Issue 47

Highlights

Introducing improvements to the fine-tuning API and expanding our custom models program

We’re adding new features to help developers have more control over fine-tuning and announcing new ways to build custom models with OpenAI.

Spotify launches personalized AI playlists that you can build using prompts | TechCrunch

AI playlists will initially become available to users on Android and iOS devices in the U.K. and Australia and will be iterated on in the months ahead.

Learning from interaction with Microsoft Copilot (web)

Microsoft researchers are taking a comprehensive and dynamic approach to help Copilot (web) continuously learn from interaction and feedback, improving the AI system and making it increasingly useful for consumers.

‘Time is running out’: can a future of undetectable deepfakes be avoided?

With more than 4,000 shares, 20,000 comments, and 100,000 reactions on Facebook, the photo of the elderly woman, sitting behind her homemade 122nd birthday cake, has unquestionably gone viral. “I started decorating cakes from five years old,” the caption reads, “and I can’t wait to grow my baking journey.”

Specials

Build GenAI Apps with Microsoft & Win!

Do you want to build generative AI applications with models like GPT-4 Turbo with Vision, win cash prizes, and get recognition at Microsoft Build 2024? Get started with Microsoft's virtual GenAI Hackathon open from April 1 - May 6th for US based participants. The first 500 teams to sign up get $1000 in Azure credits each (T&Cs apply).

Video

LLMOps with Azure - Global AI Notes

The LLMOps repository offers a structured pathway for organizations to navigate the complexities of implementing and scaling LLM applications. By understandi...

Nitya Narasimhan - Humans in AI

Join us as Nitya Narasimhan, Senior Cloud Advocate on the AI Advocacy team at Microsoft, shares her insights on the future of AI and its potential for revolu...

AI Show On Demand | Automated Safety Evaluations for Generative AI Applications

In this episode, we learn more about new metrics for automated safety evaluations in Azure AI Studio. These evaluations make it easier for organizations to systematically and comprehensively test the quality, safety, and reliability of their LLM application before deploying it in production.

Articles

AI21 Labs’ new AI model can handle more context than most

Increasingly, the AI industry is moving toward generative AI models with longer contexts. But models with large context windows tend to be compute-intensive. Ori Goshen, the CEO of AI startup AI21 Labs, asserts that this doesn’t have to be the case — and his company is releasing a generative model to prove it. Contexts

Incrementally Indexing documents with AzureAI Search Integrated Vectorisation

One can rerun the pipeline which incrementally updates the AI search vector index with new data. There is usually more data that needs to be integrated to a RAG app dynamically e.g. via a blob trigger.

Reinforcement Learning: Introduction and Main Concepts

Apart from policies, in reinforcement learning, it is common to use the notion of value functions which describe how good or bad (in terms of the expected reward) it is for the agent to be in a given state or to take a certain action given the current state. By using feedback from different states and actions, the agent gradually learns the optimal strategy to maximize the total reward over time.

Modeling Extremely Large Images with xT

As computer vision researchers, we believe that every pixel can tell a story. However, there seems to be a writer’s block settling into the field when it comes to dealing with large images. Large images are no longer rare—the cameras we carry in our pockets and those orbiting our planet snap pictures so big and detailed that they stretch our current best models and hardware to their breaking points when handling them. Generally, we face a quadratic increase in memory usage as a function of image size.

Building Local RAG Chatbots Without Coding Using LangFlow and Ollama

A Quick Way to Prototype RAG Applications Based on LangChain

GPU/ML-Enhanced Large Scale Global Routing Contest

Modern VLSI design flows demand scalable global routing techniques applicable across diverse design stages. In response, the ISPD 2024 contest pioneers the first GPU/ML-enhanced global routing competition, selecting advancements in GPU-accelerated computing platforms and machine learning techniques to address scalability challenges. Large-scale benchmarks, containing up to 50 million cells, offer test cases to assess global routers' runtime and memory scalability.

Upcoming events

Microsoft Build | May 21-23, 2024 | Seattle and Online

Learn from in-demand experts, get hands-on with the latest AI innovations, and connect with the developer community.