🎙️ Podcast Link 🎙️

So you’re writing a robotics conference paper?

Here are some practical tips to maximize the clarity, specificity & overall quality of your paper, so it gets accepted into a top #robotics#conference venue like #ICRA or #IROS.

Please reshare if you find it useful!

Tips are available in both text form & video!

Timestamps:

📌 0:00 (Introduction)
📌 0:08 (Types of Papers and Approach)
📌 1:12 (Know your Context )
📌 1:49 (Key Contributions and Specificity )
📌 2:37 (Gap Claims)
📌 3:13 (Don’t Die on Unnecessary Hills)
📌 3:59 (Three Types of Key Contributions)
📌 5:23 (Experimental Setup )
📌 6:13 (Show Weaknesses)
📌 6:38 (The Paragraph by Paragraph Summary Test)
📌 7:14 (Wrap-up)

Full Video Notes

✏️ 1) Know what type of paper you’re writing:

a) a robot doing better at some well known task or b) a completely new and novel task / application / area.

For a) Show your delta over previous contributions, theoretically and/or practically
For b) Sell the relevance to robotics

✏️ 2) Know your context: demonstrate a detailed and nuanced understanding of the application context in which your research is situated: the more specific and up-to-date, the better

✏️ 3) What are your key contributions? Be as specific as possible, broad statements are not helpful. Specificity can be around type of performance improvement, theoretical contribution, new sensoring modalities: be specific.

✏️ 4) Research gap: positively and constructively identify the research gap you’re addressing, without unnecessary punching down on previous research. Previous research has made good contributions, but there are still challenges to address – that’s what this paper is about

✏️ 5) Don’t die on any unnecessary hills: make the claims that are critical to justifying the research in the paper, and nothing more. It’s not a general soapbox: you are motivating the specific and usually narrow research direction you’ve taken in this specific paper.

✏️ 6) Know the 3 main types of contributions: 1) amazing new theory, 2) breakthroughs in performance capability / results, or 3) a self-evidently awesome new capability. Strength in one can compensate for weakness in another. But mediocrity across the board is a hard sell…

✏️ 7) Experimental setup: real robot if possible, dataset from real robots otherwise. Experiments should convince reader that your result wasn’t a fluke – typically done with multiple experiments or datasets that vary substantially. One flagship experiment, others can be smaller

✏️ 8) Candidly show your weaknesses: help researchers who follow your work understand where the remaining gaps are, & make it easy for your reader to understand not just the strengths of your approach but its shortcomings. Usually much appreciated by reviewers.

✏️ 9) Near submission: the paragraph by paragraph test. Go through para by para, check each para has a single clear and coherent message, that is critically needed for the paper. Reword / break-up / remove as necessary. You should be able to summarise each para in a short sentence!

✏️ 10) Have fun! Paper writing is an integral part of the research process. It isn’t just an end output but is also formative in helping you think about & position your research, & identify what you could do next. Good luck!