Equipping Llama with Google Query API for Improved Accuracy and Reduced Hallucination

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: March 6, 2024

Abstract This study investigates the integration of Llama 2 7b large language model (LLM) with Google Query API to enhance its accuracy and reduce hallucination instances. By leveraging real-time internet data, we aimed address limitations static training datasets improve model's performance across various processing tasks. The methodology involved augmenting 7b's architecture incorporate dynamic data retrieval from API, followed by an evaluation impact on reduction using BIG-Bench benchmark. results indicate significant improvements in both reliability, demonstrating effectiveness integrating LLMs external sources. not only marks a substantial advancement capabilities but also raises important considerations regarding bias, privacy, ethical use internet-sourced information. study's findings contribute ongoing discourse enhancing LLMs, suggesting promising direction for future research development artificial intelligence.

Language: Английский

---

Stance Detection with Collaborative Role-Infused LLM-Based Agents

Proceedings of the International AAAI Conference on Web and Social Media, Journal Year: 2024, Volume and Issue: 18, P. 891 - 903

Published: May 28, 2024

Stance detection automatically detects the stance in a text towards target, vital for content analysis web and social media research. Despite their promising capabilities, LLMs encounter challenges when directly applied to detection. First, demands multi-aspect knowledge, from deciphering event-related terminologies understanding expression styles platforms. Second, requires advanced reasoning infer authors' implicit viewpoints, as stances are often subtly embedded rather than overtly stated text. To address these challenges, we design three-stage framework COLA (short Collaborative rOle-infused LLM-based Agents) which designated distinct roles, creating collaborative system where each role contributes uniquely. Initially, multidimensional stage, configure act linguistic expert, domain specialist, veteran get multifaceted of texts, thus overcoming first challenge. Next, reasoning-enhanced debating potential stance, designate specific agent advocate it, guiding LLM detect logical connections between features tackling second Finally, conclusion final decision maker consolidates prior insights determine stance. Our approach avoids extra annotated data model training is highly usable. We achieve state-of-the-art performance across multiple datasets. Ablation studies validate effectiveness handling Further experiments have demonstrated explainability versatility our approach. excels usability, accuracy, effectiveness, versatility, highlighting its value.

Language: Английский

---

Integrating artificial intelligence with mechanistic epidemiological modeling: a scoping review of opportunities and challenges

Nature Communications, Journal Year: 2025, Volume and Issue: 16(1)

Published: Jan. 10, 2025

Integrating prior epidemiological knowledge embedded within mechanistic models with the data-mining capabilities of artificial intelligence (AI) offers transformative potential for modeling. While fusion AI and traditional approaches is rapidly advancing, efforts remain fragmented. This scoping review provides a comprehensive overview emerging integrated applied across spectrum infectious diseases. Through systematic search strategies, we identified 245 eligible studies from 15,460 records. Our highlights practical value models, including advances in disease forecasting, model parameterization, calibration. However, key research gaps remain. These include need better incorporation realistic decision-making considerations, expanded exploration diverse datasets, further investigation into biological socio-behavioral mechanisms. Addressing these will unlock synergistic modeling to enhance understanding dynamics support more effective public health planning response. Artificial has improve diseases by incorporating data sources complex interactions. Here, authors conduct use summarise methodological advancements identify gaps.

Language: Английский

---

MetaCity: Data-driven sustainable development of complex cities

The Innovation, Journal Year: 2025, Volume and Issue: unknown, P. 100775 - 100775

Published: Jan. 1, 2025

Language: Английский

---

Agentic Large Language Models for Healthcare: Current Progress and Future Opportunities

Medicine Advances, Journal Year: 2025, Volume and Issue: unknown

Published: March 3, 2025

Language: Английский

---

Higher Performance of Mistral Large on MMLU Benchmark through Two-Stage Knowledge Distillation

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: May 14, 2024

Large language models (LLM) have undergone significant transformations through the application of knowledge distillation techniques aimed at enhancing performance on complex benchmarks like MMLU. The research detailed herein introduces a novel two-stage process designed to refine capabilities Mistral Large, resulting in marked improvements both accuracy and contextual understanding. Initially, model undergoes teacher-student training phase where high-performing teacher imparts its less student model, utilizing soft hard target methods optimize transfer. This is followed by specialized refinement stage further fine-tuned tasks that require advanced cognitive skills, specifically tailored challenges presented MMLU benchmark. Quantitative results indicate substantial increase across various within benchmark, while qualitative analyses show enhanced linguistic sophistication relevance model's responses. Comparisons with baseline confirm distilled significantly outperforms traditional approaches, setting new standards for models. implications our findings suggest structured can fundamentally alter development trajectory models, making them more efficient effective diverse applications. study's approach offers scalable framework future enhancements has potential influence wide range applications artificial intelligence, from automated conversational systems sophisticated analytical tools.

Language: Английский

---

LLM-AIDSim: LLM-Enhanced Agent-Based Influence Diffusion Simulation in Social Networks

Systems, Journal Year: 2025, Volume and Issue: 13(1), P. 29 - 29

Published: Jan. 3, 2025

This paper introduces an LLM-Enhanced Agent-Based Influence Diffusion Simulation (LLM-AIDSim) framework that integrates large language models (LLMs) into agent-based modelling to simulate influence diffusion in social networks. The proposed enhances traditional by allowing agents generate language-level responses, providing deeper insights user agent interactions. Our addresses the limitations of probabilistic simulating realistic, context-aware behaviours response public statements. Using real-world news topics, we demonstrate effectiveness LLM-AIDSim topic evolution and tracking discourse, validating its ability replicate key aspects information propagation. experimental results highlight role shaping collective discussions, revealing that, over time, narrows focus conversations around a few dominant topics. We further analyse regional differences clustering across three cities, Sydney, Auckland, Hobart, how demographics, income, education levels dominance. work underscores potential as decision-support tool for strategic communication, enabling organizations anticipate understand sentiment trends.

Language: Английский

---

Agent-based modeling: Insights into consumer behavior, urban dynamics, grid management, and market interactions

Energy Strategy Reviews, Journal Year: 2025, Volume and Issue: 57, P. 101613 - 101613

Published: Jan. 1, 2025

Language: Английский

---

Integrating large language model and digital twins in the context of industry 5.0: Framework, challenges and opportunities

Robotics and Computer-Integrated Manufacturing, Journal Year: 2025, Volume and Issue: 94, P. 102982 - 102982

Published: Feb. 10, 2025

Language: Английский

---

Generative AI-Based Platform for Deliberate Teaching Practice: A Review and a Suggested Framework

Education Sciences, Journal Year: 2025, Volume and Issue: 15(4), P. 405 - 405

Published: March 24, 2025

This paper begins with a comprehensive review of the deliberate teaching practice literature related to generative AI training platforms. It then introduces conceptual framework for AI-powered system designed simulate dynamic classroom environments, allowing teachers engage in repeated, goal-oriented sessions. Leveraging recent advances large language models (LLMs) and multiagent systems, platform features virtual student agents configured demonstrate varied learning styles, prior knowledge, behavioral traits. In parallel, mentor agents—built upon same technology—continuously provide feedback, enabling adapt their strategies real time. By offering an accessible, controlled space skill development, this addresses challenge scaling personalizing teacher training. Grounded pedagogical theory supported by emerging capabilities, proposed enables educators refine methods diverse contexts through iterative practice. A detailed outline system’s main components, including agent configuration, interaction workflows, feedback loop, sets stage more personalized, high-quality experiences, contributes evolving field AI-mediated environments.

Language: Английский

---