Note START HERE — Explore why the same strain can produce different stress after different loading histories. Bilingual interactive material-memory demo.

Note EXPLORE — Compare lighter support structures under load. The browser demo uses recorded candidates; the downloadable model is linked separately.

Note LEARN MATERIAL MEMORY — 3,788 synthetic loading trajectories across staged releases, including multiaxial paths and prospectively sealed tests.

HaomingLuo/AgentFEM-DENIM

Updated 3 days ago • 33 • 2

Note RUN A NEURAL CONSTITUTIVE MODEL — Conditional DENIM v2.2 combines explicit mechanics with learned hardening. Start with capability_v4.

Note FORECAST DIFFERENT FUTURES — 64 thermal scenarios, each branching into three control histories. Full fields, sensors, controls, and scenario-grouped splits.

Note LEARN GEOMETRY-TO-FIELD MAPS — Parameterized 3D bracket elasticity data for structural surrogate experiments; use the documented geometry-grouped protocols.

HaomingLuo/AgentFEM-GINO-Bracket-128

Updated 3 days ago • 20

Note REPRODUCE A STRUCTURAL SURROGATE — GINO-based bracket models, comparison baselines, and learning-curve results. Model cards report both strengths and limits.

Note Programmable rods: 216 scenarios, three actuation families, equilibrium branches, mesh checks and native AgentFEM validation.

HaomingLuo/AgentFEM-Morph-Rod-Surrogates

Tabular Regression • Updated 3 days ago

Note Branch-conditioned shape prediction, a mechanics corrector, and bounded inverse-design examples with reproducible tests.

Note DRAW A TARGET — Explore quasi-static rod inverse design in the browser: https://lab.haoming-luo.com/morph/ . Mechanics and optimization run in the demo.

Note BUILD AND COMPARE — Assemble a pipe and examine the effect of a damper. Direct browser access: https://lab.haoming-luo.com/open-science/pipe-lab/

Note THERMAL RESEARCH — Steady temperature fields, controlled transients and layered thermal bending; choose the task before choosing the data.

Note DYNAMICS — Synthetic vibration measurements, full-field targets and state-space reconstruction baselines.

Note PIPE RESEARCH — The interactive assembly lab, component data and reduced-order solver in one place.

Note SOFT RODS — Explore target shapes, then use branch-labelled data and numerical or neural baselines.