Spring 2024 Reading Group

• Location: Info Building 809
• Time: Wednesday 4:30pm (biweekly)

Please sign up here: 2024 Spring Reading Group Schdule

We will focus on the following topics for this year's reading group

• Efficient adaptation of LLM & foundation models cross domains and tasks
• Learning to transfer for continuous adaptation (with a focus on domain dynamics learning)
• Impactful application of above topics in the medical domain

The selected paper should have been a seminal work in its filed or a representative of the state-of-the-arts that made large impact. This rule can be relaxed if the paper is highly relevant to our current line of research. We would also prefer works that are well-written with good theoretical support.

• Presenter: Xuechao Chen
• Paper: LORA: LOW-RANK ADAPTATION OF LARGE LANGUAGE MODELS
• Slides: paperreading_lora.pptx

• Presenter: Hanbing Liu
• Paper:
• Slides:

• Presenter: Jingyun Yang
• Paper: Learning to Route Among Specialized Experts for Zero-Shot Generalization
• Slides:reading_group_240410.pptx

• Presenter: Xiangyu Chen
• Paper: Merging by Matching Models in Task Parameter Subspaces
• Slides:model_merging_240515.pptx

• Presenter: Haohua Wang
• Paper: Direct Preference Optimization: Your Language Model is Secretly a Reward Model
• Slides:

This year, we adopt a new discussion format called role-playing seminar. Instead of the traditional one-to-many style, Role-playing seminar encourages everyone to participate in the discussion by taking on a specific role. The roles will rotate in each meeting. Below is a summary of what participants in each role should prepare before the seminar. For more information, please check out the creator's blogpost.

Summarize the paper and its main claims and contributions. What problem is the paper trying to solve? Why is that problem interesting and important? How does the paper solve the problem and what is the novelty?

The paper has not been published yet and is currently submitted to a top conference where you’ve been assigned as a peer reviewer. Evaluate the paper's strengths and weaknesses including the technical novelty, clarity of exposition, magnitude of impact, and reproducibility.

This paper was found buried underground in the desert. You’re an archeologist who must determine where this paper sits in the context of previous and subsequent work. Find and report on one older paper cited within the current paper that substantially influenced the current paper and explain the relationship.

You’re a researcher who is working on a new project in this area. Propose an imaginary follow-up project not just based on the current but only possible due to the existence and success of the current paper.

You work at a company or organization developing an application or product of your choice (that has not already been suggested in a prior session). Bring a convincing pitch for why you should be paid to implement the method in the paper, and discuss at least one positive and negative impact of this application.

You are a detective who needs to run a background check on one of the paper’s authors. Where have they worked? What did they study? What previous projects might have led to working on this one? What motivated them to work on this project? Feel free to contact the authors, but remember to be courteous, polite, and on-topic.