AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

SLOT-IR: Learning Disentangled Slot Representations for Infrared Spectral Unmixing

Jingru Gan, Yannah J.U. Melle, Yanqiao Zhu, Daniel Schwalbe-Koda, Wei Wang

NeurIPS 2026 · Probabilistic and scientific modeling › Molecules and proteins · approach: Equivariant and disentangled latent representations

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes SLOT-IR, a slot-based architecture that decomposes infrared spectra into disentangled per-component representations for spectral unmixing of mixtures.

Problem

Infrared spectral unmixing recovers individual component spectra from mixtures for chemical identification; linear superposition fails for liquid-phase mixtures with molecular interactions, and deep networks entangle components in shared representations.

Approach

Uses a slot-based architecture to decompose a measured IR mixture spectrum into separate slots, one per component, so that component spectra are disentangled rather than mixed in a shared latent representation.

Experimental setup

Baselines: linear unmixing methods deep unmixing networks

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy