An AI Model Built With Healthcare in Mind
Leidos and Google have collaborated to develop a healthcare-specific AI application using Google's Medical Pathways Language Model 2 (MedPaLM2). This domain-specific model offers greater accuracy and relevance for medical questions compared to general AI models. The application aims to further reduce disability benefits approval processing time from days to potentially hours.
Key features of the AI application include:
- Retrieval Augmentation Generation (RAG) for self-updating
- Human oversight to ensure accuracy
- Built on Leidos' Framework for AI Resilience and Security (FAIRS)
- Addressing potential biases in training data
Leidos is now focusing on scaling the application for production environments, processing thousands of reports weekly. Future applications may include claims management, data interoperability, and clinical decision support.
Leidos e Google hanno collaborato per sviluppare un applicativo di intelligenza artificiale specifico per la sanità utilizzando il Modello di Linguaggio Medico di Google 2 (MedPaLM2). Questo modello specifico per il settore offre maggiore accuratezza e pertinenza per le domande mediche rispetto ai modelli generali di intelligenza artificiale. L'applicazione mira a ridurre ulteriormente i tempi di approvazione dei benefici per disabilità, passando da giorni a potenzialmente ore.
Le caratteristiche chiave dell'applicazione AI includono:
- Generazione di Augmentazione di Recupero (RAG) per auto-aggiornamento
- Supervisione umana per garantire l'accuratezza
- Basata sul Framework di Resilienza e Sicurezza per l'Intelligenza Artificiale (FAIRS) di Leidos
- Affrontare i potenziali pregiudizi nei dati di addestramento
Leidos si sta ora concentrando su scalare l'applicazione per ambienti di produzione, elaborando migliaia di rapporti settimanalmente. Le applicazioni future potrebbero includere gestione dei reclami, interoperabilità dei dati e supporto alle decisioni cliniche.
Leidos y Google han colaborado para desarrollar una aplicación de inteligencia artificial específica para la salud utilizando el Modelo de Lenguaje Médico de Google 2 (MedPaLM2). Este modelo específico para el dominio ofrece mayor precisión y relevancia para preguntas médicas en comparación con los modelos generales de IA. La aplicación tiene como objetivo reducir aún más el tiempo de procesamiento de aprobación de prestaciones por incapacidad, pasando de días a potencialmente horas.
Las características clave de la aplicación de IA incluyen:
- Generación de Aumento de Recuperación (RAG) para autoactualización
- Supervisión humana para garantizar la precisión
- Construida sobre el marco de Resiliencia y Seguridad de IA (FAIRS) de Leidos
- Abordando los posibles sesgos en los datos de entrenamiento
Leidos ahora se enfoca en escalar la aplicación para entornos de producción, procesando miles de informes semanalmente. Las aplicaciones futuras pueden incluir gestión de reclamaciones, interoperabilidad de datos y soporte para decisiones clínicas.
Leidos와 Google은 Google의 의료 경로 언어 모델 2(MedPaLM2)를 사용하여 의료 전용 AI 애플리케이션을 개발하기 위해 협력했습니다. 이 도메인 특화 모델은 일반 AI 모델에 비해 의료 질문에 대해 더 높은 정확도와 관련성을 제공합니다. 이 애플리케이션은 장애 혜택 승인 처리 시간을 며칠에서 잠재적으로 몇 시간으로 단축하는 것을 목표로 하고 있습니다.
AI 애플리케이션의 주요 기능은 다음과 같습니다:
- 자동 업데이트를 위한 검색 보강 생성(RAG)
- 정확성을 보장하기 위한 인간의 감독
- Leidos의 AI 회복력 및 보안 프레임워크(FAIRS)를 기반으로 함
- 훈련 데이터의 잠재적 편향 해결
Leidos는 이제 생산 환경을 위한 애플리케이션 확장에 집중하고 있으며, 매주 수천 개의 보고서를 처리하고 있습니다. 향후 애플리케이션에는 청구 관리, 데이터 상호 운용성 및 임상 의사 결정 지원이 포함될 수 있습니다.
Leidos et Google ont collaboré pour développer une application d'IA spécifique à la santé en utilisant le Modèle de Langage Médical 2 de Google (MedPaLM2). Ce modèle spécifique au domaine offre une précision et une pertinence accrues pour les questions médicales par rapport aux modèles d'IA généraux. L'application vise à réduire encore le temps de traitement des demandes d'indemnités d'invalidité, de plusieurs jours à potentiellement quelques heures.
Les caractéristiques clés de l'application IA comprennent :
- Génération d'Augmentation de Récupération (RAG) pour l'auto-mise à jour
- Surveillance humaine pour garantir l'exactitude
- Construit sur le Cadre de Résilience et de Sécurité de l'IA (FAIRS) de Leidos
- Traitement des biais potentiels dans les données d'entraînement
Leidos se concentre désormais sur l'échelle de l'application pour les environnements de production, traitant des milliers de rapports chaque semaine. Les applications futures pourraient inclure la gestion des demandes, l'interopérabilité des données et le soutien à la décision clinique.
Leidos und Google haben zusammengearbeitet, um eine gesundheits-spezifische KI-Anwendung zu entwickeln, die auf Googles Medical Pathways Language Model 2 (MedPaLM2) basiert. Dieses domänenspezifische Modell bietet höhere Genauigkeit und Relevanz bei medizinischen Fragen im Vergleich zu allgemeinen KI-Modellen. Das Ziel der Anwendung ist es, die Bearbeitungszeit für die Genehmigung von Erwerbsminderungsleistungen weiter zu verkürzen, von Tagen auf möglicherweise Stunden.
Zu den wichtigsten Funktionen der KI-Anwendung gehören:
- Retrieval-Augmentation-Generation (RAG) für die Selbstaktualisierung
- Menschliche Aufsicht zur Sicherstellung der Genauigkeit
- Auf Leidos' Framework für KI-Resilienz und Sicherheit (FAIRS) aufgebaut
- Vermeidung potenzieller Vorurteile in den Trainingsdaten
Leidos konzentriert sich jetzt darauf, die Anwendung zu skalieren für Produktionsumgebungen, wobei wöchentlich Tausende von Berichten verarbeitet werden. Zukünftige Anwendungen könnten Schadenmanagement, Dateninteroperabilität und klinische Entscheidungsunterstützung umfassen.
- Collaboration with Google to develop a healthcare-specific AI application
- Use of domain-specific MedPaLM2 model for improved accuracy in medical contexts
- Potential to significantly reduce disability benefits approval processing time
- Implementation of RAG for continuous model improvement without retraining
- Focus on scaling the application for high-volume production environments
- None.
The collaboration between Leidos and Google on a healthcare-specific AI application marks a significant step in the integration of AI in healthcare. By leveraging Google's MedPaLM2, a domain-specific language model, they've created a potentially game-changing tool for processing disability benefit applications.
The key advantages of this approach include:
- Enhanced accuracy and relevance in medical contexts
- Faster development time (6 weeks from concept to demo)
- Improved scalability and cost-effectiveness
- Better understanding of healthcare-specific terminology and contexts
The implementation of Retrieval Augmented Generation (RAG) is particularly noteworthy, as it allows the model to stay current with the latest medical guidelines without full retraining. This feature is important in the rapidly evolving field of healthcare.
However, the true test will be in scaling this application to handle the demanding production environment Leidos operates in, processing thousands of reports weekly and millions of pages of documents. The ability to handle multimodal data (text, images, video and audio) in the future could further enhance its capabilities and applications in healthcare.
While promising, it's important to note that this project is still in the demo phase. Its success in a full-scale production environment remains to be seen and careful monitoring will be necessary to ensure it maintains its accuracy and reliability at scale.
The development of this healthcare-specific AI model addresses several critical challenges in applying AI to healthcare:
- Accuracy: The need for extremely high accuracy in healthcare (85% is insufficient) is being addressed through domain-specific training and human oversight.
- Bias: By acknowledging historical biases in medical data and actively working to prevent their propagation, the model aims for more equitable outcomes.
- Security and Reliability: The implementation of Leidos' Framework for AI Resilience and Security (FAIRS) demonstrates a commitment to safe and ethical AI deployment.
- Trust: Keeping humans in the loop for verification and validation builds trust in the AI's outputs.
The potential applications of this technology are vast, including:
- Streamlining claims management
- Addressing data interoperability issues
- Providing clinical decision support
However, it's important to maintain a balanced perspective. While the initial results are promising, the healthcare industry has seen many technological solutions fail to live up to their initial hype. The true test will be in widespread, real-world implementation and the ability to consistently deliver accurate, unbiased results while maintaining patient privacy and data security.
As this technology develops, it will be essential to monitor its performance closely, ensure ongoing regulatory compliance and continuously evaluate its impact on patient outcomes and healthcare efficiency.
NORTHAMPTON, MA / ACCESSWIRE / July 25, 2024 / Healthcare has offered some of the most promising opportunities to leverage AI, from streamlining workflows to automating routine tasks to assisting in clinical decisions. At Leidos, we have been using large language models to support our health customers even before ChatGPT made it popular in the fall of 2022. Continued interest in healthcare AI has only climbed since the relatively recent availability of generative AI (genAI) models that can process and create text, speech, images, videos, and more.
But the excitement has also been tempered by the daunting challenges of applying AI to the complex, low-margin-for-error world of healthcare. Questions have arisen about security, bias, and the level of trust that can be placed in AI's output.
Srini Iyer, senior vice president and chief technology officer for the Leidos Health & Civil Sector, and Ken Su, Google's senior outbound product manager in the CloudAI for Healthcare Group, recently sat down to discuss these opportunities and challenges for healthcare AI. The two groups have been collaborating on a leading-edge healthcare AI application, enlisting a key strategy: building the application around a domain-specific genAI model specially developed for healthcare rather than a general model.
Domain-specific vs. foundational
Iyer described how Leidos wanted to focus on improving an existing Leidos application. That application, currently based of conventional AI tools, has shortened the process of obtaining approval for disability benefits significantly from months to days.
"But we wanted to compress the approval processing time it takes even further using genAI," said Iyer. With this aim in mind, Leidos engaged Su's team as collaborators. They decided to build the new application around Google's Medical Pathways Language Model 2, or MedPaLM2-a healthcare-specific domain model.
Both the collaboration and choice of model proved to be successful. "In a span of about six weeks, we were able to go from a concept to a demo version of a full-stack genAI application," Iyer said.
PODCAST: REDEFINING HEALTHCARE WITH TRUSTWORTHY AI
A healthcare-specific genAI platform offers several advantages over a general, or foundational, model, explained Su. Because it is trained from the beginning on vast amounts of medical data, it can respond to medical questions with greater accuracy and relevance compared to foundational models, and it is more scalable and less costly.
"Foundational models are great," said Su, "but they don't quite meet the needs of healthcare applications. We're excited about where a domain-specific model can take us."
Iyer offered a small, but telling, example: "A generic AI model might assume that anything labeled positive is good, but if you tested positive in a medical exam, there are reasons to be concerned. At the same time, nobody wants to be negative in the financial domain," he explained.
Understanding the different meanings of words and other data in healthcare is one reason it's critical to use a domain-specific model.
Srini Iyer
SVP, Health & Civil Sector CTO, Leidos
However, a healthcare-specific model would know that a positive result in a healthcare setting can signal an urgent problem. "Understanding the different meanings of words and other data in healthcare is one reason it's critical to use a domain-specific model," Iyer added.
Baking in security and reliability
It's well known that genAI models can "hallucinate," that is, provide surprisingly wrong answers. Healthcare only raises the bar on the need to eliminate such errors, said Su. "Eighty-five percent accuracy is pretty good in most industries," he explained. "That won't cut it in healthcare."
To build more trustworthiness into the Leidos application, the collaborators enabled its AI model with "retrieval augmentation generation," or RAG, which allows the model to update and improve itself as new data becomes available without having to retrain the model.
"If a patient has diabetes, you'd want to make sure that anything the application does for that patient takes into account the latest diabetes guidelines," said Su.
Iyer added that further protection comes from ensuring that a human remains in the loop in anything the AI application does. "That was critical for us," he said. "You can't just let the model run on autopilot. We have a lot of people who go through the model's output to verify and validate that it's right."
Human oversight of AI is part of the Framework for AI Resilience and Security, or FAIRS, developed by Leidos. FAIRS focuses on safe and ethical AI deployment and helps ensure that AI systems are resilient and secure. "We built this whole project on that framework," said Iyer. "The security and reliability are baked in."
The collaborators were also mindful of the potential problem of bias-that an AI model may treat different types of people differently if it has been trained on data that reflects existing or past biases.
Eighty-five percent accuracy is pretty good in most industries. That won't cut it in healthcare.
Ken Su
Senior Outbound Product Manager, CloudAI for Healthcare, Google
"There used to be a lot of bias in the notes that clinicians would put in patient records, in terms of different patient demographics," explained Su. "The model is trained on some of that data, so we have to make sure that doesn't get into the model."
Scaling for future demand
Leidos is now looking to scale up the application to handle a demanding production environment. "We process thousands of reports every week and look at millions of pages of documents," said Iyer.
PODCAST: TACKLING OVERWORK IN HEALTHCARE: THE POWER OF AI
He added that although the project is only a demo now, it was built with scaling in mind-including the ability to process not just text, but also images, video, and audio.
Iyer is also already thinking about other applications in healthcare that can benefit from a domain-specific genAI model.
"The entire process of claims management is an opportunity to use AI to reduce waste and abuse," he said. "I'd also like to see if genAI can address the issue of data interoperability and eventually provide clinical decision support. There are just so many use cases for AI in healthcare; we can leverage the technology in numerous ways."
LEARN MORE ABOUT OUR TRUSTED MISSION AI APPROACH
View additional multimedia and more ESG storytelling from Leidos on 3blmedia.com.
Contact Info:
Spokesperson: Leidos
Website: https://www.3blmedia.com/profiles/leidos
Email: info@3blmedia.com
SOURCE: Leidos
View the original press release on accesswire.com
FAQ
What is the purpose of Leidos' new AI application in healthcare?
How does Leidos' healthcare AI application differ from general AI models?
What measures has Leidos (LDOS) implemented to ensure the reliability of its healthcare AI?