# Machine Learning for Scent: Learning Generalizable Perceptual Representations of Small Molecules

arXiv: https://arxiv.org/abs/1910.10685 · HTML: https://mlsanity.com/paper/1910.10685/

## Key facts

- Cited by 28 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 23 Oct 2019.

## Built on by (most cited first, 28 of 28)

- [The Vendi Score: A Diversity Evaluation Metric for Machine Learning](https://mlsanity.com/paper/2210.02410/) (2022)
- [Graph neural networks for materials science and chemistry](https://mlsanity.com/paper/2208.09481/) (2022)
- [Graph Neural Networks for Molecules](https://mlsanity.com/paper/2209.05582/) (2022)
- [A graph representation of molecular ensembles for polymer property prediction](https://mlsanity.com/paper/2205.08619/) (2022)
- [Calibration and generalizability of probabilistic models on low-data chemical datasets with DIONYSUS](https://mlsanity.com/paper/2212.01574/) (2022)
- [SmellNet: A Large-scale Dataset for Real-world Smell Recognition](https://mlsanity.com/paper/2506.00239/) (2025)
- [PUFFIN: A Path-Unifying Feed-Forward Interfaced Network for Vapor Pressure Prediction](https://mlsanity.com/paper/2307.02903/) (2023)
- [mCLM: A Modular Chemical Language Model that Generates Functional and Makeable Molecules](https://mlsanity.com/paper/2505.12565/) (2025)
- [Predicting Aqueous Solubility of Organic Molecules Using Deep Learning Models with Varied Molecular Representations](https://mlsanity.com/paper/2105.12638/) (2021)
- [New York Smells: A Large Multimodal Dataset for Olfaction](https://mlsanity.com/paper/2511.20544/) (2025)
- [Olfactory Label Prediction on Aroma-Chemical Pairs](https://mlsanity.com/paper/2312.16124/) (2023)
- [Molecule Generation and Optimization for Efficient Fragrance Creation](https://mlsanity.com/paper/2402.12134/) (2024)
- [Molecular Odor Prediction Based on Multi-Feature Graph Attention Networks](https://mlsanity.com/paper/2502.01430/) (2025)
- [Position: Olfaction Standardization is Essential for the Advancement of Embodied Artificial Intelligence](https://mlsanity.com/paper/2506.00398/) (2025)
- [Diffusion Graph Neural Networks and Dataset for Robust Olfactory Navigation in Hazard Robotics](https://mlsanity.com/paper/2506.00455/) (2025)
- [NOSE: Neural Olfactory-Semantic Embedding with Tri-Modal Orthogonal Contrastive Learning](https://mlsanity.com/paper/2604.10452/) (2026)
- [See & Sniff: Learning Visuo-Olfactory Representations](https://mlsanity.com/paper/2606.27307/) (2026)
- [GLAMOUR: Graph Learning over Macromolecule Representations](https://mlsanity.com/paper/2103.02565/) (2021)
- [From meaning to perception -- exploring the space between word and odor perception embeddings](https://mlsanity.com/paper/2203.10294/) (2022)
- [Graph Anisotropic Diffusion](https://mlsanity.com/paper/2205.00354/) (2022)
- [Data Science In Olfaction](https://mlsanity.com/paper/2404.05501/) (2024)
- [DeepNose: An Equivariant Convolutional Neural Network Predictive Of Human Olfactory Percepts](https://mlsanity.com/paper/2412.08747/) (2024)
- [Motional representation; the ability to predict odor characters using molecular vibrations](https://mlsanity.com/paper/2509.16245/) (2025)
- [SCENT: Aligning Mass Spectra with Molecular Structure for Olfactory Perception](https://mlsanity.com/paper/2605.27009/) (2026)
- [What Images Cannot Say: Language-Guided Olfactory Representation Learning](https://mlsanity.com/paper/2607.06402/) (2026)
- [Beyond Predictive Accuracy: A Reliability-Aware Audit of Molecular Representations for Human Olfaction](https://mlsanity.com/paper/2607.24848/) (2026)
- [A General-Purpose Molecular Foundation Model Transfers Across Diverse Olfactory Tasks](https://mlsanity.com/paper/2608.25893/) (2026)
- [GraphNOSE: A Graph Transformer in Olfaction](https://mlsanity.com/paper/2609.05694/) (2026)

## Abstract (arXiv)

Predicting the relationship between a molecule's structure and its odor remains a difficult, decades-old task. This problem, termed quantitative structure-odor relationship (QSOR) modeling, is an important challenge in chemistry, impacting human nutrition, manufacture of synthetic fragrance, the environment, and sensory neuroscience. We propose the use of graph neural networks for QSOR, and show they significantly out-perform prior methods on a novel data set labeled by olfactory experts. Additional analysis shows that the learned embeddings from graph neural networks capture a meaningful odor space representation of the underlying relationship between structure and odor, as demonstrated by strong performance on two challenging transfer learning tasks. Machine learning has already had a large impact on the senses of sight and sound. Based on these early results with graph neural networks for molecular properties, we hope machine learning can eventually do for olfaction what it has already done for vision and hearing.