can predict “This is not a threat”, “This is a known threat!”, and “I do not know what this is!”, using both limited in quantity of labelled samples, and extra-large datasets.
- ALERTs AI classifies known threats and can separate them from benign activities.
- ALERTs AI can identify new types of anomalies based on its rejection option.
- ALERTs AI can be used in spectrum of anomaly classification and detection, and for novel threat detection problems e.g., in the fields of cyber-security, nuclear incident response, device/hardware/vehicles fingerprinting, and in various other problems.
- ALERTs AI is a subject of a LANL provisional patent.