arXiv subjects

Jay Shin

Publications and source records attributed to Jay Shin.

3 recordsLinked to original sources

arXiv2026

TheBioCollection: Unified Pre-Training Scale LLM Corpus for Biology

The push toward large language models for biology (BioLM) has created a need for training corpora that can endow models with a genuine understanding of biology. However, existing biological resources, such as molecular databases, protein repositories, genomic annotations, single-cell atlases, and pathway databases, are scattered across heterogeneous formats and remain unorganized into a cohesive corpus for language model training. We present TheBioCollection, a 52.6B-token pre-training-scale corpus that converts these disparate resources into a unified, training-ready form spanning small molecules, proteins, genomic sequences, cells, and pathways. Beyond consolidating existing data, TheBioCollection enriches each record with tool-computed biological properties and introduces new instruction tasks for capabilities that current corpora barely cover. We pair the corpus with TheBioCollection-Eval, a matched suite probing recognition, generation, and prediction across molecular, protein, genomic, cellular, and cross-domain settings. Holding the base Gravity-16B-A3B architecture fixed, training on TheBioCollection more than doubles its overall score on TheBioCollection-Eval with gains in every domain, while leaving general linguistic ability nearly intact.

q-bio.QM↗

arXiv2025

MUG-Eval: A Proxy Evaluation Framework for Multilingual Generation Capabilities in Any Language

Evaluating text generation capabilities of large language models (LLMs) is challenging, particularly for low-resource languages where methods for direct assessment are scarce. We propose MUG-Eval, a novel framework that evaluates LLMs' multilingual generation capabilities by transforming existing benchmarks into conversational tasks and measuring the LLMs' accuracies on those tasks. We specifically designed these conversational tasks to require effective communication in the target language. Then, we simply use task success rate as a proxy for successful conversation generation. Our approach offers two key advantages: it is independent of language-specific NLP tools or annotated datasets, which are limited for most languages, and it does not rely on LLMs-as-judges, whose evaluation quality degrades outside a few high-resource languages. We evaluate 8 LLMs across 30 languages spanning high, mid, and low-resource categories, and we find that MUG-Eval correlates strongly with established benchmarks ($r$ > 0.75) while enabling standardized comparisons across languages and models. Our framework provides a robust and resource-efficient solution for evaluating multilingual generation that can be extended to thousands of languages.

cs.CL↗

arXiv2018

The Human Cell Atlas White Paper

The Human Cell Atlas (HCA) will be made up of comprehensive reference maps of all human cells - the fundamental units of life - as a basis for understanding fundamental human biological processes and diagnosing, monitoring, and treating disease. It will help scientists understand how genetic variants impact disease risk, define drug toxicities, discover better therapies, and advance regenerative medicine. A resource of such ambition and scale should be built in stages, increasing in size, breadth, and resolution as technologies develop and understanding deepens. We will therefore pursue Phase 1 as a suite of flagship projects in key tissues, systems, and organs. We will bring together experts in biology, medicine, genomics, technology development and computation (including data analysis, software engineering, and visualization). We will also need standardized experimental and computational methods that will allow us to compare diverse cell and tissue types - and samples across human communities - in consistent ways, ensuring that the resulting resource is truly global. This document, the first version of the HCA White Paper, was written by experts in the field with feedback and suggestions from the HCA community, gathered during recent international meetings. The White Paper, released at the close of this yearlong planning process, will be a living document that evolves as the HCA community provides additional feedback, as technological and computational advances are made, and as lessons are learned during the construction of the atlas.

q-bio.TO↗