Instructor: Xiang Lorraine Li
Office: 130 N Bellefield Ave, 517
Office hours: by appointment
Email: xianglli@pitt.edu (Note: please use "CS3730" at the beginning of the subject line when emailing.)
In this course, we will dive into the recent advancements in language, knowledge, and reasoning, analyzing and discussing relevant literature. The primary goal of this class is to cultivate your ability (learn) to read numerous papers, ask critical questions, identify limitations, and ultimately synthesize new ideas.
Prerequisites: Basic understanding of probability and linear algebra is required. Familiarity or experience with natural language process and machine learning is recommended.
<aside> 1️⃣ For each class, we’ll assign two required readings to discuss in class: we’ll go deep into Paper 1 in the discussion so it requires careful reading, and you can skim Paper 2 time-permitting. We’ll have three presentations with assigned presenters to play a subset of the following five roles [1] in each class:
</aside>
Mandatory roles:
🌻 Original Author [8-10 min]. You’re among the original authors of Paper 1. You’ll give a 8-10min short, high-quality presentation at a conference.
👩🏽🔬 Scientific Peer Reviewer [5-7 min]. Paper 1 has not been published yet and is currently under review at a top conference where you’ve been assigned as a peer reviewer. Complete a full review of this paper answering all prompts of the official review form of the top venue in this research area (e.g., ACL/NAACL/EMNLP/NeurIPS/ICLR/…). This includes recommending whether to accept or reject the paper. Review template could be found here. Please do maintain a polite and considerate voice when sharing limitations about the work (as if the readers of the reviews might be yourself!) and even better if you can suggest constructive and concrete action items to improve the paper.
Free-style roles (Choose one role from the following three roles)
🍜 Imaginative Researcher [5-7 min]. You’re a researcher who is working on a new project in the area of Paper 1. 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.
🏺Archaeologist [5-7 min]. Paper 1 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 one newer paper that cites this current paper.
🦦 Original Author of a Related Paper [8-10 min]. You’re among the original authors of a related Paper 2. You’ll give a 5-10 min short, high-quality presentation at a conference.
Each role-playing presentation will be completed by an individual student. We will use a shared Google Slides for all presenters to streamline the presentation. Each role-playing presentation will begin with a title page that states which role they are playing and the name(s) of the presenter(s). The presentation material can be borrowed from elsewhere so long as the original source is acknowledged.
<aside> 2️⃣ For people who are not assigned as the official presenters, you will serve as the discussion assistant of the class:
</aside>
Complete a written artifacts ****on any paper of your choosing from among the required and optional readings (to be shared with the class).
📝 A study (summary) note of a paper created in some note-taking app (e.g., Notion)
📑 A blog post of a paper or a group of papers that reads like a tutorial
🙋♀️ An opinion piece reflecting your own thoughts and perspectives
📕 A scribe of the class discussion (key questions and discussions)
Each written artifact will be completed by an individual student. It should be at least one paragraph (minimum 100 words) in length within one week of the class takes place. The written artifact should not contain any plagiarism (e.g., a blog post elsewhere or ChatGPT opinions on the same paper/topic cannot be submitted as is or even after paraphrasing). Please keep in mind that all your written artifacts will be shared with the class.