Generative Artificial Intelligence: A Historical Perspective

National Science Review, Год журнала: 2025, Номер 12(5)

Опубликована: Фев. 21, 2025

ABSTRACT Generative artificial intelligence (GAI) has recently achieved significant success, enabling anyone to create texts, images, videos and even computer codes while providing insights that might not be possible with traditional tools. To stimulate future research, this work provides a brief summary of the ongoing historical developments in GAI over past 70 years. The achievements are grouped into four categories: (i) rule-based generative systems follow specialized rules instructions, (ii) model-based algorithms produce new content based on statistical or graphical models, (iii) deep methodologies utilize neural networks learn how generate from data (iv) foundation models trained extensive datasets capable performing variety tasks. This paper also reviews successful applications identifies open challenges posed by remaining issues. In addition, describes potential research directions aimed at better utilizing, understanding harnessing technologies.

Язык: Английский

---

Extracting interaction kernels for many-particle systems by a two-phase approach

Physics of Fluids, Год журнала: 2025, Номер 37(3)

Опубликована: Март 1, 2025

This paper presents a two-phase method for learning interaction kernels of stochastic many-particle systems. After transforming trajectories every particle into the density function by kernel estimation method, first phase our approach combines importance sampling with an adaptive threshold strategy to identify key terms in function, while second uses whole dataset refine coefficients. During implementation mean-field equation plays role reformulating task extracting learnable regression problem. We demonstrate outstanding performance through extensive numerical examples, including interacting systems cubic potential, power-law repulsion-attraction double-well and piecewise linear as well two-dimensional radially symmetric potential.

Язык: Английский

---

Large language model agents can use tools to perform clinical calculations

npj Digital Medicine, Год журнала: 2025, Номер 8(1)

Опубликована: Март 17, 2025

Large language models (LLMs) can answer expert-level questions in medicine but are prone to hallucinations and arithmetic errors. Early evidence suggests LLMs cannot reliably perform clinical calculations, limiting their potential integration into workflows. We evaluated ChatGPT's performance across 48 medical calculation tasks, finding incorrect responses one-third of trials (n = 212). then assessed three forms agentic augmentation: retrieval-augmented generation, a code interpreter tool, set task-specific tools (OpenMedCalc) 10,000 trials. Models with access showed the greatest improvement, LLaMa GPT-based demonstrating 5.5-fold (88% vs 16%) 13-fold (64% 4.8%) reduction responses, respectively, compared unimproved models. Our findings suggest that machine-readable, may help overcome LLMs' limitations calculations.

Язык: Английский

---

An AI Researchers' Perspective: At the Crossroad of LLMs, Agent-Based Modelling, and Complex Systems

Physics of Life Reviews, Год журнала: 2025, Номер unknown

Опубликована: Март 1, 2025

Язык: Английский

---

Exploring Large Language Model-Driven Agents for Environment-Aware Spatial Interactions and Conversations in Virtual Reality Role-Play Scenarios

Опубликована: Март 8, 2025

Язык: Английский

---

Deferred prefill for throughput maximization in LLM inference

Опубликована: Март 30, 2025

Язык: Английский

---

CCA: collaborative competitive agents for image editing

Frontiers of Computer Science, Год журнала: 2025, Номер 19(11)

Опубликована: Апрель 5, 2025

Язык: Английский

---

MRAgent: an LLM-based automated agent for causal knowledge discovery in disease via Mendelian randomization

Briefings in Bioinformatics, Год журнала: 2025, Номер 26(2)

Опубликована: Март 1, 2025

Abstract Understanding causality in medical research is essential for developing effective interventions and diagnostic tools. Mendelian Randomization (MR) a pivotal method inferring through genetic data. However, MR analysis often requires pre-identification of exposure-outcome pairs from clinical experience or literature, which can be challenging to obtain. This poses difficulties clinicians investigating causal factors specific diseases. To address this, we introduce MRAgent, an innovative automated agent leveraging Large Language Models (LLMs) enhance knowledge discovery disease research. MRAgent autonomously scans scientific discovers potential pairs, performs inference using extensive Genome-Wide Association Study We conducted both human evaluations compare different LLMs operating provided proof-of-concept case demonstrate the complete workflow. MRAgent’s capability conduct large-scale analyses represents significant advancement, equipping researchers with robust tool exploring validating relationships complex Our code public at https://github.com/xuwei1997/MRAgent.

Язык: Английский

---

Application of large language models in medicine

Nature Reviews Bioengineering, Год журнала: 2025, Номер unknown

Опубликована: Апрель 7, 2025

Язык: Английский

---