Page 85 of 130 • 1560 Total Articles

createLiveAI

Continue exploring the latest AI breakthroughs, technology insights, and industry analysis. Page 85 of our comprehensive AI news collection.

📰 Latest Intelligence

Showing 12 articles on page 85 of 130

Live feed
Business
📄 MarkTechPost

Baidu Open Sources ERNIE 4.5: LLM Series Scaling from 0.3B to 424B Parameters

Baidu has open-sourced its ERNIE 4.5 series, a comprehensive family of foundation models designed for advanced language understanding, reasoning, and generation, with variants ranging from 0.3 billion to 424 billion parameters. The models incorporate both dense and Mixture-of-Experts (MoE) architectures, with the largest MoE models efficiently scaling parameters by activating only a subset of experts per token, thus balancing model capacity and computational efficiency. Built on Baidus multi-stage pretraining pipeline, ERNIE 4.5 models are trained on a vast corpus of 5

Research
📄 MarkTechPost

OMEGA: A Structured Math Benchmark to Probe the Reasoning Limits of LLMs

The article introduces OMEGA, a structured math benchmark designed to evaluate the reasoning capabilities of large language models (LLMs) in mathematical problem-solving, particularly focusing on their ability to generalize beyond learned patterns. While models like DeepSeek-R1 demonstrate proficiency in Olympiad-level mathematics through chain-of-thought reasoning, their reliance on limited techniques such as rule repetition and pattern recognition hampers their performance on more complex, creative tasks that require genuine mathematical insight. Current evaluation methods and datasets are insufficient for thoroughly assessing the true reasoning skills of LLMs, as they often conflate surface-level pattern matching with

General
📄 MarkTechPost

Building Advanced Multi-Agent AI Workflows by Leveraging AutoGen and Semantic Kernel

The tutorial demonstrates the integration of AutoGen and Semantic Kernel with Googles Gemini Flash model by developing GeminiWrapper and SemanticKernelGeminiPlugin classes that enable multi-agent orchestration leveraging Geminis generative capabilities. This approach configures specialist agents using AutoGens ConversableAgent API alongside Semantic Kernels decorated functions to perform tasks such as text analysis, summarization, code review, and creative problem-solving, resulting in an advanced, adaptable AI assistant framework.

Google AI
Read More
Research
📄 Towards Data Science

Lessons Learned After 6.5 Years Of Machine Learning

After 6.5 years of extensive research and experimentation, significant insights have been gained into the evolving landscape of machine learning, emphasizing the importance of deep work, emerging trends, and data-driven approaches. This period has highlighted critical developments in understanding model performance, optimization techniques, and the integration of large-scale data to enhance AI capabilities, paving the way for more robust and efficient machine learning systems.

Machine Learning
Read More
Business
📄 MarkTechPost

UC San Diego Researchers Introduced Dex1B: A Billion-Scale Dataset for Dexterous Hand Manipulation in Robotics

UC San Diego researchers have introduced Dex1B, a large-scale dataset comprising one billion samples designed to advance dexterous hand manipulation in robotics. This dataset aims to address the significant challenge of collecting diverse, high-quality data necessary for training effective control models, which has historically limited progress due to the complexity of robotic hands and the limitations of existing data collection methods like human demonstrations and reinforcement learning. The development of Dex1B represents a critical step toward enabling more robust and generalizable learning-based approaches for dexterous manipulation, leveraging extensive data to improve the physical feasibility and diversity of generated manipulation behaviors. This

Robotics
Read More
Business
📄 MarkTechPost

Tencent Open Sources Hunyuan-A13B: A 13B Active Parameter MoE Model with Dual-Mode Reasoning and 256K Context

Tencent's Hunyuan team has unveiled Hunyuan-A13B, an open-source large language model leveraging a sparse Mixture-of-Experts (MoE) architecture that efficiently balances performance and computational cost by activating only 13 billion parameters out of 80 billion during inference. The model incorporates advanced features such as Grouped Query Attention (GQA), a 256K token context window, and a dual-mode reasoning framework that switches between fast and slow thinking modes, enhancing its capability for complex reasoning and long-context tasks. Built with a fine-grained MoE design, Hunyuan-A13

Transformers
Read More
Technology
📄 MarkTechPost

Getting started with Gemini Command Line Interface (CLI)

Google has introduced the Gemini CLI, a sophisticated command-line tool that integrates multimodal AI capabilities directly into developer workflows. This tool enables querying and editing extensive codebases beyond traditional token limits, generating applications from visual inputs such as PDFs and sketches, and automating complex operational tasks like pull request management and rebasing, thereby significantly enhancing productivity and efficiency. Built on Googles latest AI models, Gemini CLI also facilitates seamless integration with external media generation tools like Imagen, Veo, and Lyria, and leverages Google Search within the terminal environment. Its deployment requires Node.js installation, followed by a straightforward npm

Google AI
Read More

Page 85 of 130 • Showing articles 1009-1020 of 1560