Focus AI - Edition #101
#Microsoft #OpenAI #Whisper #OpenXLA #Lettria
Focus AI, the last AI news in 3 minutes
Edition #101 - October 28th 2022
Feel free to subscribe or to share it.
This week:
🌬 OpenAI: Introducing Whisper
🛒 Instacart: The Journey to Real-Time Machine Learning
📀 AI Chips Spark International Tension
🚜 Microsoft: Smarts for Farms
💸 New €5M fundraising for the Deeptech startup Lettria
And about code:
[Tutorial 📝] Python 3.11: Cool New Features for You to Try
[Video 📺] MLRun: Speed Demo in 5 Minutes
[Course 🧑🏫] ML Observability Fundamentals
[Repo 🧑💻] OpenXLA, a community-driven and modular open-source compiler
[Lib 💻] OpenVino by Intel
[Live 📺] Beyond Jupyter Notebooks
== News ==
🌬 Introducing Whisper
Open AI has trained and is open-sourcing a neural net called Whisper that approaches human level robustness and accuracy on English speech recognition.
Whisper is an automatic speech recognition (ASR) system trained on 680,000 hours of multilingual and multitask supervised data collected from the web. It shows that the use of such a large and diverse dataset leads to improved robustness to accents, background noise and technical language.
Moreover, it enables transcription in multiple languages, as well as translation from those languages into English. They are open-sourcing models and inference code to serve as a foundation for building useful applications and for further research on robust speech processing.
Check out the paper, model card, and code to learn more details and to try out Whisper.
🛒 Lessons Learned: The Journey to Real-Time Machine Learning at Instacart
As real-time AI gains popularity, it’s valuable to dig into how data science teams are solving the challenges of transitioning to real-time pipelines.
Here’s a very informative post on how Instacart transitioned from batch prediction to online prediction.
📀 AI Chips Spark International Tension
New U.S. restrictions on chip sales aim to hamper China’s AI efforts.
The U.S. government published sweeping limits on sales of processors that involve U.S. designs and technology to Chinese businesses. U.S. officials stated that the restrictions are meant to prevent China from militarizing AI.
The rules block sales of certain processors as well as U.S.-made equipment used to design and manufacture them. This includes high-end graphics processing units (GPUs) and other processors optimized for machine learning.
🚜 Smarts for Farms
The next green revolution may be happening in the server room.
Microsoft open-sourced AI tools designed to help farmers cut costs and improve yields.
FarmVibes-AI includes systems that analyze overhead imagery and sensor data to guide farm operations.
Farmers tend to welcome innovations that improve yields and cut costs. They’re also famously self-sufficient, performing repairs and installing upgrades to their equipment. As self-driving tractors and precision-ag systems take root, they’re great candidates to become early adopters of industry-focused platforms that make it easy for anyone to build useful AI applications.
💸 New €5M fundraising for the Deeptech startup Lettria
Lettria, the platform specializing in automating natural language processing, supported by Sia Partners since its creation, announced a fundraising round of 5 million euros from Galion.exe, Jezby Ventures (Octave Klaba, OVH), and Business Angels.
With this new round of financing, Lettria, who already counts major groups including the likes of La Poste and Leroy Merlin amongst its customers, intends to continue developing its no-code AI platform and international expansion.
== Code & Tools ==
Let’s dive in into more technical news.
[Tutorial 📝] Python 3.11: Cool New Features for You to Try
Python 3.11 was published on October 24, 2022. This latest version of Python is faster and more user-friendly. After seventeen months of development, it’s now ready for prime-time use.
As in every version, Python 3.11 comes with lots of improvements and changes. You can see a list of all of them in the documentation. Here, you’ll explore the coolest and most impactful new features.
[Video 📺] MLRun: Speed Demo in 5 Minutes
Want to orchestrate and accelerate your AI projects with a simple open-source tool?
We can show you how, in under 5 minutes! MLOps Solution Architecht Nicholas Schenone's speed MLRun demo will show you how you can take an integrative approach to manage your ML pipeline, from early development to management at scale, in any production environment.
[Course 🧑🏫] ML Observability Fundamentals
As more machine learning models are deployed into production, it’s imperative to have the right skillset to monitor, troubleshoot, and explain model performance.
This self-paced, ML Observability Fundamentals Course is designed help data scientists and ML practitioners gain confidence taking their models from research to production. Through their pre-recorded instructor videos, checkpoint questions and unit labs, you will gain a hands-on understanding of how to identify where a model is underperforming, troubleshoot model and data issues, and how to proactively mitigate future degradations.
Course Curriculum:
Performance tracing
Drift Detection
Data Quality Management
Feature Importance
Explainability…
[Repo 🧑💻] OpenXLA, a community-driven and modular open-source compiler
The OpenXLA compiler is a community-driven and modular ML compiler. It will enable efficient optimization and deployment of ML models from most major frameworks to any hardware backend notably CPUs, GPUs, and ML ASICs.
[Lib 💻] OpenVino by Intel
OpenVINO™ is an open-source toolkit for optimizing and deploying AI inference.
Boost deep learning performance in computer vision, automatic speech recognition, natural language processing and other common tasks
Use models trained with popular frameworks like TensorFlow, PyTorch and more
Reduce resource demands and efficiently deploy on a range of Intel® platforms from edge to cloud
[Live 📺] Beyond Jupyter Notebooks
This live is an interactive workshop on how to set up your new computer to build and deploy industry-standard machine learning applications so that you’re ready to maximize your impact during your first week on the job.
Greg from FourthBrain will guide you through version control, conda, jupyter, docker in less than an hour.
#Microsoft #OpenAI #Whisper #OpenXLA #Lettria
A big thanks to our sources: https://jack-clark.net/, https://www.actuia.com/, https://thevariable.com/news/, https://techcrunch.com/, https://read.deeplearning.ai/the-batch/
What about you? Have you noticed something else?
Don’t miss the next news!
Have a good week-end,
Maxime 🙃 from Toulouse, France with 🌺