Stanford on Machine Unlearning Stanford AI Lab researcher dives into Machine Unlearning in 2024, or the art of removing the influences of training data from a trained model: A really insightful view into the growing emerging need for "machine unlearning" in machine learning systems, particularly as models and data sets grow. This ML unlearning is defined as the removal of specific data influences from a trained model to address issues such as privacy, outdated information, and unsafe content. There are a few interesting unlearning techniques (or perhaps more as "categories") ranging across 1) exact, 2) differential privacy, 3) empirical specific, 4) empirical underspecific, and 5) few-shot prompting. ML unlearning is clearly a topic that will continue growing for the immediate term, supported by initiatives such as the recent NeurIPS 2023 "ML unlearning challenge". |
|
|---|
|
StackOverflow and OpenAI An interesting development in the AI race with Stack Overflow and OpenAI announcing a partnership with OverflowAPI: This is an interesting collaboration, it's outlined as StackOverflow making their data available for OpenAI's models and services, most likely seeing an appearance in OpenAI models. However there has already been quite a mixed response online across various forums (hackernews, twitter, even linkedin) where active users are challenging this decision and even deactivating their accounts due to differing perspectives. It will certainly be interesting to see how this develops, however one thing that will be certain is that we will see interesting innovations in the space of developer productivity. |
|
|---|
|
Open Time Series Large Models Recently the release of time-series foundation models across Amazon, Google and beyond have highlighted challenges such as lack of benchmarks, datasets and even models; the MOMENT initiative aims to tackle this: A great initiative aiming to provide a compiled dataset for time-series foundation models, standardised benchmark to evaluate foundation models, and a suite of time-series foundation models that excel in multiple time-series contexts. |
|
|
|---|
|
Scraping with AI (Prompts) It was only about time until AI came into web scraping with ScrapeGraphAI: An interesting new initiative bringing specialised LLM agentic design into web scraping through graph-based logic. Quite interesting to see interfaces in SDKs that require prompting to specify the tasks, which open questions of reproducibility and debugging that will be interesting to see as production adoption of these tools increasese. Furthermore support for "multiple LLM backends" also opens further considerations that although improve its ease, may have further nuances on how prompt engineering would be optimised as the libraries are expanded through maintenance. |
|
|
|---|
|
The Illustrated Word2Vec The Illustrated Word2vec is a classic in ML for the key foundation concept of embeddings - this article provides a fantastic visual and intuitive overview: Word2vec has been quite influential upon its release, so it's a great opportunity for ML practitioners to brush up on how language embeddings underpin various applications, from language tasks to recommendation systems used by major companies. This is also a great overview of the technical details, including the mechanics and training processes of Word2vec, Continuous Bag of Words and Skipgram architectures, as well as advanced techniques like negative sampling. |
|
|---|
|
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
