arXiv subjects

Douglas Jiang

Publications and source records attributed to Douglas Jiang.

3 recordsLinked to original sources

arXiv2026

NeuroPlastic: A Plasticity-Modulated Optimizer for Biologically Inspired Learning Dynamics

Optimization algorithms are fundamental to modern deep learning, yet most widely used methods rely on update rules based primarily on local gradient statistics. We introduce NeuroPlastic, a plasticity-modulated optimizer that augments gradient-based updates with an adaptive multi-signal modulation mechanism inspired by multi-factor synaptic plasticity, a concept from neurobiology. NeuroPlastic dynamically scales gradient updates using interacting components that capture gradient, activity-like, and memory-like statistics, forming a lightweight modulation layer compatible with standard deep learning training pipelines. Across image classification benchmarks, NeuroPlastic consistently improves over a controlled gradient-only ablation, with more pronounced gains on the Fashion-MNIST benchmark and in reduced-data regimes. In transfer experiments on CIFAR-10 with ResNet-18, the method remains stable and competitive without retuning. These results suggest that multi-signal plasticity-inspired modulation can provide a useful extension to conventional gradient-driven optimization, particularly when learning signals are limited or noisy, and offer a promising direction for gradient-based methods in deep learning.

cs.LG↗

arXiv2025

Contrastive Learning Enhances Language Model Based Cell Embeddings for Low-Sample Single Cell Transcriptomics

Large language models (LLMs) have shown strong ability in generating rich representations across domains such as natural language processing and generation, computer vision, and multimodal learning. However, their application in biomedical data analysis remains nascent. Single-cell transcriptomic profiling is essential for dissecting cell subtype diversity in development and disease, but rare subtypes pose challenges for scaling laws. We present a computational framework that integrates single-cell RNA sequencing (scRNA-seq) with LLMs to derive knowledge-informed gene embeddings. Highly expressed genes for each cell are mapped to NCBI Gene descriptions and embedded using models such as text-embedding-ada-002, BioBERT, and SciBERT. Applied to retinal ganglion cells (RGCs), which differ in vulnerability to glaucoma-related neurodegeneration, this strategy improves subtype classification, highlights biologically significant features, and reveals pathways underlying selective neuronal vulnerability. More broadly, it illustrates how LLM-derived embeddings can augment biological analysis under data-limited conditions and lay the groundwork for future foundation models in single-cell biology.

q-bio.GN↗

arXiv2025

Bridging Large Language Models and Single-Cell Transcriptomics in Dissecting Selective Motor Neuron Vulnerability

Understanding cell identity and function through single-cell level sequencing data remains a key challenge in computational biology. We present a novel framework that leverages gene-specific textual annotations from the NCBI Gene database to generate biologically contextualized cell embeddings. For each cell in a single-cell RNA sequencing (scRNA-seq) dataset, we rank genes by expression level, retrieve their NCBI Gene descriptions, and transform these descriptions into vector embedding representations using large language models (LLMs). The models used include OpenAI text-embedding-ada-002, text-embedding-3-small, and text-embedding-3-large (Jan 2024), as well as domain-specific models BioBERT and SciBERT. Embeddings are computed via an expression-weighted average across the top N most highly expressed genes in each cell, providing a compact, semantically rich representation. This multimodal strategy bridges structured biological data with state-of-the-art language modeling, enabling more interpretable downstream applications such as cell-type clustering, cell vulnerability dissection, and trajectory inference.

q-bio.GN↗