Academic Research

Main project dashboard |full

Overview

Academic research often involves publishing peer-reviewed papers and journals, especially in top-tier and world-renowned conferences. An initial idea has to be developed into a paper that makes a significant contribution to the AI and machine learning community. Check out my Google Scholar page to view my publications.

Challenge

Peer-reviewed publications usually require high-quality paper submissions, as top-tier conferences have high paper rejection rates. The process can be mentally challenging due to time constraints and the demanding comments from reviewers.

Methodology

  1. Research Design: Set up a hypothesis and identify the research goal, such as a new framework, theoretical advancement with proofs, new model architecture, software application, or improved algorithms.
  2. Proof-of-Concepts: Conduct small experiments to validate or disprove the hypothesis. If the hypothesis is weak, it’s beneficial to fail fast.
  3. Proofs and Experiments: Provide mathematical definitions, theorems, and proofs to make the paper self-contained. Run experiments to compare with existing benchmarks.
  4. Paper Drafting: Write clearly about the introduction, methodology, literature review, and experiments.
  5. Submission and Rebuttal: Submit the paper for review and respond to reviewers’ comments during the rebuttal period.
  6. Camera Ready and Presentation: If the paper is accepted, prepare the camera-ready paper, a poster, and slides for the conference presentation.

Tools

Research collaboration is crucial in an academic setting, and different team members usually handle different parts of the methodology mentioned above. It is also essential to use the right tech stack for experiments. This includes: Python with PyTorch for development, GPU and cloud for model training, Git and GitHub for version control and collaboration, continuous integration/continuous deployment (CI/CD) practices for testing and deployment.

Results

The peer-reviewing paper submission process is repeated with improvements to the paper quality by addressing reviewers’ comments and team feedback. This cycle continues until the paper is of sufficient quality to be accepted for publication.