日立製作所は、生成AIの活用を促進するためにGenerative AIセンターを設立し、ITとOTの両面からAIを活用しています。2024年度には、社内の業務改革と外販向けのソリューション提供を計画しており、さまざまな取り組みが進行中です。生成AIの責任ある活用のための社内ガイドラインや人材育成にも力を入れています。
Tag: IT AI ML
「LINE WORKS DAY 24」で大成建設が活用事例を発表、現場の職人への教育コストをかけず生産性を向上
LINE WORKSは2024年5月28日に「LINE WORKS DAY 24」を開催し、増田氏と大竹氏が登壇。最新動向やAI機能の強化、新製品発表などが話題に。声と文字を融合した新アプリも開発中。大成建設ではLINE WORKSを導入し、安全管理や情報共有で大きな効果を発揮している。
What We Learned from a Year of Building with LLMs (Part I)
The authors of this post will be hosting a virtual event on June 20th and a Generative AI Success Stories Superstream on June 12th. The series explores lessons and methodologies essential for building products based on Large Language Models (LLMs), focusing on tactical, operational, and strategic aspects. The authors have diverse experience and aim to provide practical guidance based on their firsthand experiences.
Small Language Models (SLMs): The Next Frontier For The Enterprise
Microsoft’s CEO, Satya Nadella, praised generative AI progress and emphasized the strategic value of small language models (SLMs) alongside large language models (LLMs). While LLMs offer broad capabilities, SLMs tailored to enterprise needs provide enhanced accuracy, cost-effectiveness, and security. Despite challenges, SLMs show promise as a valuable tool for enterprise generative AI.
DeepSeek-V2: An Efficient and Economical Mixture-of-Experts LLM
Large Language Models (LLMs) have advanced text generation but face computational challenges. DeepSeek-V2, with 236 billion parameters and efficient architecture, balances performance and efficiency. Multi-head Latent Attention reduces memory usage, and DeepSeekMoE optimizes expert utilization. The model aligns with human preferences and outperforms its predecessor, demonstrating its effectiveness across various domains and languages.
How to deploy LLM-Powered Algorithmic Trading Agents?
The world of trading comprises three types: Category 1 – uninformed investors, Category 2 – systematic traders, and Category 3 – algorithmic traders. The ultimate goal is to make algorithmic trading accessible to everyone through AI-powered agents. NexusTrade.io is developing semi-automated and fully-autonomous trading agents to achieve this, while considering associated risks.
Meta AI Review: A Convenient but Unimpressive Virtual Assistant
Meta AI, with integration into major platforms, offers text and image generation. However, it faces criticism for frequent hallucinations and the need for clarifications. The AI excels in shopping and recipe assistance, but struggles in research and travel planning. While convenient, it falls short compared to competitors, with mixed performance across different tasks.
ChatGPT 4 Review: A Smarter AI Chatbot, but It’ll Cost You
The review of ChatGPT 4.0 highlights its benefits, such as improved comprehension and nuanced responses, with potential downsides including slow response times and a lack of internet connection. Despite limitations, it excels in providing accurate information for product research, travel recommendations, and complex inquiries. The $20 monthly fee may be justified for specialized and in-depth queries.
Maximizing AI Potential: GPT-4o Unlocks New Dimensions in Human-AI Interaction and Innovation
GPT-4o marks a significant shift in AI, enabling real-time conversation, surpassing previous boundaries, and enhancing user experiences across various industries. It democratizes AI access, enhances safety, and promotes collaboration, while empowering developers with affordable AI tools. Continuous innovation and future enhancements point to a future of extensive AI collaboration, revolutionizing our environment.
Unlocking the Power of ChatGPT for Enhanced Workplace Productivity
The workplace benefits of modern technology are clear, with ChatGPT standing out as a versatile tool leveraging artificial intelligence. It can assist with communication, content creation, data analysis, coding, and research. Integrating it effectively into workflows unlocks new levels of efficiency and collaboration. Clear prompts and responsible usage are crucial for maximizing its utility.
This AI newsletter is all you need #99
This week saw the release of Alphafold-3 and GPT-4o, advancing AI capabilities in protein prediction and multimodal interactions. Alphafold-3 can model folding patterns and chemical structures across biomolecules, while GPT-4o enables real-time interaction with speech, images, and video. These advancements have significant implications for drug development and AI accessibility.
Here’s how Apple can compete against OpenAI and Google
Since Siri’s debut in 2011, Apple has faced tough competition in AI, with OpenAI and Google launching advanced voice assistants. Apple may license OpenAI’s GPT-4o or Google’s Gemini to enhance Siri’s capabilities. Additionally, Apple’s potential for on-device training and personalized neural networks could revolutionize AI. Leveraging user data and federated learning may further boost Siri’s effectiveness.
Generative AI on AWS with Bedrock
Amazon Bedrock is a managed service providing access to high-performing foundation models from leading AI companies. Users can experiment with different models, fine-tune them, and integrate knowledge bases and agents for enhanced features. It also offers the capability to create custom models and train them using specific datasets to improve performance for particular tasks.
The rise of Generative AI-driven design patterns
Generative AI has sparked a revolution in feature design, shifting from chatbots to diverse applications. Notion AI and Grammarly lead the charge, enhancing content creation and message personalization. Summarization and insight recognition are also growing, as AI predicts outcomes and fosters empathy. The landscape is evolving rapidly, opening new possibilities for innovation.