У нас вы можете посмотреть бесплатно MediBridge AI: Real-Time Clinical Decision Support with BigQuery & Gemini или скачать в максимальном доступном качестве, видео которое было загружено на ютуб. Для загрузки выберите вариант из формы ниже:
Если кнопки скачивания не
загрузились
НАЖМИТЕ ЗДЕСЬ или обновите страницу
Если возникают проблемы со скачиванием видео, пожалуйста напишите в поддержку по адресу внизу
страницы.
Спасибо за использование сервиса ClipSaver.ru
📝 Description See how MediBridge AI turns raw oncology records into doctor-ready insights in seconds, using only Google Cloud tools. In this full walkthrough you’ll learn how I built a complete, production-style pipeline, no external servers required: 🧬 Semantic Search • Generate 384-dimension embeddings with MiniLM • Use *BigQuery VECTOR\_SEARCH* to instantly find clinically similar cancer cases from the TCGA dataset 🤖 AI-Powered Care Cards • Call ML.GENERATE_TEXT with Gemini Flash directly inside SQL • Produce structured JSON summaries with staging, recommended treatments, and follow-up plans ⚡ Real-Time Workflow • Combine search + generation into a single BigQuery pipeline • Deliver tumor-board-ready “Care Cards” in under 10 seconds, cutting case-review time from hours 📊 What You’ll See • Step-by-step notebook demo • Flask + Cloud Run web app interface • Performance profiling and visualization of latency across vector search, AI generation, and data processing ⏱️ Timestamps 0:00 Introduction – MediBridge AI Clinical Engine 0:31 Step 1 – Imports & Parameters 0:44 Step 2 – Authenticate 🔒 with Google Cloud 0:52 Step 3 – Sanity-Check Source Table 1:02 Step 4 – Load Embedding Model 1:16 Step 5 – Build Embeddings Table 1:28 Step 6 – Create Vector Index 1:38 Step 7 – Semantic Search (VECTOR_SEARCH) 1:56 Step 8 Part 1 – AI-Generated Clinical Guidance 2:15 AI vs. Traditional SQL 2:32 Step 8 Part 2 – Enhanced Clinical Care Card 2:46 Step 9 – End-to-End Workflow Demo 🔥 3:01 Tumor-Board-Ready Care Card 💖 3:06 Performance Analysis 3:22 Flask Web-App Demo 4:04 Conclusion & Impact ✔️ 🔗 Resources & Links • GitHub Repo: https://github.com/tuba89/medibridge-ai • Kaggle Notebook: https://www.kaggle.com/code/assiaben/... • Medium Blog Post: / medibridge-ai-semantic-search-vector • Dataset: The Cancer Genome Atlas (TCGA) – de-identified cases Tech Stack: BigQuery ML · Gemini Flash · VECTOR_SEARCH · Python · Sentence-Transformers · Flask/Cloud Run ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ Built with ❤️ by Assia — Google BigQuery AI Hackathon 2025 Submission ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- #BigQueryAI #MLGenerateText #VectorSearch #DataPipeline #AI #HealthTech #OpenSource