Healthcare AI

University of Pittsburgh researchers have developed PCAP–Backdoor, software capable of attacking packet capture (PCAP) datasets through a “backdoor,” bypassing deep learning (DL)-based intrusion detection systems (IDS).

This invention is a multi-agent retrieval-augmented large language model (LLM) architecture that classifies free-form text comments into precise operational categories and subcategories, providing an auditable reasoning for its decision.

Two built-in ethical checkpoints aligned with four core pillars of AI ethics and professional codes ensure that client resources and treatment plans produced with generative AI maintain safety, reliability and professionalism.

It stands out by integrating ethical oversight directly into both prompt design and output review, a feature often missing in general AI guidelines.

University of Pittsburgh research has developed Perepex, an AI-powered educational platform designed to facilitate video-based learning and assessment between teachers and students.

By leveraging AI for real-time transcription, cognitive analysis, and personalized feedback, Perepex aims to improve educational outcomes and streamline the assessment process.

Perepex is a software platform that utilizes AI to enhance video-based learning and assessment.

International Journal of Artificial Intelligence in Education, 35(2), 879–913.

- Precision Behavioral Health Analytics: Integration into platforms and software focused on improving patient outcomes through real-world data and artificial intelligence.

NeuroSmart is an AI-driven clinical decision support system (CDSS) designed to optimize stroke management by providing personalized, real-time diagnostic and treatment recommendations.

js, allows clinicians to input patient data and receive AI-generated recommendations with confidence scores and visual analytics.

- Enhances clinical decision-making with AI-driven insights tailored to patient-specific factors.

Utilizing artificial intelligence (AI), SLAID analyzes source material, generates customized presentation outlines based on specified contexts, and evaluates recorded presentations for alignment with the outline and cognitive demand levels.

SLAID is a software tool that leverages AI and Natural Language Processing (NLP) to assist users in creating and analyzing presentations.

Key features include document upload/search, AI-powered document analysis, contextual presentation outline generation, an editable checklist, integrated recording tools, AI-driven presentation analysis, feedback and reporting, and secure video handling.

This invention is a computational technology utilizing a machine learning model to accurately predict single-cell types from standard hematoxylin and eosin (H&E) stained digital pathology images.

• AI safety and efficacy assessments.

As the integration of generative AI into healthcare advances, it is crucial to ensure the safety and efficacy of these models through rigorous human assessments.

This invention is a novel software solution that processes in-situ infrared (IR) camera data during laser powder bed fusion (LPBF) to identify and locate manufacturing defects.

It is increasingly used in patient management, clinical research, and the development of artificial intelligence (AI) tools to support clinical decision making.

Novel software has been developed to collect data from existing EHR databases for use in Azure Healthcare API with the potential to improve healthcare efficiency, clinical research, and AI tool development using large existing databases.

• AI tool development using Azure Healthcare API.

This technology comprises a specialized codebook designed to annotate clinical data associated with an ICU rounding checklist based on the ABCEDF bundle.

AI-dBPE correlated inversely with age, suggesting AI-dBPE may be a suitable, quantifiable measure of BPE classification.

Designed using an AI-based model to segment images of breast tissue into fibroglandular or fatty tissue, this method can quantify BPE, an indicator of blood flow in the breast tissue.

Using low energy with recombined images and AI segmentation of fibroglandular and fatty tissue, overlay of these segments onto corresponding RI images can standardize the measurement of BPE.

University of Pittsburgh researcher has developed spatiAlytica, an AI-powered system designed for the seamless analysis and visualization of single-cell or spot-based spatial transcriptomics data.