Create abstract for TMS 2027

[Copilot]

Adds a submission-ready abstract for the TMS 2027 Annual Meeting & Exhibition, derived from the project's existing tensegrity energy-absorption work and kept consistent with idetc-abstract.tex.

Added — tms-2027-abstract.md

• Submission metadata — TMS 2027 (Orlando, Mar 14–18, 2027), abstract deadline Jul 1, 2026, plain-text ≤150-word format, oral preference.
• Target symposium — AI-Enabled Materials Processing: Integrating Accelerated Experimental Workflows and Processing-Aware Machine Learning (covers Bayesian optimization, surrogate models, closed-loop experimental workflows).
• Title / authors — Closed-Loop Bayesian Optimization of Multi-Material 3D-Printed Tensegrity-Inspired Energy Absorbers; full tensegrity team in first-paper author order: Marcus Madsen*, Audrey Christiansen*, Jinkwan Han*, Jeffrey R. Hill† (presenting), Sterling G. Baird† (all Department of Mechanical Engineering, BYU), with * and † equal-contribution footnotes.
• Abstract body — 141-word plain-text summary of the multi-material FDM (PLA + TPU) design–print–test loop driven by multi-objective BO on measured force/SEA/efficiency.

Added — Edison symposium-fit feedback (edison-trajectories/tms-symposium-fit/ + scripts/edison/)

• Submitted an Edison ANALYSIS task (16895002) with the abstract and the symposium call-for-abstracts flyer attached, asking how well the abstract fits the current symposium and whether an additive-manufacturing-focused symposium would fit better. Committed the fetched answer, full task dump, README headline, and reusable submit_tms_symposium_fit.py / fetch_tms_symposium_fit.py driver scripts.
• Headline: as written the abstract is a weak fit (~4/10) for the current symposium Fetch actual designs/models for various existing tensegrity structures #21 (which wants processing as a primary design variable). Best as-written home is AM-track symposium Increase focus on experimentation, strengthen and disperse mentoring throughout proposal #3 "Additive Manufacturing Modeling, Simulation, and AI: Microstructure, Mechanics, and Process" (~8.5/10, minimal edits). To credibly keep Fetch actual designs/models for various existing tensegrity structures #21 (~7/10), add FDM process variables (interface overlap, PLA/TPU nozzle temperature, layer height, print speed) plus a process-sensitive response to the BO loop. Backups: Explore the variables and base structures that we'd look at for the TPU + PETG #23 and apply overleaf updates from Jeff #5.

Notes for review

• Word count verified at 141 (TMS limit 150).
• Presenting author, oral-vs-poster preference, and symposium choice are easy to adjust before submission.
• The abstract is left unchanged pending a team decision between moving to AM symposium Increase focus on experimentation, strengthen and disperse mentoring throughout proposal #3 (as-written) or adding print-parameter optimization to stay in Fetch actual designs/models for various existing tensegrity structures #21; ready-to-paste rewrites for both paths are included in the Edison answer.

[sgbaird]

Noting that we need to get this submitted by July 1st

Cc @me-madsen @achris0520 @ctrhjk

[me-madsen]

This hits the head well on our current direction in research and concisely summarizes the important notes we would want to emphasize. From my perspective this abstract looks good. @sgbaird, the only potential addition I could think of might be adjusting the author list to include the rest of the tensegrity team, but I'm unsure if that's necessary for the abstract portion or not.

[sgbaird]

Thanks @me-madsen for catching. @copilot all authors should be added here. All affiliations are dept mech eng @ BYU.

author order mirroring that of the first paper (forgot which PR that is): Marcus*, Audrey*, Jinkwan*, Jeff†, Sterling†

• === equal contribution
  † === equal contribution

[Copilot]

Updated the author list in bd84227 to Marcus Madsen*, Audrey Christiansen*, Jinkwan Han*, Jeffrey R. Hill† (presenting), Sterling G. Baird†, all Department of Mechanical Engineering, BYU, with * and † equal-contribution footnotes.

[sgbaird]

@copilot attached is the flyer for that symposium. There is no materials optimization in this workflow (i.e., no optimization of print parameters). Perhaps that could be something we incorporate.. (i.e., various print parameters), which would then make it more of a materials problem, but weakly so (no change to composition, only to processing parameters). Send to edison for feedback on the fit in this symposium, and whether there might be another symposium (e.g., additive manufacturing focused) where it would better fit?

TMS2027-CFA-Flyer-021.pdf

[Copilot]

@copilot attached is the flyer for that symposium. There is no materials optimization in this workflow (i.e., no optimization of print parameters). Perhaps that could be something we i...

Sent to Edison (ANALYSIS task 16895002, with the abstract + this CFA flyer attached). Feedback committed in afec245 under edison-trajectories/tms-symposium-fit/ (full answer in the *.md, headline in the README.md).

Bottom line from Edison:

• As written, the abstract is a weak fit (~4/10) for symposium Fetch actual designs/models for various existing tensegrity structures #21 "AI-Enabled Materials Processing." That CFA explicitly wants processing history as a primary design variable, but our BO optimizes geometry/topology — not process parameters — so it's vulnerable in reviewer triage (not desk-reject territory, but the closing "processing-aware ML" line over-promises).
• Best home as written: AM-track symposium Increase focus on experimentation, strengthen and disperse mentoring throughout proposal #3, "Additive Manufacturing Modeling, Simulation, and AI: Microstructure, Mechanics, and Process" (~8.5/10). Near-zero edits — just soften "processing-aware." Its scope explicitly covers ML/AI-for-AM, AM mechanics, UQ, and build-failure handling without requiring processing to be the primary variable.
• To credibly keep Fetch actual designs/models for various existing tensegrity structures #21 (~7/10), make processing real: add 2–4 FDM process variables to the BO decision vector — most defensibly interface overlap/interpenetration, PLA and/or TPU nozzle temperature, layer height (also print speed) — plus at least one process-sensitive response beyond the crash metrics (interfacial-failure incidence / bond-strength proxy / void fraction / dimensional fidelity), and reframe failures as manufacturability constraints. This matches your "weakly a materials problem (processing varies, composition fixed)" intuition.
• Backups: Explore the variables and base structures that we'd look at for the TPU + PETG #23 "Algorithms Development in MSE" (if BO/qNEHVI method is the headline) and apply overleaf updates from Jeff #5 "Designing Complex Microstructures Through AM". No dedicated architected-materials, polymer-AM, or AM-mechanical-behavior symposium appears in the TMS 2027 flyer set.

The answer includes ready-to-paste ~114-word Path-A (move to #3, keep geometry story) and Path-B (#21 + process variables) rewrites, both under the 150-word cap. This is a scope decision (Path B means actually adding print-parameter optimization to the campaign), so I left the abstract unchanged — tell me which path and I'll update the title/symposium/body accordingly.

[sgbaird]

@claude provide recommendations for the top 5 symposia

[claude]

Claude finished @sgbaird's task in 1m 53s —— View job

---

Top 5 symposia recommendations — TMS 2027

• Read Edison symposium-fit analysis + abstract
• Cross-checked the TMS 2027 program (ProgramMaster) for additional candidates
• Ranked top-5 with fit scores + rationale

⏰ Heads-up: the abstract deadline is today (Jul 1, 2026). This ranking is built to let you pick a home and paste immediately.

Ranking combines Edison's flyer-level analysis (edison-trajectories/tms-symposium-fit/, which vetted the actual CFA text of #3/#21/#5/#23) with two additional data-driven symposia I confirmed are on the TMS 2027 program. Fit is judged against the paper as written — architected multi-material FDM energy absorbers optimized by closed-loop experiment-driven BO, with geometry/topology (not process parameters) as the decision variables.

#	Symposium	Fit (as-written)	Edit effort	Verdict
1	Additive Manufacturing Modeling, Simulation, and AI: Microstructure, Mechanics, and Process (flyer #3, AM track)	8.5/10	Near-zero	🥇 Best home as-is. Scope explicitly covers ML/AI-for-AM, AM mechanics, UQ, and failed-build handling — none of which require processing to be the primary variable. Only edit: soften the closing "processing-aware ML" line.
2	AI/ML/Data Informatics for Materials Discovery: Bridging Experiment, Theory, and Modeling (Data-Driven track)	~7.5/10	Low	🆕 Not in Edison's original set. Strong home for the closed-loop, experiment-driven BO methodology — the "bridging experiment and modeling" framing fits a surrogate-in-the-loop physical campaign directly, no process-variable requirement. Verify CFA scope before committing.
3	AI-Enabled Materials Processing: Integrating Accelerated Experimental Workflows and Processing-Aware ML (flyer #21 — current target)	4/10 now → 7/10 if revised	High (scope change)	⚠️ Current target, but its CFA wants processing history as a primary design variable. Credible only if you actually add 2–4 FDM process variables to the BO loop (interface overlap, PLA/TPU nozzle temp, layer height, print speed) + a process-sensitive response (interfacial-failure incidence / bond-strength proxy / void fraction / dimensional fidelity). That's a real campaign-scope decision, not a wording tweak.
4	Algorithms Development in Materials Science and Engineering (flyer #23, Data-Driven track)	~6/10	Medium (reframe)	Backup if the method is the headline — lead with qNEHVI / noisy MOBO / feasibility-constrained BO / experiment orchestration and de-emphasize the absorber application. Your paper currently reads application-first, so this is a stretch unless you re-center it.
5	Designing Complex Microstructures Through Additive Manufacturing (flyer #5, AM track)	~5/10 → 6.5/10 reframed	Medium	Weakest of the AM options — flyer is metal/process-structure-property centric, and a PLA/TPU architected absorber isn't a microstructure-engineering story. Only worth it reframed around interface morphology/defect state. #3 is strictly cleaner.

Honorable mention (not ranked — verify scope): 3D Printing of Scaffolds and Porous Materials (AM track) is the closest thing TMS 2027 has to an architected/porous-AM bucket; could suit the energy-absorber geometry story, but I haven't read its CFA. Edison confirmed there is no dedicated architected-materials, polymer-AM, or AM-mechanical-behavior symposium in the TMS 2027 set.

Bottom line

• Path A (recommended, zero-risk for today's deadline): submit to Increase focus on experimentation, strengthen and disperse mentoring throughout proposal #3 essentially as-is — just replace the final clause ("…practical lessons for accelerated, processing-aware machine learning…") with something like "…data-efficient closed-loop optimization of additively manufactured architected energy absorbers." A ready 114-word rewrite is in edison-trajectories/tms-symposium-fit/tms-symposium-fit-16895002-*.md (lines ~228–231).
• Consider Restructure MRG proposal to match grant format with Edison Scientific-informed revisions #2 if you'd rather headline the BO/closed-loop methodology than AM mechanics — but confirm the CFA scope first given the deadline.
• Path B (Fetch actual designs/models for various existing tensegrity structures #21) is only honest if the team commits to adding print-parameter optimization to the campaign. Not a today-afternoon change.

Want me to prepare the abstract for a specific choice (update the target-symposium line + swap in the matching rewrite + re-verify the word count)? Just say which number.

Sources: TMS 2027 program (ProgramMaster) · TMS 2027 meeting
· Branch: copilot/create-abstract-for-tms-2027

[sgbaird]

@claude update the abstract to mention that we'll also vary processing parameters (temperature, speed, layer height)

[claude]

Claude finished @sgbaird's task in 1m 43s —— View job

---

Update abstract — add processing-parameter optimization

• Read the abstract + Edison symposium-fit analysis
• Update abstract body to jointly optimize architecture and FDM process parameters (nozzle temperature, print speed, layer height)
• Verify word count ≤150 (now 149)
• Commit and push (725c887)

Done. The BO loop in tms-2027-abstract.md now jointly tunes architecture and FDM process parameters:

…uses Bayesian optimization to jointly tune architecture and FDM process parameters—nozzle temperature, print speed, and layer height— in a design–print–test workflow operating on physical measurements…

To fit the added clause under the word cap, I compressed the design-space list (dropped tension-element cross-section) and trimmed a couple of filler words. Word count is 149/150.

Note on symposium fit: this change is exactly the Path B that the Edison analysis recommended for keeping symposium #21 "AI-Enabled Materials Processing" (raises the fit from ~4/10 → ~7/10, since processing is now a real optimization variable rather than just a framing word). The abstract still targets #21, which now matches the content. If you want to push the fit higher, Edison also suggested adding a process-sensitive response beyond the crash metrics (e.g. interfacial-failure incidence or a bond-strength proxy) and framing print failures as manufacturability constraints — but that's a further scope decision I left alone. Just say the word.

[sgbaird]

@claude would any of the following individuals likely be receptive to a PLA + TPU multi-material application or are they primarily focused on other types of AM? (e.g., metal, ceramic, etc.)

• Jing Zhang, Purdue University
• Li Ma, Johns Hopkins Applied Physics Laboratory
• Charles Fisher, Office of Naval Research
• Brandon McWilliams, US Army Research Laboratory
• Yeon-Gil Jung, Changwon National University

[claude]

Claude finished @sgbaird's task in 1m 43s —— View job

---

Are these individuals receptive to a PLA + TPU (polymer) multi-material story?

• Researched each person's AM focus
• Assessed receptivity to a polymer (PLA + TPU) architected-absorber story
• Ranked below

Short answer: this is a metal/ceramic-dominated organizer list. None of the five is a polymer / FDM / soft-material specialist — so your entry point with all of them is the AI / BO / ICME-modeling method and the architected-structure mechanics, not the PLA/TPU chemistry. Material-agnostic "AI-for-AM" framing can land; a "look at our thermoplastic material system" framing will not.

Person	Primary AM focus	Receptive to PLA+TPU polymer?	Best hook for us
Jing Zhang (Purdue)	Metals + ceramics — LPBF, ceramic slurry extrusion, process-structure-property modeling. Editor-in-Chief of Int. J. Additive Manufacturing; co-edited a broad AM textbook (with Jung, below).	Moderate	Broadest scope of the group. As an AM-journal EIC he thinks material-agnostically; the closed-loop BO / process-property-performance framing fits even if polymers aren't his bench work.
Brandon McWilliams (ARL)	Metals — UHS steel, Al/Mg alloys, WE43. But explicitly works on micro-lattice / architected structures and compressive failure modes via LPBF.	Low–moderate	The architected energy-absorber + multi-scale-structure angle overlaps his lattice/crush-behavior interest. Material is wrong (he's metal), but the mechanics + structure-optimization story is his closest match.
Li Ma (JHU APL)	Metals, computational — ICME, physics-based microstructure prediction for LPBF, ML+simulation for AM qualification/certification.	Low	Her hook is methodology: ML-in-the-loop, UQ, and surrogate modeling for AM. She cares about the BO/data-efficiency machinery, not the polymer.
Charles Fisher (ONR)	Metals — naval structural alloys, ICME for AM, nickel-aluminum-bronze. Mission-driven (shipbuilding).	Low	Program-officer breadth means he sees polymer work, but his technical center of gravity is structural metals. Energy-absorption-for-protection framing is the only bridge.
Yeon-Gil Jung (Changwon)	High-temperature ceramics — thermal barrier coatings, lanthanum zirconate, spark-plasma sintering.	Lowest	A room-temperature thermoplastic crush absorber is essentially orthogonal to his TBC/high-temp-ceramics program. Weakest fit of the five.

Bottom line

• No one here is a polymer/FDM champion. If the symposium these five organize hinges reviewers on material system, PLA+TPU is a headwind.
• Your two viable bridges are (1) the AI/closed-loop-BO/ICME method — resonates most with Zhang, Ma, Fisher — and (2) the architected-lattice crush mechanics — resonates most with McWilliams.
• This reinforces the earlier symposium-fit read: lead the abstract with the data-efficient closed-loop optimization method + architected energy-absorption mechanics, and treat "PLA + TPU multi-material" as an implementation detail of a general workflow rather than the selling point. That framing is what makes a metal/ceramic-oriented organizing committee comfortable accepting it.
• ⚠️ Caveat: this is inferred from each person's published research focus, not from stated symposium scope. If these are the organizers of a specific TMS 2027 symposium, the written CFA scope is the authority — a material-agnostic "AI-for-AM / modeling & simulation" CFA (like flyer Increase focus on experimentation, strengthen and disperse mentoring throughout proposal #3) would happily take the paper regardless of these individuals' personal metal/ceramic leanings.

Sources: Jing Zhang (Purdue/IUPUI lab) · Zhang & Jung, Additive Manufacturing textbook · Li Ma (JHU APL) · APL ML+simulation for AM · Charles Fisher (ONR/NSWC Carderock) · Brandon McWilliams (ARL, lattice AM) · Yeon-Gil Jung (Changwon, TBC ceramics)
· Branch

[sgbaird]

@me-madsen let's stick with "AI-Enabled Materials Processing: Integrating Accelerated Experimental Workflows and Processing-Aware ML"

[sgbaird]

@claude for the "accelerated experimental workflow" aspect, add something that points out we're building something closer to a self-driving lab:

• full data and metadata provenance tracked in the cloud, remotely accessible
• direct Python integration between the multi-material 3D printer and our orchestration software
• human-in-the-loop software architecture

Word count is tight, so you don't have to put all of this in, but it can be summarized / help point out that it's an actually accelerated workflow that is increasing in terms of level of autonomy

[claude]

Claude finished @sgbaird's task in 1m 53s —— View job

---

Update abstract — frame the accelerated workflow as a self-driving lab

• Read the abstract + prior symposium-fit context
• Add self-driving-lab framing (cloud-tracked data/metadata provenance, direct Python printer↔orchestration integration, human-in-the-loop, rising autonomy)
• Verify word count ≤150 (now 150/150)
• Commit and push (653ca03)

Done. I added a dedicated sentence to the abstract body that reframes the "accelerated experimental workflow" as an emerging self-driving lab:

…Moving toward a self-driving lab, we couple printer and orchestration software via direct Python integration, track full data and metadata provenance in the cloud, and keep a human in the loop, raising autonomy.…

This folds in all three points you listed, compressed to fit the cap:

• Cloud provenance → "track full data and metadata provenance in the cloud"
• Direct Python integration → "couple printer and orchestration software via direct Python integration"
• Human-in-the-loop + increasing autonomy → "keep a human in the loop, raising autonomy"

Staying under 150 words: the new sentence added ~30 words, so I tightened the rest without losing content — trimmed the intro ("suspended within a continuous" → "in a continuous"; "exhibit" → "offer"; "per unit mass" → "per mass"; "too large to explore by trial and error" → "too large for trial and error"), shortened "in a design–print–test workflow operating on physical measurements, without relying on calibrated finite-element simulation" → "from physical measurements, not calibrated simulation", and condensed the closing sentence. Word count is now 150/150 exactly.

No change to the target symposium — still #21 "AI-Enabled Materials Processing" (per the team decision above). If 150/150 feels too tight to submit safely, say the word and I'll shave a word or two for headroom.
· Branch: copilot/create-abstract-for-tms-2027