variational autoencoder

Discovery of microbial intergenic features with genomic language modeling and multimodal search

The paper identifies microbial intergenic features using a sparse autoencoder and multimodal search, revealing previously uncharacterized regulatory DNA and structured RNA elements.

Sequence Generation and Phylogenetic Inference with Generative Flow Networks

This paper introduces AncestorGFN, a method using Generative Flow Networks for alignment-free sequence generation and phylogenetic inference, showing it can recover ancestral relationships and generate novel sequences.

TLS-Sim: An Open-Source Toolkit for Individualized Transcranial Light Stimulation Modeling with Deep-Learning-Accelerated Simulation

TLS-Sim provides a toolkit for individualized transcranial light stimulation modeling using deep-learning-accelerated Monte Carlo simulations to improve efficiency and accuracy.

Improved Stability and Interpretability of Motor Modules Computed with an Autoencoder

This paper introduces an autoencoder-based method for computing motor modules from EMG data, showing improved stability and biomechanical interpretability compared to non-negative matrix factorization.

Concurrent model evidence computation and posterior sampling in continuous attractor network subspaces

This study shows continuous attractor networks can simultaneously compute model evidence and posterior sampling in distinct subspaces, advancing understanding of neural computation.

Preferred visual experiences provide intuitive descriptions of the functional properties of the cortical navigation network

The paper generates video clips that maximize activity in cortical navigation regions, revealing dynamic scenes that reflect functional properties of these regions.

Generalizable Cysteine Quantification in Pea Cultivars from SERS Spectra Using AI

The paper uses AI to quantify cysteine in pea cultivars from SERS spectra, finding that a 1D-CNN outperforms other models in accuracy and generalization.

Modelling interpretable patient-level representationsfrom structured and simple multimodal data

The paper introduces FACTMx, a variational framework that jointly models structured and simple patient-level data to learn interpretable representations, showing improved performance in simulations and applications to cancer and cardiovascular cohorts.

An Integrated Single-Nucleus Atlas Resolves Cell-Type-Specific Programs and Molecular Subtypes in Alzheimer's Disease

The paper constructs a single-nucleus transcriptomic atlas of Alzheimer's disease and identifies cell-type-specific molecular subtypes and therapeutic targets using a new framework.

On the use of variational autoencoders for biomedical data integration

The paper applies variational autoencoders to integrate biomedical data, explores model dynamics, and evaluates methods using synthetic and real datasets.

scDRP: Disentangled representation learning for predicting single-cell responses to perturbations and estimating individual treatment effects

scDRP estimates individual treatment effects and predicts single-cell responses to perturbations using disentangled representation learning with a sparsity regularized β-VAE.

GDA-Pred: Generative AI-Driven Data Augmentation for Improved Prediction of IL-6 and IL-13 Inducing Peptides

GDA-Pred improves prediction of IL-6 and IL-13 inducing peptides by using generative AI to augment small datasets, enhancing classifier performance through data expansion.

scDisent: regulatory-aware disentangled representation learning for multi-omic single-cell analysis

scDisent separates regulatory and expression variables in multi-omic single-cell data, improving integration and interpretability.

Cooperative Modular Representation Learning for Lung Adenocarcinoma Survival Prediction from Transcriptomic and Clinical Data

This paper introduces CMRL, a multimodal framework for LUAD survival prediction that integrates transcriptomic and clinical data, achieving better performance than existing methods.

A Cytokine Receptor Signaling Atlas Reveals How STAT Mosaics Fine-Tune T Cell Function

A fully orthogonal IL-2 receptor platform isolates 36 cytokine receptors in T cells; receptor-specific STAT combinations specify distinct fates, and STAT3-dominant programs give curative antitumor responses.

RIFT-VAE: grammar-conditioned pretraining and latent-space optimization for RNA inverse folding

RIFT-VAE improves RNA inverse folding by combining grammar-conditioned pretraining with latent-space optimization, achieving high structural recovery on benchmark datasets.

MOFTy: Multimodal Gaussian Process Factor Analysis with Numerical Information Field Theory

MOFTy disentangles spatial structures and recovers cross-modal patterns in multi-omics data by separating smooth gradients from fine-scale heterogeneity.

A PLUM Job: Peptide modeLs for Understanding and engineering antiMicrobial therapeutics

PLUM generates antimicrobial peptides with controlled function and length, achieving high yields and diverse sequences with favorable predicted properties.

LEN-Seek: Fast and scalable ligand binding-site similarity search in the latent space of an SE(3)-invariant graph VAE

LEN-Seek is a method for ligand binding-site similarity search using a graph VAE that encodes structural information into a latent space, achieving faster and more scalable results than existing methods.

Detection of Spatially Aberrant Cells in Spatial Transcriptomics Data by Conformal Prediction

The paper introduces SPADE, a framework that detects spatially aberrant cells in spatial transcriptomics data using variational autoencoders and conformal prediction.

Learning Discrete Cell and Niche Codes from Spatial Transcriptomics Using Dual Residual Vector Quantization

SQUINT learns discrete cell and niche codes from spatial transcriptomics data using residual vector quantization, outperforming baselines in identification and cross-section integration.

The Synthetic Fidelity-Stability Framework (SFSF): A Systematic Multi-Dimensional Benchmark of Synthetic Clinical Laboratory Data Generators

The paper benchmarks 17 synthetic data generators for clinical lab data, finding no single model excels across all metrics, and highlights ForestDiffusion's best privacy-utility balance.

Solving High-Dimensional Population Balance Equations via Dynamics-Preserving Autoencoders

The paper introduces a hybrid framework combining mechanistic models and machine learning to solve high-dimensional population balance equations for cell-state dynamics.

Whole-brain modeling of dynamic causal circuits in human cognition using amortized variational inference

This paper introduces MDSI-AVI, a framework for modeling whole-brain causal interactions in fMRI data, revealing new insights into working memory mechanisms.

XSSDense: Time-resolved X-ray Solution Scattering Density Reconstruction Using a Variational Autoencoder

XSSDense reconstructs electron density maps from X-ray scattering data using a variational autoencoder and genetic algorithm, capturing conformational heterogeneity in proteins.

Learning with Recurrence Geometric AI in Spatial Transcriptomics

This paper introduces RGAI, a geometric deep-learning framework for cross-domain analysis of spatial transcriptomics, enabling alignment of latent tissue states across diverse datasets.

Invasion status stratifies the composition of the human pancreatic cancer perineural niche

Xenium and single-nucleus profiling of 14 human pancreatic tumors shows invaded nerve neighborhoods enriched for classical cancer cells, myCAFs and lipid-handling macrophages but depleted of lymphocytes.

Variational inference in coupled models of amino acid substitution

The paper develops variational inference methods for amino acid coevolution models, showing improved efficiency and accuracy over existing approaches.

CENO: A Genome-Scale World Model for Evolutionary Sequence Interpretation and Programmable Regulatory Design

CENO is a long-context generative model for genomic sequence understanding and design, showing improved variant prediction and regulatory sequence generation across species.

Steering Sequence Generation in Protein Language Models through Iterative Lookback Monte Carlo Sampling

This paper introduces ILMC, a sampling strategy for protein language models that improves diversity while steering generation toward desired properties.