Skip to content

Add 16 refs for the Kiewit PLF proposal and companion whitepaper - #74

Merged
jeremymanning merged 1 commit into
masterfrom
add-kiewit-plf-refs
Apr 19, 2026
Merged

Add 16 refs for the Kiewit PLF proposal and companion whitepaper#74
jeremymanning merged 1 commit into
masterfrom
add-kiewit-plf-refs

Conversation

@jeremymanning

Copy link
Copy Markdown
Member

Summary

Adds 16 new bibliography entries supporting the Kiewit proposal (a Dartmouth Provost Leadership Fellows 2026–2027 application) and its companion whitepaper on distributed diffusion-LM inference.

All entries were individually web-verified against primary sources and formatted to pass python bibcheck.py verify cdl.bib.

Historical Dartmouth references

key paper verified via
KemeKurt68 Kemeny & Kurtz, "Dartmouth time-sharing", Science 162(3850):223–228 (1968) PubMed 5675464
Keme72 Kemeny, Man and the Computer (Scribner's, 1972) Open Library OL5282840M
McCaEtal06 McCarthy, Minsky, Rochester & Shannon, AI Magazine 27(4):12–14 (2006 reprint of 31 Aug 1955 proposal) AAAI OJS PDF (page range read directly)
Rank18 Rankin, A People's History of Computing in the United States (HUP, 2018) HUP catalog
Kurt81 Kurtz, "BASIC", in History of Programming Languages, Wexelblat ed., Academic Press, pp. 515–537 ACM DL 10.1145/800025.1198404

Kurt81 uses Force = {True} because bibcheck's sentence-case rule would render the acronym "BASIC" as "Basic".

Systems / distributed-inference references

key paper
KwonEtal23 PagedAttention, SOSP 2023 (pp. 611–626)
LiEtal24 DistriFusion, CVPR 2024 Highlight
HanEtal24 David helps Goliath, NAACL 2024 long (pp. 8385–8400)
BorzEtal23a Petals demo, ACL 2023 demos (pp. 558–568)
BorzEtal23b Follow-up NeurIPS 2023 paper on distributed inference/fine-tuning of LLMs over the internet (pp. 12312–12331). Distinct from the ACL demo above; note the different transliteration of "Ryabinin" (this paper) vs "Riabinin" (the ACL demo)
ShihEtal23 Parallel sampling of diffusion models (ParaDiGMS), NeurIPS 2023 Spotlight (pp. 4263–4276)

Diffusion-LM references

key paper
YeEtal25 Dream 7B (arXiv:2508.15487)
NieEtal25 LLaDA (arXiv:2502.09992)
TrauEtal25 DCUtR NAT-traversal measurement campaign (arXiv:2510.27500). The project's local refs.bib had abbreviated the author list with "and others"; this entry carries the verified 4-author list
BradNakk25 Classifier-free guidance is a predictor-corrector, TMLR 2025 (OpenReview forum zrWNtzSZsf)
RazaEtal26 When are two scores better than one? (full title includes "investigating ensembles of diffusion models"), TMLR 2026 (OpenReview forum 4iRx9b0Csu)

Integrity check

python bibcheck.py verify cdl.bib
→ looks good!

Test plan

  • python bibcheck.py verify cdl.bib passes cleanly on this branch
  • Each new entry's primary source URL was opened and the field values were compared line-by-line
  • No existing cdl.bib entries were modified
  • No duplicate keys introduced (bibcheck duplicate check passed)
  • Key-naming convention: 4-author-prefix + 2-digit year, with a/b suffixes on collision (BorzEtal23a / BorzEtal23b)

🤖 Generated with Claude Code

All entries were individually web-verified against primary sources
(AAAI OJS, ACM DL, ACL Anthology, CVPR Open Access, NeurIPS
proceedings, arXiv, HUP catalog, Open Library, PubMed) and
formatted to bibcheck verify conventions.

Historical Dartmouth references:
  - KemeKurt68: Kemeny & Kurtz, "Dartmouth time-sharing",
    Science 162(3850):223-228 (1968). Verified via PubMed 5675464.
  - Keme72: Kemeny, Man and the Computer (Scribner's, 1972).
    Verified via Open Library OL5282840M.
  - McCaEtal06: McCarthy, Minsky, Rochester, Shannon, AI Magazine
    27(4):12-14 (2006 reprint of 31 Aug 1955 Dartmouth proposal).
    Page range 12-14 confirmed by reading the AAAI OJS PDF.
  - Rank18: Rankin, A People's History of Computing in the United
    States (HUP, 2018).
  - Kurt81: Kurtz, "BASIC", in History of Programming Languages,
    Wexelblat ed., Academic Press 1981, pp. 515-537. ACM DL
    10.1145/800025.1198404 . Uses Force=True because bibcheck's
    sentence-case rule would render the acronym "BASIC" as "Basic".

Systems / distributed-inference references:
  - KwonEtal23: PagedAttention, SOSP 2023, pp. 611-626.
  - LiEtal24: DistriFusion, CVPR 2024 Highlight.
  - HanEtal24: David helps Goliath, NAACL 2024 long, pp. 8385-8400.
  - BorzEtal23a: Petals demo, ACL 2023 demos, pp. 558-568.
  - BorzEtal23b: Follow-up NeurIPS 2023 paper on distributed
    inference/fine-tuning of LLMs over the internet, pp. 12312-12331.
    Distinct from the ACL demo; note the different transliteration
    of "Ryabinin" (this paper) vs "Riabinin" (the ACL demo).
  - ShihEtal23: Parallel sampling of diffusion models (ParaDiGMS),
    NeurIPS 2023 Spotlight, pp. 4263-4276.

Diffusion-LM references:
  - YeEtal25: Dream 7B (arXiv:2508.15487).
  - NieEtal25: LLaDA (arXiv:2502.09992).
  - TrauEtal25: DCUtR NAT-traversal measurement campaign
    (arXiv:2510.27500). Original grant-project refs.bib abbreviated
    the author list with "and others"; this entry carries the
    verified 4-author list.
  - BradNakk25: Classifier-free guidance is a predictor-corrector,
    TMLR 2025 (OpenReview forum zrWNtzSZsf).
  - RazaEtal26: When are two scores better than one? (full title
    includes "investigating ensembles of diffusion models"), TMLR
    2026 (OpenReview forum 4iRx9b0Csu).

Integrity check: python bibcheck.py verify cdl.bib -> "looks good!"

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
@jeremymanning
jeremymanning merged commit b929127 into master Apr 19, 2026
1 check passed
paxtonfitzpatrick pushed a commit to paxtonfitzpatrick/CDL-bibliography that referenced this pull request Aug 8, 2026
The PR ContextLab#74 / ContextLab#75 / ContextLab#76 commits inadvertently introduced @comment{...}
blocks into cdl.bib via bibtexparser's serialization. bibtex's parser
chokes on the embedded URLs and commas inside those blocks (the
\"Folding@home, treating it\" pattern is the most direct trigger),
causing exit-code-2 failures for any downstream document that runs
bibtex against the master cdl.bib.

The fix: strip @comment{...} entirely. The audit-trail content these
blocks carried is preserved in the original PR descriptions and in
each submodule commit message, which is the appropriate place for
that history. cdl.bib is now machine-clean again.

Verification:
  python bibcheck.py verify cdl.bib -> looks good!  (6206 entries)
  bibtex (against a downstream document)  -> exit 0, no errors

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant