HyCoSeq: Contextual Hyperbolic Representation Learning for Genomic Sequences
Abstract
Hyperbolic geometry provides a natural inductive bias for genomic representation learning, but existing hyperbolic genomic models primarily use Lorentz convolutions to learn local sequence representations, while their residual pathways do not directly aggregate full Lorentz representations. We propose HyCoSeq, a contextual hyperbolic representation learning framework for genomic sequences. HyCoSeq incorporates weighted Lorentzian residual aggregation into multi-curvature Lorentz encoding, allowing full Lorentz representations to participate directly in geometry-consistent local aggregation. It further introduces a bidirectional long short-term memory network that integrates information from both sequence directions to learn contextual relationships among local representations at different positions within a genomic sequence, thereby extending local hyperbolic convolutional encoding to sequence-level contextualized representations. Extensive experiments across diverse genomic tasks show that HyCoSeq outperforms existing hyperbolic baselines and, without large-scale genomic pretraining, achieves competitive performance against substantially larger pretrained DNA language models.
Related
- Successor representation supports structural learning of syllable sequences
- Multimodal Protein Retrieval via Joint Representation Learning from Sequences and Cryo-EM Density Maps
- Neural substrates related to memory consolidation of learning multiple motor sequences during wakeful rest
- Facilitation and Attenuation of Learning to Novel and Familiar Stimuli Through Contextual Mismatch: Experimental Evidence and a Computational Model
- High-dimensional hippocampus representation facilitates learning of low-dimensional abstract representation and supports generalization
BibTeX
@misc{zeng2026hycoseq,
title = {{HyCoSeq: Contextual Hyperbolic Representation Learning for Genomic Sequences}},
author = {Chenhao Zeng and Zhibin Pu and Shufei Ge},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.16925},
eprint = {2609.16925},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.16925},
}
RIS
TY - UNPB
TI - HyCoSeq: Contextual Hyperbolic Representation Learning for Genomic Sequences
AU - Chenhao Zeng
AU - Zhibin Pu
AU - Shufei Ge
PY - 2026
DA - 2026/09/15
PB - arXiv
DO - 10.48550/arxiv.2609.16925
UR - https://arxiv.org/abs/2609.16925
AB - Hyperbolic geometry provides a natural inductive bias for genomic representation learning, but existing hyperbolic genomic models primarily use Lorentz convolutions to learn local sequence representations, while their residual pathways do not directly aggregate full Lorentz representations. We propose HyCoSeq, a contextual hyperbolic representation learning framework for genomic sequences. HyCoSeq incorporates weighted Lorentzian residual aggregation into multi-curvature Lorentz encoding, allowing full Lorentz representations to participate directly in geometry-consistent local aggregation. It further introduces a bidirectional long short-term memory network that integrates information from both sequence directions to learn contextual relationships among local representations at different positions within a genomic sequence, thereby extending local hyperbolic convolutional encoding to sequence-level contextualized representations. Extensive experiments across diverse genomic tasks show that HyCoSeq outperforms existing hyperbolic baselines and, without large-scale genomic pretraining, achieves competitive performance against substantially larger pretrained DNA language models.
ER -