Undergrad Research

학부연구 주제

Jay_J 2024. 3. 20. 04:20

연구 프로젝트 설명:
Research Project (high level) Description: Motivation: Many undergraduate courses, especially senior-level courses, involve project-based learning. For example, the Computer Science department at VT has many capstone courses which are designed “to synthesize and integrate skills and knowledge acquired throughout the CS undergraduate curriculum, and which includes a significant design experience, where teamwork and written and oral communication are a key part of that design experience [link].“In such project-based courses, typically, professors ask students to give milestone presentations so that professors can check their progress, give feedback to each team, and allow students to see how other teams have made progress. By nature, getting feedback from milestone presentations can benefit the team in gaining ideas for improving the project and successful final artifacts. Some instructors use a peer-assessment system so that peers’ evaluations will collectively determine the team’s grade for the presentation. Peer assessment has proven to have positive learning outcomes. Importantly, peer assessment is a social process; some claim that anonymity might have advantages. While it can be beneficial for students to learn from evaluating other students’ work, give constructive feedback, and receive feedback from others, students typically do not feel comfortable allowing their peers to determine their grades [link]. Students’ grading may be inconsistent across individuals, and some can even game the system in favor of their grades. If we only include peer-review feedback without grading, eliciting participation can be another challenge for instructors. PI’s teaching practice to address the problem: PI Lee has developed and applied a novel peer-review feedback system in his Creative Computing Studio capstone course that he taught four times. In this class activity, the peer-assessment system has been modified; the class gives feedback to a presenting team, and the presenting team evaluates the feedback they received from their peers based on the quality of the feedback. For example, Whenever there is a group presentation, the PI asks everyone in the class to give feedback to the presenting teams. For example, suppose there are ten teams of four students in a project-based class. In that case, each student will have to pay attention to everyone else’s presentations (9 presentations) except their own presentation and leave feedback in an online survey. The online survey has three types of feedback entries: “I like (positives),” “I wish (improvements),” and “What if (suggestions)” sections. Once those responses are collected, the PI anonymizes the feedback and gives feedback to the team. Then the team has to evaluate each peer’s feedback based on simple criteria. (e.g., is it specific? Is it justified? is it actionable? Is it positive? [link]). Then one student will get their grades from 9 different teams, eventually becoming the student’s final grade for the assignment (5% of the course grade). This crowdsourced peer-review system has been developed over four iterations of the same course and is now used consistently. Here’s a sample spreadsheet that one team received from their peers. The PI anticipates multiple benefits of this system. First, students can practice giving constructive feedback, as their grade depends on the quality of the feedback. Second, each team may take the crowdsourced feedback more seriously than the instructor’s feedback as they find some common theme that emerges from multiple students. Third, the entire class can develop a sense of community as they give feedback to their peers over the semester, witnessing the progress of each team as a result of forced peer-review systems.

 

간단히 말해, 현재 Virginia Tech 의 학부 Computer Science 시스템은 졸업 전 필수 이수 과목인 'Capstone Course'를 듣도록 하고 있는데, 이 과목은 이론보다는 프로젝트 기반의 수업이다. 각 팀은 4명정도의 팀원으로 구성되어 있으며, 한 수업당 4-50명정도의 학생이 강의에 참여한다. 이 수업은 위에서 언급했듯 프로젝트 베이스이며, 각 팀들은 학기 시작부터 프로젝트를 구상하고, 모델링, 프로토타입, 와이어 프레임 등 여러가지 milestone을 거쳐 최종 프로젝트 발표에 이르는것이 Virginia Tech의 Capstone과정 중 하나이다. 이 과정중, 각 팀들은 서로 다른 팀의 발표 내용이나 프로젝트 진행 상황에 대해서 피드백을 남겨야 하는데, 그 피드백은 여러가지로 나뉘어서 채점된다. 

  •  피드백이 정당한가?
  • 피드백이 설득력이 있는가?
  • 피드백이 발표자의 요점을 잘 파악했는가?
  • 피드백이 너무 포괄적이지 않고 디테일 한가?

등등으로 채점을 해야 한다.

 

하지만 현재 시스템에서는 이러한 모든 일들을 Google Docs를 통해 하나하나 일일히 해야 한다. 교수님은 이러한 것들을 위해 학생들을 하나하나 각기 다른 팀에 배분해야 하며, 학생들이 피드백을 남길 때마다 교수님이 직접 체크해야 한다. 이것은 상당히 불편한 일이다. 대부분의 서비스들이 연동 되는 현 시점에 이렇게 귀찮은 방법으로 하는건 비효율적이다.

 

그래서!

이번 프로젝트의 목적은 위의 일을 번거롭지 않게 처리 할 수 있는 캡스톤 피드백 시스템을 개발 해내는 것이다. 이 시스템은 유저들간의 피드백을 주고 받을 수 있어야 하며, 일일히 채점하는 수고를 덜도록 학생들이 피드백을 평가하면 자동으로 채점이 되어 데이터베이스에 점수를 남기는 것이다.

 

여기까지가 기본적인 이번 연구 주제의 내용이며, 앞으로 추가 사항이 있을때마다 추가 해 나가도록 하겠다.