Polymer Chemistry

Paramus Polymer brings computational chemistry, ML models, and domain ontologies together to accelerate polymer development with chemically verified workflows rather than generic AI output.

Enter copilot, openAI and Claude for polymer scientists: real structure handling, validated reaction chemistry, and predictive models spanning kinetics, rheology, and thermo-mechanical behavior.

The best of both worlds: large-language-model capability fused with chemical exactness for real polymer design and validated structure–property prediction.

Paramus POLY

The Paramus.ai Polymer Chemistry Solution accelerates polymer discovery through AI-powered models and predictive analytics.

from $599 USD*

* named user, per year

Order here:

Buy
Contact

Included

The polymer package from paramus contains:

Models

Models (Polymer-property prediction / generation / simulation)

  • TransPolymer — Transformer-based language model for polymer property prediction. Pretrained on large unlabeled polymer corpora and fine-tuned across multiple property benchmarks.
  • PolyNC — Unified “natural & chemical language” model (text-to-text) for predicting polymer properties via multi-task regression + classification.
  • PolyTAO — Transformer-Assisted Oriented pre-trained model for conditional polymer generation (“inverse design”) — generates polymers satisfying 15 predefined fundamental properties.
  • wD‑MPNN (weighted Directed Message Passing Neural Network) — Graph-based model adapted to polymer ensembles (accounts for monomer stoichiometry, chain architecture, degree of polymerization) for property prediction.
  • Polyply — Python suite generating parameters & coordinates for atomistic and coarse-grained polymer molecular-dynamics (MD) simulations (force-field and topology agnostic). Useful for bridging ML and MD workflows.

Data

Datasets (Polymer Data)

  • PL1M — ~1 million polymer entries, combining computed and experimental polymer properties.
  • RadonPy — All-atom molecular dynamics data for ~1,070 amorphous polymers, includes properties like density, heat capacity, thermal conductivity, refractive index under given conditions.
  • Open Macromolecular Genome (OMG) — Database of ~12 million constitutional repeating units derived from ~77 000 commercially available monomers, supporting generative polymer design.
  • OMG‑Property DB — Monomer-level properties for ~12 million synthetically accessible polymers (includes quantum chemistry & xTB computed descriptors, Flory–Huggins parameters, electronic properties, etc.).
  • VipEA — Dataset of vertical ionization potentials and electron affinities for over 10,000 polymers/copolymers (computed via xTB), useful for electronic-property prediction and ML benchmarking.
  • PolyIE — Annotated corpus of ~146 full-text polymer-material research articles, with named entities (compounds, properties, measurement contexts) to support information extraction and knowledge-graph building.

Scroll to Top