Have a hard drive full of sequential data like audio, text, images, or even DNA and can’t figure out how to make the most of it? Have an innovative research solution or business objective in mind, but stuck in a quagmire of summarizing your data into a few condensed features? Need broad solutions that cover a number of tasks – for example, generating conversational responses based on your text or audio data?
Join the DSI for a hands-on AI Summer of leveraging the state-of-the-art transformer architecture in deep learning! Our 4-week session will be held May 10th – June 3rd on Mondays, Wednesdays, and Fridays from 9am – 12pm CT. In this short course, you’ll learn how to use, train, and share your models using your own research or project data.
Why transformers?
The practice of data science is changing with the rise of powerful, flexible and capable models called transformers. With the infusion of context, domain understanding, and standardized feature spaces, transformers allow you to leverage the richness of your complex datasets for enhanced insights and discovery. Transformers allow you to solve many different problems using a single pre-trained model, sometimes with no additional training on your data (“zero-shot” solutions). From text, to images, to audio (“textless natural language processing”), you can identify objects in images, identify concepts in spoken conversations, ask questions about tables of data, and more, all using the same type of model – transformers.
Who should attend?
Faculty, staff, post-docs, graduate students, and undergraduates from all disciplines are invited to attend! Not an expert in AI, deep learning, or programming? No problem. This short course is suitable for beginners and novices alike, divided into two parts, from 9am-12pm on each scheduled day:
- Introduction to Python (May 10th – 12th): An optional, hands-on, 3-day series for attendees new to programming or new to Python. This series is specifically targeted towards the Python skills and tools which will used in the Transformer Deep Dive session.
- Transformer Deep Dive (May 16th – June 3rd): A 3-week intensive on transformers, held Mondays, Wednesdays, and Fridays.
Researchers are encouraged to bring or have active access to their datasets to use during the workshops.
What will I learn?
This short course will provide an immersive experience in using transformers. You’ll gain:
- A high-level understanding of transformers, their building blocks, and their relationship to other deep learning architectures
- Insight into the capabilities and applications of transformers for your project purpose
- Hands-on experience in training and deploying transformer models
- Guided practice in training and using transformer models with text, audio, image, and multimodal data
- Dedicated office hour time to discuss your research project with DSI staff data scientists
- An opportunity to build new partnerships with faculty and students in other disciplines for collaboration, undergraduate research, immersion, and data sharing
This short course is an interactive, hands-on experience with live coding. Come prepared to hop on Google Colab or your chosen computational platform and code along with us during each session!
How do I apply?
Register here to attend! Contact Jesse Spencer-Smith or Charreau Bell for more information.
We look forward to seeing you there!