October Newsletter

Hi everyone-

Well, September certainly seemed to disappear pretty rapidly (along with the sunshine sadly). And dramatic events keep accumulating, from the sad death of the Queen, together with epic coverage of ‘the queue‘, to dramatic counter offensives in the Ukraine, to unprecedented IMF criticism of the UK government’s tax-cutting plans. Perhaps time for a breather, with a wrap up data science developments in the last month.

Following is the October edition of our Royal Statistical Society Data Science and AI Section newsletter. Hopefully some interesting topics and titbits to feed your data science curiosity. (If you are reading this on email and it is not formatting well, try viewing online at http://datasciencesection.org/)

As always- any and all feedback most welcome! If you like these, do please send on to your friends- we are looking to build a strong community of data science practitioners. And if you are not signed up to receive these automatically you can do so here.

Industrial Strength Data Science October 2022 Newsletter

RSS Data Science Section

Committee Activities

The RSS 2022 Conference, held on 12-15 September in Aberdeen was a great success. The Data Science and AI Section’s session ‘The secret sauce of open source’ was undoubtedly a highlight (we are clearly biased!) but all in all lots of relevant, enlightening and entertaining talks for a practicing data scientist. See David Hoyle’s commentary here (also highlighted in the Members section below).

Following hot on the heels of our July meetup, ‘From Paper to Pitch‘ we were very pleased with our latest event, “IP Freely, making algorithms pay – Intellectual property in Data Science and AI” which was held on Wednesday 21 September 2022. A lively and engaging discussion was held including leading figures such as Dr David Barber (Director of the UCL Centre for Artificial Intelligence ) and Professor Noam Shemtov (Intellectual Property and Technology Law at Queen Mary’s University London).

The AI Standards Hub, an initiative that we reported on earlier this year, led by committee member Florian Ostmann, will see its official launch on 12 October. Part of the National AI Strategy, the Hub’s new online platform and activities will be dedicated to knowledge sharing, community building, strategic research, and international engagement around standardisation for AI technologies. The launch event will be livestreamed online and feature presentations and interactive discussions with senior government representatives, the Hub’s partner organisations, and key stakeholders. To join the livestream, please register before 10 October using this link (https://tinyurl.com/AIStandardsHub). 

Martin Goodson (CEO and Chief Scientist at Evolution AI) continues to run the excellent London Machine Learning meetup and is very active with events. The next event is on October 12th when Aditya Ramesh, Researcher at OpenAI, will discuss (the very topical) “Manipulating Images with DALL-E 2“. Videos are posted on the meetup youtube channel – and future events will be posted here.

This Month in Data Science

Lots of exciting data science going on, as always!

Ethics and more ethics…

Bias, ethics and diversity continue to be hot topics in data science…

"Interactive deepfakes have the capability to impersonate people with realistic interactive behaviors, taking advantage of advances in multimodal interaction. Compositional deepfakes leverage synthetic content in larger disinformation plans that integrate sets of deepfakes over time with observed, expected, and engineered world events to create persuasive synthetic histories"
"We argue that the upcoming regulation might be particularly important in offering the first and most influential operationalisation of what it means to develop and deploy trustworthy or human-centred AI. If the EU regime is likely to see significant diffusion, ensuring it is well-designed becomes a matter of global importance.."
"Most of the problems you will face are, in fact, engineering problems. Even with all the resources of a great machine learning expert, most of the gains come from great features, not great machine learning algorithms. So, the basic approach is:
1. make sure your pipeline is solid end to end
2. start with a reasonable objective
3. add common­sense features in a simple way
4. make sure that your pipeline stays solid.
This approach will make lots of money and/or make lots of people happy for a long period of time. Diverge from this approach only when there are no more simple tricks to get you any farther. Adding complexity slows future releases."
  • Finally, we can also try and build ‘fairness’ into the underling algorithms, and machine learning approaches. For instance, this looks to be an excellent idea – FairGBM
"FairGBM is an easy-to-use and lightweight fairness-aware ML algorithm with state-of-the-art performance on tabular datasets.

FairGBM builds upon the popular LightGBM algorithm and adds customizable constraints for group-wise fairness (e.g., equal opportunity, predictive equality) and other global goals (e.g., specific Recall or FPR prediction targets)."

Developments in Data Science Research…

As always, lots of new developments on the research front and plenty of arXiv papers to read…

"Even the largest neural networks make errors, and once-correct predictions can become invalid as the world changes. Model editors make local updates to the behavior of base (pre-trained) models to inject updated knowledge or correct undesirable behaviors"
"We consider an ability to be emergent if it is not present in smaller models but is present in larger models. Thus, emergent abilities cannot be predicted simply by extrapolating the performance of smaller models. The existence of such emergence implies that additional scaling could further expand the range of capabilities of language models"
  • DeepMind have released Menagerie – “a collection of high-quality models for the MuJoCo physics engine”: looks very useful for anyone working with physics simulators
  • Finally, another great stride for the open source community this time from LAION – a large scale open source version of CLIP (a key component of image generation models that computes representations of images and texts to measure similarity)
We replicated the results from openai CLIP in models of different sizes, then trained bigger models. The full evaluation suite on 39 datasets (vtab+) are available in this results notebook and show consistent improvements over all datasets.

Stable-Dal-Gen oh my…

Lots of discussion about the new breed of text-to-image models (type in a text prompt/description and an -often amazing- image is generated) with three main models available right now: DALLE2 from OpenAI, Imagen from Google and the open source Stable-Diffusion from stability.ai.

"minGPT tries to be small, clean, interpretable and educational, as most of the currently available GPT model implementations can a bit sprawling. GPT is not a complicated model and this implementation is appropriately about 300 lines of code (see mingpt/model.py). All that's going on is that a sequence of indices feeds into a Transformer, and a probability distribution over the next index in the sequence comes out. The majority of the complexity is just being clever with batching (both across examples and over sequence length) for efficiency."

Real world applications of Data Science

Lots of practical examples making a difference in the real world this month!

"By using our latest AI model, Multitask Unified Model (MUM), our systems can now understand the notion of consensus, which is when multiple high-quality sources on the web all agree on the same fact. Our systems can check snippet callouts (the word or words called out above the featured snippet in a larger font) against other high-quality sources on the web, to see if there’s a general consensus for that callout, even if sources use different words or concepts to describe the same thing. We've found that this consensus-based technique has meaningfully improved the quality and helpfulness of featured snippet callouts."
“One of the motivations of this work was our desire to study systems that learn models of datasets that is represented in a way that humans can understand. Instead of learning weights, can the model learn expressions or rules? And we wanted to see if we could build this system so it would learn on a whole battery of interrelated datasets, to make the system learn a little bit about how to better model each one"

How does that work?

Tutorials and deep dives on different approaches and techniques

"Deep learning is sometimes referred to as “representation learning” because its strength is the ability to learn the feature extraction pipeline. Most tabular datasets already represent (typically manually) extracted features, so there shouldn’t be a significant advantage using deep learning on these."

Practical tips

How to drive analytics and ML into production

Bigger picture ideas

Longer thought provoking reads – lean back and pour a drink! …

“We’re not trying to re-create the brain,” said David Ha, a computer scientist at Google Brain who also works on transformer models. “But can we create a mechanism that can do what the brain does?”
"A common finding is that with the right representation, the problem becomes much easier. However, how to train the neural network to learn useful representations is still poorly understood. Here, causality can help. In causal representation learning, the problem of representation learning is framed as finding the causal variables, as well as the causal relations between them.."
"As we’ve seen, the nature of algorithms requires new types of tradeoff, both at the micro-decision level, and also at the algorithm level. A critical role for leaders is to navigate these tradeoffs, both when the algorithm is designed, but also on an ongoing basis. Improving algorithms is increasingly a matter of changing rules or parameters in software, more like tuning the knobs on a graphic equalizer than rearchitecting a physical plant or deploying a new IT system"
"Lucas concludes his essay by stating that the characteristic attribute of human minds is the ability to step outside the system. Minds, he argues, are not constrained to operate within a single formal system, but rather they can switch between systems, reason about a system, reason about the fact that they reason about a system, etc. Machines, on the other hand, are constrained to operate within a single formal system that they could not escape. Thus, he argues, it is this ability that makes human minds inherently different from machines."

Fun Practical Projects and Learning Opportunities

A few fun practical projects and topics to keep you occupied/distracted:

Covid Corner

Apparently Covid is over – certainly there are very limited restrictions in the UK now

Updates from Members and Contributors

  • David Hoyle has published an excellent review of the recent RSS conference, highlighting the increasing relevance to practicing Data Scientists- well worth a read
  • The ONS are keen to highlight the last of this year’s ONS – UNECE Machine Learning Groups Coffee and Coding session on 2 November 2022 at 1400 – 1530 (CEST) / 0900 – 1030 (EST) when Tabitha Williams and Brittny Vongdara from Statistics Canada will provide an interactive lesson on using GitHub, and an introduction to Git. For more information and to register, please visit the Eventbrite page (Coffee and Coding Session 2 November). Any questions, get in touch at ML2022@ons.gov.uk


The Job market is a bit quiet over the summer- let us know if you have any openings you’d like to advertise

  • EvolutionAI, are looking to hire someone for applied deep learning research. Must like a challenge. Any background but needs to know how to do research properly. Remote. Apply here

Again, hope you found this useful. Please do send on to your friends- we are looking to build a strong community of data science practitioners- and sign up for future updates here.

– Piers

The views expressed are our own and do not necessarily represent those of the RSS

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: