AI4Research open seminar

While recent advances in AI are enabling next generation data-driven life science, it suffers from issues in generalization and interpretability which further hinders critical applications in healthcare. The human-in-the-loop approach provides an intermediate way to build trust between AI models and users, and allows progressive adoption of AI tools in data-driven life science applications. In this talk, we will introduce our efforts in creating a computational platform (ImJoy: https://imjoy.io) for improving interactivity and scalability of AI software, building AI tools for enabling human-in-the-loop annotation and model training, and creating the bioimage model zoo (https://bioimage.io) for making AI models accessible to everyone. You will also see live demos on how we can use massive NLP models (OpenAI Codex) for generating Python code from plain English. We envision that in the near future life scientists can work with AI models in a human-in-the-loop setting and perform data exploration and analysis tasks without coding.