 | Top 3 Challenges in Prod ML The Top 3 Challenges in Production Machine Learning include 1) ML Monitoring, 2) Data, and 3) Showcasing business value: The challenges highlighted in production machine learning seem to resonate quite a lot with what we see in practice; the top 3 challenges in are: 1) Monitoring - Establishing standardised and robust monitoring for ML systems; 2) Data - Access to relevant data for training (which is aligned data in production inference), and; 3) Impact - Showcasing business impact and business value on usecases. Further challenges highlighted include: 4) Inconsistency of training and experimentation environments; 5) Building production-grade ML pipelines; 6) Gaps in tooling and support for model productionisation, and 7) Governance and Domain Risks. These are really important insights from 2024 Survey on The State of Production ML - if you have a chance we would be grateful if you could spend a few minutes on the survey, as you'll contribute valuable information about the machine learning tools and platforms you use in your production ML development. Your input will help create a comprehensive overview of common practices, tooling preferences, and challenges faced when deploying models to production, ultimately benefiting the entire ML community 🚀 |
|
|
|---|
|
Google DeepMind Weather Forecast Google DeepMind has released an exciting foundation model for weather forecasting - this is a hybrid approach that brings together traditional physics-simulations and deep learning to surpass the current best models for weather and climate forecasting: The Google team provides a great deep dive into how they implemented their new NeuralGCM model using JAX to overcome the limitations of pure physics or machine learning approaches. This approach is showcased as achieving higher accuracy than state-of-the-art models whilst being faster, more cost-effective, producing more detailed forecasts, and doing everything with a simplified codebase (vs 1m+ line FORTRAN codebase). This is actually published together with the full code available, as well as the pre-trained models, as well as a fully-fledged benchmarking suite - this is certainly an exciting time for this space. |
|
|
|---|
|
PlanetScale Embeddings SQL Vector embeddings have been exploding in popularity, and similarly we have seen a similar growth on vector databases - this week we see an interesting new challenger with PlanetScale introducing native support for AI embedding vectors in its MySQL-compatible database; this is quite exciting as it comes as an integrated index that supports all relational database features, eliminating the need for a separate vector database. The vector search functionality is built upon Microsoft Research's SPANN and SPFresh algorithms, which in themselves deserve their own deep dive, as they have enabled integration into MySQL's default storage engine (InnoDB); this means that inserts, updates, and deletes of vector data are immediately reflected in the vector index as part of SQL transactions, ensuring ACID compliance. This is quite an interesting advancement, as we are able to see the foundational DB field and emerging technologies converge to ensure standardisation and best practice to simplify architecture and operations while leveraging familiar SQL features. |
|
|---|
|
Google Shopping 10m Embedding Vector Embedding Platform Marco has released a Google Shopping Dataset with 10m products for benchmarking of multi-modal ranking evaluation, together with a new approach to improve multi-modal retrieval and ranking. This is quite an interesting deep dive into their "Generalized Contrastive Learning" framework for optimization of retrieval and ranking by encoding multiple data types (like text and images) and incorporating fine-grained relevance directly into embeddings. From the research they share, it seems this framework extends CLIP-style models to improve multi-modal retrieval and ranking whilst addressing limitations of existing approaches through unified representations of documents composed of multiple fields, enhancing intra-modal understanding, and optimizing embeddings for efficient storage in vector databases. It is great to see initiatives that bring together novel approaches together with benchmarking frameworks that encourage further breakthroughs. |
|
|
|---|
|
META Llama LLM Optimisation Efficiency in LLMs is becoming growingly important; Meta has now released quantized versions of its Llama 3.2 1B and 3B models optimized for mobile devices which they suggest up to 4× speedup + 56% reduction in model size + 41% less memory usage compared to the original models: This is quite an interesting approach using QLoRA + post-training quantization method which doesn't require the original training data. It is quite insightful to see that the race to larger accurate models is almost marching the race to smaller accurate models! |
|
|
|---|
|
Upcoming MLOps Events The MLOps ecosystem continues to grow at break-neck speeds, making it ever harder for us as practitioners to stay up to date with relevant developments. A fantsatic way to keep on-top of relevant resources is through the great community and events that the MLOps and Production ML ecosystem offers. This is the reason why we have started curating a list of upcoming events in the space, which are outlined below. Upcoming conferences where we're speaking:
Other upcoming MLOps conferences in 2024:
In case you missed our talks:
|
|
|---|
| | |
Check out the fast-growing ecosystem of production ML tools & frameworks at the github repository which has reached over 10,000 ⭐ github stars. We are currently looking for more libraries to add - if you know of any that are not listed, please let us know or feel free to add a PR. Four featured libraries in the GPU acceleration space are outlined below. - Kompute - Blazing fast, lightweight and mobile phone-enabled GPU compute framework optimized for advanced data processing usecases.
- CuPy - An implementation of NumPy-compatible multi-dimensional array on CUDA. CuPy consists of the core multi-dimensional array class, cupy.ndarray, and many functions on it.
- Jax - Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more
- CuDF - Built based on the Apache Arrow columnar memory format, cuDF is a GPU DataFrame library for loading, joining, aggregating, filtering, and otherwise manipulating data.
If you know of any open source and open community events that are not listed do give us a heads up so we can add them! |
|
|---|
| | |
As AI systems become more prevalent in society, we face bigger and tougher societal challenges. We have seen a large number of resources that aim to takle these challenges in the form of AI Guidelines, Principles, Ethics Frameworks, etc, however there are so many resources it is hard to navigate. Because of this we started an Open Source initiative that aims to map the ecosystem to make it simpler to navigate. You can find multiple principles in the repo - some examples include the following: - MLSecOps Top 10 Vulnerabilities - This is an initiative that aims to further the field of machine learning security by identifying the top 10 most common vulnerabiliites in the machine learning lifecycle as well as best practices.
- AI & Machine Learning 8 principles for Responsible ML - The Institute for Ethical AI & Machine Learning has put together 8 principles for responsible machine learning that are to be adopted by individuals and delivery teams designing, building and operating machine learning systems.
- An Evaluation of Guidelines - The Ethics of Ethics; A research paper that analyses multiple Ethics principles.
- ACM's Code of Ethics and Professional Conduct - This is the code of ethics that has been put together in 1992 by the Association for Computer Machinery and updated in 2018.
If you know of any guidelines that are not in the "Awesome AI Guidelines" list, please do give us a heads up or feel free to add a pull request!
|
|
|---|
| | |
| | | | The Institute for Ethical AI & Machine Learning is a European research centre that carries out world-class research into responsible machine learning. | | | | |
|
|
|---|
|
|
This email was sent to You received this email because you are registered with The Institute for Ethical AI & Machine Learning's newsletter "The Machine Learning Engineer"
|
| | | | |
|
|
|---|
|
© 2023 The Institute for Ethical AI & Machine Learning |
|
|---|
|
|
|
The Institute for Ethical AI & Machine Learning
