Next-Level Science using AI AssistantsOnline Event

27 Jun 2024, 09:00 → 11 Jul 2024, 12:00 Europe/Berlin

Zoom

Markus Gyger (Gyger Training), Daniel Mertens (DKFZ Heidelberg), Axel Schumacher (KIT), Nicole Merkle (KIT), Sikha Ray (KIT), Thorsten Auth (Forschungszentrum Jülich)

Next-Level Science using AI Assistants

Would you like to

• Work more efficiently and effectively than ever before?
• Save time on literature search and literature assessment?
• Produce easy-to-read text for cover letters, social media, for grants and scientific papers?
• Brainstorm more effectively on experimental strategies with the help of AI?

 

AI-based tools revolutionize science and enable researchers to work more efficiently and effectively. However, it is important for researchers to use these tools judiciously and with an understanding of their limitations. The course will kick off with an explanation of what GPTs are and how to use them. In the following, optimization of prompting using context and framework will be introduced and applied.

You will discuss how to address complex tasks and documents, such as extracting key claims, summarizing, brainstorming ideas, and finding sources using, e.g,. https://perplexity.ai.

You will discuss enhanced literature review, connectivity of papers, and interactive summaries of key information using, e.g., https://elicit.org, https://www.explainpaper.com, https://app.litmaps.com, https://researchrabbit.ai.

You will discuss aspects of texting, such as optimizing text, generating scientific text de novo (paragraphs, titles, cover letters, ...), generating text for social media ("Tweets", LinkedIn posts, ...) using, e.g., https://www.deepl.com/write, https://chat.openai.com/chat.

You will discuss how to use artificial intelligence for brainstorming, such as experiment planning, strategic planning, and career development, using, e.g., https://chat.openai.com/chat.

You will get a teaser for programming using artificial intelligence, from zero to implementation.

Registration, Participation, and Certificates

This course is organized by HIDSS4Health, BIF-IGS, and IHRS BioSoft. Registration is required to attend and places are limited, participation is free or charge for members of the three schools. In case of overbooking, priority will be given to fellows and faculty of the three schools. Participants that attend all three half-day online sessions and contribute to the homework will receive a certificate of attendance.

The course has started, and registration is closed.

Course Format

The course will be offered online in an interactive format. The trainers will use questions, material, text, and ideas provided by the participants to create an interactive learning experience. Content suggestions and questions will be implemented live online as input into the AI tools. Case studies and live exercises will help the participants to apply the concepts and tools to their own research. In-between the three live online lectures, tasks will be assigned to groups of participants and the trainers will be available for questions.

---

This event is organized by

International Helmholtz Research School of Biophysics and Soft Matter    |    BioInterfaces International Graduate School

Helmholtz Information & Data Science School for Health

LLM24.pdf

Thursday 27 June

09:00 → 12:00 Live Session: Introduction to Large Language Models

09:00 Introduction to Large Language Models

This introductory module will provide scientists with a fundamental understanding of large language models such as GPT and how they work. Given the increasing integration of AI in scientific research, understanding the underpinnings of such models and recognizing their pitfalls will enable participants to leverage their capabilities and mitigate limitations.

1. What is “GPT”? An explanation and overview of “Generative Pretrained Transformers”.
2. The architecture of large language models: Understanding transformer networks.
3. Mechanism of operation: How does GPT understand and generate text?
4. Strengths and limitations of large language models in scientific research.
5. Pitfalls to avoid when using large language models, and how to manage biases in AI.

10:30 How to use Large Language Models

Building on the knowledge of large language models, this module will dive into practical applications, demonstrating how scientists can use AI tools for text production and analysis Most important is the introduction of participants to advanced prompting techniques, which are critical for obtaining desired outputs.

1. Practical uses of large language models in life sciences.
2. Best practices for texting with AI models.
3. Techniques for effective text analysis.
4. Mastering the art of prompting for better outcomes.
5. Interactive session: Hands-on prompting exercise with AI models.

Wednesday 3 July

13:30 → 16:30 Live Session: Literature Review and Brainstorming

13:30 Literature review

AI is revolutionizing literature review and makes it time-effective and more objective because more literature can be assessed at once. This module explores how AI can be used to enhance literature review processes, ultimately aiding in more thorough and efficient research.

1. The role of AI in literature reviews.
2. Techniques for AI-powered text mining in scientific literature.
3. Tools and resources for AI-based literature reviews.
4. Case studies of successful AI application in literature review.
5. Hands-on session: Using AI tools for literature search and review.

15:00 Brainstorming and Structuring

AI can be used for comprehensive brainstorming and can provide structure, helping researchers develop innovative ideas and frameworks. This module covers AI's role in idea generation and organization, showcasing how it can inspire creativity and promote logical thinking in scientific research.

1. Introduction to AI-aided brainstorming.
2. AI tools for idea generation and mind-mapping.
3. How AI can help structure and organize thoughts.
4. Techniques for extracting value from AI-generated content.
5. Hands-on session: Engaging with AI for brainstorming and idea organization.

Thursday 11 July

09:00 → 12:00 Live Session: Advanced applications and Science Communication

09:00 From Zero to Programming using AI-Tools

AI is not only for programmers. This module is designed to empower scientists with no programming background to use AI tools effectively. It will use AI-tools to take you from nothing to a running implementation. The major challenge is changing the mindset towards “everything is possible”.

1. Introduction to the “twoTools” approach to programming with AI.
2. Introduction to the disruptive mindset of programming with AI tools.
3. Hands-on exercises on using AI to write simple code.
4. Troubleshooting common issues using AI.
5. Optional: Q&A session with expert AI programmers.

10:30 Science Communication with AI

This module delves into the application of artificial intelligence to significantly simplify science communication. Given the critical importance of clearly presenting complex scientific concepts, AI serves as an essential tool in achieving this objective. Participants will explore generative AI's capabilities to create texts and visuals that enhance the dissemination of their research to a broader audience.

1. Utilizing AI for Clear Communication: Apply generative AI to transform complex information into clear, engaging content and articulate intricate scientific issues straightforwardly and concisely.
2. AI-Driven Text and Image Generation: Discover how to use AI tools to produce concise texts and visuals that clarify and emphasize scientific concepts for a lay audience.
3. Idea Generation for Communication: Explore AI's potential to suggest images, analogies, and examples that make complicated scientific topics more accessible and relatable to a broad audience.