Petuum Symphony brings a structured and industralized process for enterprises to create AI solutions. With Symphony, enterprises can quickly turn raw data in various forms stored across siloed hardware into fully operationalized AI solutions without coding.
At the core, Symphony provides a gallery of customizable building blocks for every step, from data processing, machine learning, to deployment. Our intuitive design tools make it possible to graphically assemble these building blocks into operational AI solutions without coding, while our library of industry specific reference designs and training courses help any enterprise to get started quickly.
A key component of Symphony is its distributed Operating System. It supports AI deployment on any hardware such as laptop, datacenter, cloud, Internet of Things (IoT), and edge computing. It also allows users to build and experiment very large data at high speed using commodity computers.
The Petuum healthcare software system, PetuumMed, provides end-to-end artificial intelligence functions needed in healthcare applications to augment clinical practice and increase patient engagement. PetuumMed offers immediate access to accurate and actionable insights and streamlines workflow to ease EHR-related physician burnout. With state-of-the-art natural language processing (NLP) and computer vision algorithms such as supervised learning, deep learning, unsupervised learning, adversarial learning, and probabilistic graphical models, PetuumMed enables real-time prediction and point-of-care assistance across the care continuum.
PetuumMed offers a robust information distillation pipeline that standardizes and transforms multi-modal, large-scale, unstructured healthcare data into canonical, compact and informative representations. These representations allow PetuumMed to effectively extract critical information from lengthy blocks of free text, populate designated areas in medical charts and generate insurance codes, and write abstractive summaries of past medical histories or imaging reports, to reduce physicians' administrative burden with EHRs. PetuumMed can also provide second opinions on disease diagnoses and treatment options based on clinical notes or medical images such as x-rays, CT scan, and others, thus facilitating medical reconciliation and reducing medical errors.
By using industrial-strength AI design processes and standardized building blocks, PetuumMed achieves state-of-the-art accuracy and performance over solutions built by expert AI teams, as measured across a comprehensive suite of benchmarks including Medical Report Generation, Critical Information Extraction, Disease Classification and Treatment Recommendation.
The Petuum Manufacturing and IoT software system, PetuumIoT, provides end-to-end artificial intelligence functions needed by users at various levels, across manufacturing, energy, process, discrete & utilities industry verticals. PetuumIoT provides sophisticated ML models such as supervised learning, deep learning, unsupervised learning, transfer learning, reinforcement learning, and Bayesian learning that best fit real-world industrial IoT scenarios – ranging from anomaly detection and process optimization to rare event forecast and human-machine interactions. The PetuumIoT system harnesses data generated by billions of IoT-enabled devices in real-time, including unstructured data in the form of text, log files, manual entry, graphs, and audio or video feeds, to create practical AI solutions that increase productivity, reduce downtime and enable smart logistical and intelligent field service.
PetuumIoT allows data scientists, and knowledge engineers in asset-intensive enterprises like Oil & Gas, Metals & Mining, Chemicals and Utilities to quickly ingest data, develop and deploy predictive models at scale across the entire value chain or communities to foster new service-based business models. PetuumIoT orchestrates computing resources to implement enterprise-wide optimization strategies, enabling easier regulatory compliance, more efficient operation, and more intelligent safety systems for resource management.