property-prediction

torch-molecule is a deep learning package for molecular discovery, designed with an sklearn-style interface for property prediction, inverse design and representation learning.

Open-source AI infrastructure for materials science

Chemical representation learning paper in Digital Discovery

CrysXPP: An Explainable Property Predictor for Crystalline Materials (NPJ Computational Materials - 2022)

Open Knowledge Enrichment for Long-tail Entities, WWW 2020

Predicting properties of small molecules using MPNN on QM9 dataset

An integrated Python package for molecular descriptor generation, data processing, model training, and hyper-parameter optimization.

EGAT - Edge Featured Graph Attention Networks for Property Prediction

Machine learning algorithm implementation in materials science

P2MAT - A python based user interface to predict melting point and boiling point of chemical compounds.

AI-powered real estate platform for property price prediction, recommendations, and market insights using Streamlit and Machine Learning.

CPU-only perimeter sanity filter for AI-predicted quantum-chemical properties (dipole, HOMO/LUMO, gap). Agnostic of the source model; flags violations of Koopmans theorem, molecular point-group symmetry, topology and physical ranges. Heuristic linter, not a DFT calculator.

NP-specific chemical language models (NPCLMs) for molecule generation and property prediction using state-space and transformer architectures.

Pittsburgh Single Family Home Prediction with Non-Conventional Data Streams

Smart, user-friendly property valuation app leveraging bulk price prediction, market summaries, feature insights, and top picks ,powered by Streamlit and machine learning.

A TensorFlow-based machine learning framework for predicting zeolite properties, particularly framework density, using neural networks and composite building unit fingerprints

ML for predicting the compressive strength of SCMs

Property prediction of metal-organic frameworks (MOFs) based on their molecular/crystal structures. Based on open source datasets.

🧭 Australian Property Orientation Finder

LLM benchmark for property prediction of proteins, protein pairs, and small molecules. Tasks from PEER, CALM, and MoleculeACE with regression, binary, binned, multiclass, and multilabel framings.