[Review Needed] 2026-04-r-variants and 2026-04-m-caas

[lalithaar]

@sanand0 @mvark I have added the drafts to the site to make it easier to review:

• https://isrl-research.github.io/pub/2026-04-r-variants/
• https://isrl-research.github.io/pub/2026-04-m-caas/

We are looking forward for your review on this, please help us see where it breaks. Can you please let me know a review timeline that works on your end ?, so we can plan around it appropriately.

If overleaf is more helpful for inline comments, it's still open too.

[lalithaar]

The Groundsource pipeline has four steps, not three. The fourth — location reconciliation — matches extracted location strings against WebRef entity codes, a pre-existing verified index. That step exists because Google had something to reconcile against. In the IFID implementation, no equivalent canonical index exists yet; that's precisely what IFID is building toward. So CAAS collapses steps 3 and 4 into "spatial precision" not as an error but because step 4 has no analogue in the current IFID pipeline — the reconciliation infrastructure is what the EMF framework is constructing. The CAAS paper is implicitly describing what was done, not what Groundsource did.

That said, the description in Section 4.2 is about Groundsource specifically, so collapsing their four steps into three is technically a misrepresentation of their pipeline even if the parallel holds structurally. The cleaner fix is to represent Groundsource's four steps accurately, then explicitly note in Section 4.3 (convergence) or a footnote that IFID's pipeline currently performs steps 1–3 and the reconciliation step (4) is the work IFID is building toward — which actually strengthens the paper's framing of IFID as a foundation-building project rather than weakening it.

[mvark]

@lalithaar I can only see the CAAS Project in Overleaf. Can you please provide access to "Data Acquisition and Ingredient Extraction" in Overleaf so that I can add my review comments? I will share my feedback within 2 days.

[lalithaar]

@lalithaar I can only see the CAAS Project in Overleaf. Can you please provide access to "Data Acquisition and Ingredient Extraction" in Overleaf so that I can add my review comments? I will share my feedback within 2 days.

It's saved as data-paper.tex under the same project, please check. Looking forward to your feedback