Normal view
-
cs.AI, q-bio.NC updates on arXiv.org
-
SciEvalKit: An Open-source Evaluation Toolkit for Scientific General Intelligence
arXiv:2512.22334v1 Announce Type: new Abstract: We introduce SciEvalKit, a unified benchmarking toolkit designed to evaluate AI models for science across a broad range of scientific disciplines and task capabilities. Unlike general-purpose evaluation platforms, SciEvalKit focuses on the core competencies of scientific intelligence, including Scientific Multimodal Perception, Scientific Multimodal Reasoning, Scientific Multimodal Understanding, Scientific Symbolic Reasoning, Scientific Code Gene
-
(Multiomics OR Omics) AND (Lung OR gastric OR Hepatocellular)
-
Causal relationship of immune cell characteristics in hepatocellular carcinoma: A multi-omics analysis based on Mendelian randomization
Medicine (Baltimore). 2025 Dec 5;104(49):e45942. doi: 10.1097/MD.0000000000045942.ABSTRACTThe tumor immune microenvironment of hepatocellular carcinoma (HCC) is complex, yet the causal relationship between immune cell subpopulations and HCC risk remains incompletely elucidated. This study aims to systematically evaluate the causal association between immune cell subpopulations and HCC using Mendelian randomization (MR) analysis, and to validate the biological mechanisms underlying these associat
Causal relationship of immune cell characteristics in hepatocellular carcinoma: A multi-omics analysis based on Mendelian randomization
Medicine (Baltimore). 2025 Dec 5;104(49):e45942. doi: 10.1097/MD.0000000000045942.
ABSTRACT
The tumor immune microenvironment of hepatocellular carcinoma (HCC) is complex, yet the causal relationship between immune cell subpopulations and HCC risk remains incompletely elucidated. This study aims to systematically evaluate the causal association between immune cell subpopulations and HCC using Mendelian randomization (MR) analysis, and to validate the biological mechanisms underlying these associations through multi-omics data. Bidirectional two-sample MR analysis was performed to examine causal relationships between 731 immune cell subpopulations and HCC. Inverse-variance weighting (IVW) served as the primary analysis method, with robustness validation through Bayesian weighted MR (BWMR) and machine learning algorithms. Therefore, for significantly associated immune subpopulations, independent analyses of gene expression, prognosis, and tumor immune microenvironment were conducted using HCC data from the Cancer Genome Atlas (TCGA) LIHC cohort. MR analysis and validation identified 21 immune cell subpopulations with significant causal associations to HCC risk. Among these, 12 were identified as risk factors, and 9 as protective factors. Validation in the TCGA cohort revealed that risk-associated immune subpopulations were predominantly enriched for markers of T cell exhaustion and immunosuppressive microenvironments, whereas protective subpopulations likely represented a distinct regulatory B cell subset whose function was associated with the anti-inflammatory factor interleukin-10. This study genetically confirms that specific immune cell functional subpopulations constitute causal risk factors for HCC. These subpopulations exert their effects by shaping distinct tumor immune microenvironments. These findings provide novel mechanisms for understanding the immunopathogenesis of HCC and identify potential targets for developing novel immune intervention strategies.
PMID:41366997 | DOI:10.1097/MD.0000000000045942
-
cs.AI, q-bio.NC updates on arXiv.org
-
AI Application in Anti-Money Laundering for Sustainable and Transparent Financial Systems
arXiv:2512.06240v1 Announce Type: new Abstract: Money laundering and financial fraud remain major threats to global financial stability, costing trillions annually and challenging regulatory oversight. This paper reviews how artificial intelligence (AI) applications can modernize Anti-Money Laundering (AML) workflows by improving detection accuracy, lowering false-positive rates, and reducing the operational burden of manual investigations, thereby supporting more sustainable development. It fu
AI Application in Anti-Money Laundering for Sustainable and Transparent Financial Systems
-
cs.AI, q-bio.NC updates on arXiv.org
-
SciAgent: A Unified Multi-Agent System for Generalistic Scientific Reasoning
arXiv:2511.08151v2 Announce Type: replace Abstract: Recent advances in large language models have enabled AI systems to achieve expert-level performance on domain-specific scientific tasks, yet these systems remain narrow and handcrafted. We introduce SciAgent, a unified multi-agent system designed for generalistic scientific reasoning-the ability to adapt reasoning strategies across disciplines and difficulty levels. SciAgent organizes problem solving as a hierarchical process: a Coordinator A
SciAgent: A Unified Multi-Agent System for Generalistic Scientific Reasoning
-
cs.AI, q-bio.NC updates on arXiv.org
-
SciAgent: A Unified Multi-Agent System for Generalistic Scientific Reasoning
arXiv:2511.08151v1 Announce Type: new Abstract: Recent advances in large language models have enabled AI systems to achieve expert-level performance on domain-specific scientific tasks, yet these systems remain narrow and handcrafted. We introduce SciAgent, a unified multi-agent system designed for generalistic scientific reasoning-the ability to adapt reasoning strategies across disciplines and difficulty levels. SciAgent organizes problem solving as a hierarchical process: a Coordinator Agent
SciAgent: A Unified Multi-Agent System for Generalistic Scientific Reasoning
-
cs.AI, q-bio.NC updates on arXiv.org
-
VaultGemma: A Differentially Private Gemma Model
arXiv:2510.15001v2 Announce Type: replace-cross Abstract: We introduce VaultGemma 1B, a 1 billion parameter model within the Gemma family, fully trained with differential privacy. Pretrained on the identical data mixture used for the Gemma 2 series, VaultGemma 1B represents a significant step forward in privacy-preserving large language models. We openly release this model to the community