Google’s AI will auto-generate ads
Plus: LLMs to think like human. ToolLLM masters 16k+ real-word APIs.
Hello Engineering Leaders and AI Enthusiasts!
Welcome to the 75th edition of The AI Edge newsletter. This edition brings you “Google’s AI will auto-generate ads.”
And a huge shoutout to our incredible readers. Your support is invaluable to us!😊
In today’s edition:
🤖 Google’s AI will auto-generate ads
🧠LLMs to think more like a human for answer quality
🛠️ ToolLLM masters 16k+ real-word APIs
💡 Knowledge Nugget: Large Language Models and Nearest Neighbors by
Let’s go!
Google’s AI will auto-generate ads
Google Ads has introduced a new feature that uses AI to generate advertisements on its platform automatically. The feature utilizes Large Language Models and generative AI to create campaign workflows based on prompts from marketers.
Google Ads can analyze landing pages, successful queries, and approved headlines to generate new creatives. The company also highlighted its commitment to privacy and introduced enhanced privacy features like Privacy Sandbox.
Why does this matter?
Using LLMs and Generative AI, this AI tool for auto-generated ads will save time, ensure privacy, and empower small businesses to leverage AI. Integrating generative AI in content creation also promises exciting possibilities beyond advertising.
LLMs to think more like a human for answer quality
This research introduces "Skeleton-of-Thought" (SoT), a method to decrease the generation latency of large language models. SoT guides LLMs first to generate the skeleton of the answer and then complete the contents of each skeleton point in parallel.
This approach provides significant speed-up (up to 2.39x across 11 different LLMs) and can potentially improve answer quality regarding diversity and relevance. SoT is an initial attempt at optimizing LLMs for efficiency and encouraging them to think more like humans for better answers.
Research by: Microsoft Research And Department of Electronic Engineering, Tsinghua University.
Why does this matter?
By emulating human-like thinking processes, LLMs can deliver more natural and contextually appropriate answers, enhancing their practical applications across various domains, such as NLP, customer support, and information retrieval. This advancement brings us closer to creating AI systems that can interact with users more effectively, making them more valuable tools in our everyday lives.
ToolLLM masters 16k+ real-word APIs
ToolLLM is a framework that enhances the tool-use capabilities of open-source LLMs by training them to follow human instructions to use external tools (APIs). The framework includes a dataset called ToolBench, which contains instructions for using over 16,000 real-world APIs.
A depth-first search-based decision tree (DFSDT) is used to improve the planning and reasoning capabilities of the LLMs. An automatic evaluator called ToolEval is also developed to assess the performance of the LLMs. The results show that the trained LLM, ToolLLaMA, can execute complex instructions and generalize to unseen APIs, performing comparably to closed-source LLMs like ChatGPT.
Why does this matter?
ToolLLM, can execute complex instructions and perform comparably to closed-source models like ChatGPT. And it bridges the gap between language models and practical tool usage, making them more versatile and valuable for various applications.
Knowledge Nugget: Large Language Models and Nearest Neighbors
This thoughtful article by
explores using nearest-neighbor methods in the context of large language models. He highlights the beauty of simple techniques like nearest neighbor algorithms and discusses their potential for making significant contributions based on foundational or classic approaches. Nearest neighbor algorithms, though not as popular as before, are still widely used in practice, and the k-Nearest Neighbor algorithm is recommended as a benchmark for predictive performance in classification projects.
(A k-nearest neighbor classifier with k=5.)
The article also provides additional resources on improving computational performance for nearest-neighbor methods.
Why does this matter?
This article showcases a simple yet effective method. It demonstrates that foundational techniques can still be competitive in low-resource scenarios and highlights the potential of alternative approaches.
What Else Is Happening❗
📱 ChatGPT Android app is now available in all countries. (Link)
💻 Dell and Nvidia join hands for Gen AI solutions. (Link)
🗣️ Google will update Assistant with similar tech like ChatGPT. (Link)
🌟 Incredible response of Liama 2, 150K+ downloads in just a week! (Link)
👨💼 Read how OpenAI’s Sam Altman links 2 hot tech trends with his new Worldcoin! (Link)
🔫 ZeroEyes can detect guns in public spaces before shootings occur. (Link)
🛠️ Trending Tools
BgRem: Create unique images, art, and videos with AI. Remove unwanted objects or backgrounds. Try it for free!
MyLikenessAI: Monetize and control your likeness in AI. Free engine integration for GenAI companies.
RankWizard: Programmatic SEO writer that writes bulk content using AI. Backed by real research for fresh and accurate information
eBookerAI: Generate ebooks and audiobooks with AI. Refine content in a WYSIWYG editor. Export to PDF, EPUB, or MP3.
42Signals: AI-powered ecommerce insights tool for brands to track, optimize, and grow their online presence.
Double Subtitles: Mobile-first AI video editor with auto-transcription. Separate subtitles by word or sentence.
RewriteX: Make your text easier to understand with AI. Rewrite text directly without leaving your text window.
Robofy AI: Create a smart and powerful chatbot from your website’s content in minutes. Answer visitor queries in seconds.
😁 Tickle Tuesday
Wow! The meme-scape evolution is here.
Image-to-video AI will take the fun to new heights!
That's all for now!
By subscribing to The AI Edge, you join the company of readers from Moody’s, Vonage, Voya, WEHI, Cox, INSEAD, and other reputable organizations.
Thanks for reading, and see you tomorrow. 😊