Neural Network Architecture


IANN leverages a pre-trained healthcare auto-generated corpus and transfer learning to provide semantic meaning to healthcare terminology



IANN’s deep learning based Natural Language Processing identifies information from text and narratives using patterns similar to humans

Unlike traditional NLP IANN  functions more like

the human brain leveraging true semantic understanding of terms




IANN at work in Radiology

Auto-Impression Assistant IANN has been trained to understand reporting to the extent that it can improve productivity

and quality of radiology reports

Radiology Protocol Assistant IANN can automate the process of protocoling an upcoming exam

reducing the need for manual protocoling

Radiology Analytic Assistant 

IANN has been trained to structure reports, allowing users to perform analytics on vast volumes of unstructured data



Inference Analytics Offers New Approach to Covid-19 Searches:

Standard Search A lexical search engine employing our word embedding model, and using our narrative cloud, that uses synonyms and related terminology to provide targeted reference materials

Research Search This is Semantic Search and is based on our proprietary ability to conceptualize text. This allows data searches, such as "Are there geographic variations in the rate of COVID-19 spread?". Results highlight the articles that relate to this question, without requiring specific, structured search terms



IANN at work with Clinical Notes

Readmission Risk Assistant IANN has been trained to highlight the risk of readmission based on EMR data

  • IANN has been trained to understand narratives and unstructured information in notes in order to predict the risk of readmission with greater than 80% accuracy 

  • Application of this risk prediction algorithm during hospitalization guides optimal care and planning

  • Readmission of a patient within 30 days of discharge can result in penalties to the care provider



IANN at work in Revenue

IANN has been trained to work within the revenue cycle management processes

Identifies appropriate ICD-10 codes by analyzing EMR information

Facilitates  the revenue cycle process, reducing claims rejections, need for resubmission and lag time until payment

Enhances compliance with federal billing regulations