AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

xTrimoGene: An Efficient and Scalable Representation Learner for Single-Cell RNA-Seq Data

Jing Gong, Minsheng Hao, Xingyi Cheng, Xin Zeng, Chiming Liu, Jianzhu Ma, Xuegong Zhang, Taifeng Wang, Le Song

NeurIPS 2023 · Architectures and efficiency › RNNs and normalization · approach: Distributed LLM inference systems · 30 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use asymmetric encoder-decoder processing only non-zero unmasked genes in the encoder instead of full-length TransformerxTrimoGene · pre-training compute (5M samples, 5 epochs)total train FLOPs (resource vs Transformer): 100% -> 3.4%Improved
use auto-discretization of expression values instead of round bin, up bin or equal bin (ablation)xTrimoGene · cell clustering (ARI, NMI, HOMO, CP, SIL)clustering metricsImproved
use regression (MSE) masked objective instead of classification (ablation)xTrimoGene pre-training · cell clusteringclustering metricsImproved
supervise zero-valued positions during masking (ablation)xTrimoGene pre-training · cell clusteringclustering metricsImproved
raise masking ratio from 15% to about 30%, then further to 90% (ablation)xTrimoGene pre-training · cell clusteringclustering metricsMixed
replace encoder-only Performer with asymmetric encoder-decoder (ablation)xTrimoGene · masked value recovery and cell clusteringPearson correlation / clusteringImproved
scale xTrimoGene from 3M to 10M to 100M parameters (ablation)xTrimoGene · pre-training losslossImproved
replace scBERT with xTrimoGene for cell type annotationscBERT · Zheng68KF1 score: 0.6695 -> 0.7354Improved
replace CellTypist with xTrimoGene on Zheng68K precisionCellTypist · Zheng68Kprecision: 0.7454 -> 0.7335Hurt
add xTrimoGene embeddings to GEARSGEARS · Norman perturb-seq top-20 DE genesMSE: -14.8% MSEImproved
use xTrimoGene embeddings in DeepDDSDeepDDS · drug combination synergy predictionROCImproved

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