AI with MATLAB: 3 things that happened in 2024
Today is the last day of 2024. Most of you are busy with non-AI, non-programming, non-thinking-about-work preparations for the new year. Nevertheless, I wanted to wrap up this year by looking back at exciting things that happened for AI with MATLAB. I set myself the goal to pick only three highlights to write about, and this was the hardest part of writing this blog post.
Without further ado, here are my three picks for AI with MATLAB in 2024.
File Exchange Repository: Large Language Models (LLMs) with MATLAB
The Large Language (LLMs) with MATLAB repository was created to provide you with the code you need to access and interact with LLMs using MATLAB. You can connect MATLAB to the OpenAI® Chat Completions API (which powers ChatGPT™), Ollama™ (for local LLMs), and Azure® OpenAI services.
By interfacing the OpenAI API from your MATLAB environment, you can use models (such as GPT-4 and GPT-4 Turbo) for a variety of NLP tasks, including building your own chatbot and sentiment analysis. To learn more, read this blog post.
By connecting MATLAB to a local Ollama server, you have access to popular local LLMs, such as llama3, mistral, and gemma. Local LLMs are great for NLP tasks, such as retrieval-augmented generation (RAG) that can enhance the LLM accuracy by using your own data. To learn more, read this blog post.
Conversion of deep learning models between MATLAB, PyTorch, TensorFlow, and ONNX
To facilitate cross-platform and cross-team collaboration when designing AI-enabled systems, Deep Learning Toolbox integrates with PyTorch®, TensorFlow™, and other Python®-based frameworks. With every release this integration is enhanced and improved.
With MATLAB R2024b, you can now import deep learning models from PyTorch 2.0 by using the importNetworkFromPyTorch function. To learn more about interoperation, see Convert Deep Learning Models between PyTorch, TensorFlow, and MATLAB.
Another exciting enhancement is that you can now simulate and test PyTorch, TensorFlow, ONNX, and custom Python models within systems by using the Simulink co-execution blocks. This allows you to quickly iterate on your design, assess model behavior, and test system performance.
W-shaped development process. Credit: EASA, Daedalean
AI-enabled engineered systems are being increasingly adopted in safety-critical industries like aerospace, automotive, and manufacturing, where ensuring reliability and safety is vital. In 2024, Lucas García completed a blog post series on “Verification and Validation for AI”. This series delves into the meticulous process of ensuring AI models are both reliable and robust, especially in the context of safety-critical applications.
In 2024, important tools were introduced in MATLAB to help you build confidence in AI. Two of these are a function for explainability for object detection and a repository for performing constrained deep learning.
Large Language Models with MATLAB
File Exchange Repository: Large Language Models (LLMs) with MATLAB
The Large Language (LLMs) with MATLAB repository was created to provide you with the code you need to access and interact with LLMs using MATLAB. You can connect MATLAB to the OpenAI® Chat Completions API (which powers ChatGPT™), Ollama™ (for local LLMs), and Azure® OpenAI services.
By interfacing the OpenAI API from your MATLAB environment, you can use models (such as GPT-4 and GPT-4 Turbo) for a variety of NLP tasks, including building your own chatbot and sentiment analysis. To learn more, read this blog post.
By connecting MATLAB to a local Ollama server, you have access to popular local LLMs, such as llama3, mistral, and gemma. Local LLMs are great for NLP tasks, such as retrieval-augmented generation (RAG) that can enhance the LLM accuracy by using your own data. To learn more, read this blog post.
Enhanced Integration with Python
Conversion of deep learning models between MATLAB, PyTorch, TensorFlow, and ONNX
To facilitate cross-platform and cross-team collaboration when designing AI-enabled systems, Deep Learning Toolbox integrates with PyTorch®, TensorFlow™, and other Python®-based frameworks. With every release this integration is enhanced and improved.
With MATLAB R2024b, you can now import deep learning models from PyTorch 2.0 by using the importNetworkFromPyTorch function. To learn more about interoperation, see Convert Deep Learning Models between PyTorch, TensorFlow, and MATLAB.
Another exciting enhancement is that you can now simulate and test PyTorch, TensorFlow, ONNX, and custom Python models within systems by using the Simulink co-execution blocks. This allows you to quickly iterate on your design, assess model behavior, and test system performance.
Verification and Validation for AI
W-shaped development process. Credit: EASA, Daedalean
AI-enabled engineered systems are being increasingly adopted in safety-critical industries like aerospace, automotive, and manufacturing, where ensuring reliability and safety is vital. In 2024, Lucas García completed a blog post series on “Verification and Validation for AI”. This series delves into the meticulous process of ensuring AI models are both reliable and robust, especially in the context of safety-critical applications.
In 2024, important tools were introduced in MATLAB to help you build confidence in AI. Two of these are a function for explainability for object detection and a repository for performing constrained deep learning.
The Year Ahead
Happy New Year! Let’s look forward to a year with more exciting AI tools and AI system integration, bridging the gap between innovation and real-world applications.
댓글
댓글을 남기려면 링크 를 클릭하여 MathWorks 계정에 로그인하거나 계정을 새로 만드십시오.