Towards Data Science

Progress in AI has been accelerating dramatically in recent years, and even months. It seems like every other day, there’s a new, previously-believed-to-be-impossible feat of AI that’s achieved by a world-leading lab. And increasingly, these breakthroughs have been driven by the same, simple idea: AI scaling.

For those who haven’t been following the AI scaling sage, scaling means training AI systems with larger models, using increasingly absurd quantities of data and processing power. So far, empirical studies by the world’s top AI labs seem to suggest that scaling is an open-ended process that can lead to more and more capable and intelligent systems, with no clear limit.

And that’s led many people to speculate that scaling might usher in a new era of broadly human-level or even superhuman AI — the holy grail AI researchers have been after for decades.

And while that might sound cool, an AI that can solve general reasoning problems as well as or better than a human might actually be an intrinsically dangerous thing to build.

At least, that’s the conclusion that many AI safety researchers have come to following the publication of a new line of research that explores how modern AI systems tend to solve problems, and whether we should expect more advanced versions of them to perform dangerous behaviours like seeking power.

This line of research in AI safety is called “power-seeking”, and although it’s currently not well understood outside the frontier of AI safety and AI alignment research, it’s starting to draw a lot of attention. The first major theoretical study of power seeking was led by Alex Turner, who’s appeared on the podcast before, and was published in NeurIPS (the world’s top AI conference), for example.

And today, we’ll be hearing from Edouard Harris, an AI alignment researcher and one of my co-founders in the AI safety company (Gladstone AI). Ed’s just completed a significant piece of AI safety research that extends Alex Turner’s original power-seeking work, and that shows what seems to be the first experimental evidence suggesting that we should expect highly advanced AI systems to seek power by default.

What does power seeking really mean though? And does all this imply for the safety of future, general-purpose reasoning systems? That’s what this episode will be all about.

***

Intro music:

- Artist: Ron Gelinas

- Track Title: Daybreak Chill Blend (original mix)

It’s no secret that a new generation of powerful and highly scaled language models is taking the world by storm. Companies like OpenAI, AI21Labs, and Cohere have built models so versatile that they’re powering hundreds of new applications, and unlocking entire new markets for AI-generated text.

In light of that, I thought it would be worth exploring the applied side of language modelling — to dive deep into one specific language model-powered tool, to understand what it means to build apps on top of scaled AI systems. How easily can these models be used in the wild? What bottlenecks and challenges do people run into when they try to build apps powered by large language models? That’s what I wanted to find out.

My guest today is Amber Teng, and she’s a data scientist who recently published a blog that got quite a bit of attention, about a resume cover letter generator that she created using GPT-3, OpenAI’s powerful and now-famous language model. I thought her project would be make for a great episode, because it exposes so many of the challenges and opportunities that come with the new era of powerful language models that we’ve just entered.

So today we’ll be exploring exactly that: looking at the applied side of language modelling and prompt engineering, understanding how large language models have made new apps not only possible but also much easier to build, and the likely future of AI-powered products.

***

Intro music:

- Artist: Ron Gelinas

- Track Title: Daybreak Chill Blend (original mix)

Imagine you’re a big hedge fund, and you want to go out and buy yourself some data. Data is really valuable for you — it’s literally going to shape your investment decisions and determine your outcomes.

But the moment you receive your data, a cold chill runs down your spine: how do you know your data supplier gave you the data they said they would? From your perspective, you’re staring down 100,000 rows in a spreadsheet, with no way to tell if half of them were made up — or maybe more for that matter.

This might seem like an obvious problem in hindsight, but it’s one most of us haven’t even thought of. We tend to assume that data is data, and that 100,000 rows in a spreadsheet is 100,000 legitimate samples.

Today, we live in the era of AI scaling. It seems like everywhere you look people are pushing to make large language models larger, or more multi-modal and leveraging ungodly amounts of processing power to do it.

But although that’s one of the defining trends of the modern AI era, it’s not the only one. At the far opposite extreme from the world of hyperscale transformers and giant dense nets is the fast-evolving world of TinyML, where the goal is to pack AI systems onto small edge devices.

Deep learning models — transformers in particular — are defining the cutting edge of AI today. They’re based on an architecture called an artificial neural network, as you probably already know if you’re a regular Towards Data Science reader. And if you are, then you might also already know that as their name suggests, artificial neural networks were inspired by the structure and function of biological neural networks, like those that handle information processing in our brains.

So it’s a natural question to ask: how far does that analogy go? Today, deep neural networks can master an increasingly wide range of skills that were historically unique to humans — skills like creating images, or using language, planning, playing video games, and so on. Could that mean that these systems are processing information like the human brain, too?

To explore that question, we’ll be talking to JR King, a CNRS researcher at the Ecole Normale Supérieure, affiliated with Meta AI, where he leads the Brain & AI group. There, he works on identifying the computational basis of human intelligence, with a focus on language. JR is a remarkably insightful thinker, who’s spent a lot of time studying biological intelligence, where it comes from, and how it maps onto artificial intelligence. And he joined me to explore the fascinating intersection of biological and artificial information processing on this episode of the TDS podcast.

***

Intro music:

- Artist: Ron Gelinas

- Track Title: Daybreak Chill Blend (original mix)

It’s no secret that the US and China are geopolitical rivals. And it’s also no secret that that rivalry extends into AI — an area both countries consider to be strategically critical.

Two ML researchers with world-class pedigrees who decided to build a company that puts AI on the blockchain. Now to most people — myself included — “AI on the blockchain” sounds like a winning entry in some kind of startup buzzword bingo. But what I discovered talking to Jacob and Ala was that they actually have good reasons to combine those two ingredients together.

At a high level, doing AI on a blockchain allows you to decentralize AI research and reward labs for building better models, and not for publishing papers in flashy journals with often biased reviewers.

And that’s not all — as we’ll see, Ala and Jacob are taking on some of the thorniest current problems in AI with their decentralized approach to machine learning. Everything from the problem of designing robust benchmarks to rewarding good AI research and even the centralization of power in the hands of a few large companies building powerful AI systems — these problems are all in their sights as they build out Bittensor, their AI-on-the-blockchain-startup.

Ala and Jacob joined me to talk about all those things and more on this episode of the TDS podcast.

---

Intro music:

- Artist: Ron Gelinas

- Track Title: Daybreak Chill Blend (original mix)

As you might know if you follow the podcast, we usually talk about the world of cutting-edge AI capabilities, and some of the emerging safety risks and other challenges that the future of AI might bring. But I thought that for today’s episode, it would be fun to change things up a bit and talk about the applied side of data science, and how the field has evolved over the last year or two.

AI scaling has really taken off. Ever since GPT-3 came out, it’s become clear that one of the things we’ll need to do to move beyond narrow AI and towards more generally intelligent systems is going to be to massively scale up the size of our models, the amount of processing power they consume and the amount of data they’re trained on, all at the same time.

That’s led to a huge wave of highly scaled models that are incredibly expensive to train, largely because of their enormous compute budgets. But what if there was a more flexible way to scale AI — one that allowed us to decouple model size from compute budgets, so that we can track a more compute-efficient course to scale?

That’s the promise of so-called mixture of experts models, or MoEs. Unlike more traditional transformers, MoEs don’t update all of their parameters on every training pass. Instead, they route inputs intelligently to sub-models called experts, which can each specialize in different tasks. On a given training pass, only those experts have their parameters updated. The result is a sparse model, a more compute-efficient training process, and a new potential path to scale.

Google has been pushing the frontier of research on MoEs, and my two guests today in particular have been involved in pioneering work on that strategy (among many others!). Liam Fedus and Barrett Zoph are research scientists at Google Brain, and they joined me to talk about AI scaling, sparsity and the present and future of MoE models on this episode of the TDS podcast.

***

Intro music:

- Artist: Ron Gelinas

- Track Title: Daybreak Chill Blend (original mix)

There’s an idea in machine learning that most of the progress we see in AI doesn’t come from new algorithms of model architectures. instead, some argue, progress almost entirely comes from scaling up compute power, datasets and model sizes — and besides those three ingredients, nothing else really matters.

Through that lens the history of AI becomes the history f processing power and compute budgets. And if that turns out to be true, then we might be able to do a decent job of predicting AI progress by studying trends in compute power and their impact on AI development.

Generating well-referenced and accurate Wikipedia articles has always been an important problem: Wikipedia has essentially become the Internet's encyclopedia of record, and hundreds of millions of people use it do understand the world.

But over the last decade Wikipedia has also become a critical source of training data for data-hungry text generation models. As a result, any shortcomings in Wikipedia’s content are at risk of being amplified by the text generation tools of the future. If one type of topic or person is chronically under-represented in Wikipedia’s corpus, we can expect generative text models to mirror — or even amplify — that under-representation in their outputs.

Through that lens, the project of Wikipedia article generation is about much more than it seems — it’s quite literally about setting the scene for the language generation systems of the future, and empowering humans to guide those systems in more robust ways.

That’s why I wanted to talk to Meta AI researcher Angela Fan, whose latest project is focused on generating reliable, accurate, and structured Wikipedia articles. She joined me to talk about her work, the implications of high-quality long-form text generation, and the future of human/AI collaboration on this episode of the TDS podcast.

--- 

Intro music:

- Artist: Ron Gelinas

- Track Title: Daybreak Chill Blend (original mix)

Trustworthy AI is one of today’s most popular buzzwords. But although everyone seems to agree that we want AI to be trustworthy, definitions of trustworthiness are often fuzzy or inadequate. Maybe that shouldn’t be surprising: it’s hard to come up with a single set of standards that add up to “trustworthiness”, and that apply just as well to a Netflix movie recommendation as a self-driving car.

So maybe trustworthy AI needs to be thought of in a more nuanced way — one that reflects the intricacies of individual AI use cases. If that’s true, then new questions come up: who gets to define trustworthiness, and who bears responsibility when a lack of trustworthiness leads to harms like AI accidents, or undesired biases?

Through that lens, trustworthiness becomes a problem not just for algorithms, but for organizations. And that’s exactly the case that Beena Ammanath makes in her upcoming book, Trustworthy AI, which explores AI trustworthiness from a practical perspective, looking at what concrete steps companies can take to make their in-house AI work safer, better and more reliable. Beena joined me to talk about defining trustworthiness, explainability and robustness in AI, as well as the future of AI regulation and self-regulation on this episode of the TDS podcast.

Intro music:

- Artist: Ron Gelinas

- Track Title: Daybreak Chill Blend (original mix)

Until recently, very few people were paying attention to the potential malicious applications of AI. And that made some sense: in an era where AIs were narrow and had to be purpose-built for every application, you’d need an entire research team to develop AI tools for malicious applications. Since it’s more profitable (and safer) for that kind of talent to work in the legal economy, AI didn’t offer much low-hanging fruit for malicious actors.

But today, that’s all changing. As AI becomes more flexible and general, the link between the purpose for which an AI was built and its potential downstream applications has all but disappeared. Large language models can be trained to perform valuable tasks, like supporting writers, translating between languages, or write better code. But a system that can write an essay can also write a fake news article, or power an army of humanlike text-generating bots.

More than any other moment in the history of AI, the move to scaled, general-purpose foundation models has shown how AI can be a double-edged sword. And now that these models exist, we have to come to terms with them, and figure out how to build societies that remain stable in the face of compelling AI-generated content, and increasingly accessible AI-powered tools with malicious use potential.

Imagine, for example, an AI that’s trained to identify cows in images. Ideally, we’d want it to learn to detect cows based on their shape and colour. But what if the cow pictures we put in the training dataset always show cows standing on grass?

In that case, we have a spurious correlation between grass and cows, and if we’re not careful, our AI might learn to become a grass detector rather than a cow detector. Even worse, we could only realize that’s happened once we’ve deployed it in the real world and it runs into a cow that isn’t standing on grass for the first time.

So how do you build AI systems that can learn robust, general concepts that remain valid outside the context of their training data?

That’s the problem of out-of-distribution generalization, and it’s a central part of the research agenda of Irina Rish, a core member of the Mila— Quebec AI Research institute, and the Canadian Excellence Research Chair in Autonomous AI. Irina’s research explores many different strategies that aim to overcome the out-of-distribution problem, from empirical AI scaling efforts to more theoretical work, and she joined me to talk about just that on this episode of the podcast.

***

Intro music:

- Artist: Ron Gelinas

- Track Title: Daybreak Chill Blend (original mix)

Google the phrase “AI over-hyped”, and you’ll find literally dozens of articles from the likes of Forbes, Wired, and Scientific American, all arguing that “AI isn’t really as impressive at it seems from the outside,” and “we still have a long way to go before we come up with *true* AI, don’t you know.”

Amusingly, despite the universality of the “AI is over-hyped” narrative, the statement that “We haven’t made as much progress in AI as you might think™️” is often framed as somehow being an edgy, contrarian thing to believe.

All that pressure not to over-hype AI research really gets to people — researchers included. And they adjust their behaviour accordingly: they over-hedge their claims, cite outdated and since-resolved failure modes of AI systems, and generally avoid drawing straight lines between points that clearly show AI progress exploding across the board. All, presumably, to avoid being perceived as AI over-hypers.

Why does this matter? Well for one, under-hyping AI allows us to stay asleep — to delay answering many of the fundamental societal questions that come up when widespread automation of labour is on the table. But perhaps more importantly, it reduces the perceived urgency of addressing critical problems in AI safety and AI alignment.

Yes, we need to be careful that we’re not over-hyping AI. “AI startups” that don’t use AI are a problem. Predictions that artificial general intelligence is almost certainly a year away are a problem. Confidently prophesying major breakthroughs over short timescales absolutely does harm the credibility of the field.

But at the same time, we can’t let ourselves be so cautious that we’re not accurately communicating the true extent of AI’s progress and potential. So what’s the right balance?

That’s where Sam Bowman comes in. Sam is a professor at NYU, where he does research on AI and language modeling. But most important for today’s purposes, he’s the author of a paper titled, “When combating AI hype, proceed with caution,” in which he explores a trend he calls under-claiming — a common practice among researchers that consists of under-stating the extent of current AI capabilities, and over-emphasizing failure modes in ways that can be (unintentionally) deceptive.

Sam joined me to talk about under-claiming and what it means for AI progress on this episode of the Towards Data Science podcast.

***

Intro music: 

- Artist: Ron Gelinas

- Track Title: Daybreak Chill Blend (original mix)

Until very recently, the study of human disease involved looking at big things — like organs or macroscopic systems — and figuring out when and how they can stop working properly. But that’s all started to change: in recent decades, new techniques have allowed us to look at disease in a much more detailed way, by examining the behaviour and characteristics of single cells.

One class of those techniques now known as single-cell genomics — the study of gene expression and function at the level of single cells. Single-cell genomics is creating new, high-dimensional datasets consisting of tens of millions of cells whose gene expression profiles and other characteristics have been painstakingly measured. And these datasets are opening up exciting new opportunities for AI-powered drug discovery — opportunities that startups are now starting to tackle head-on.

Until recently, AI systems have been narrow — they’ve only been able to perform the specific tasks that they were explicitly trained for. And while narrow systems are clearly useful, the holy grain of AI is to build more flexible, general systems.

2021 has been a wild ride in many ways, but its wildest features might actually be AI-related. We’ve seen major advances in everything from language modeling to multi-modal learning, open-ended learning and even AI alignment.

So, we thought, what better way to take stock of the big AI-related milestones we’ve reached in 2021 than a cross-over episode with our friends over at the Last Week In AI podcast.

***

Intro music:

- Artist: Ron Gelinas

- Track Title: Daybreak Chill Blend (original mix)

Imagine for a minute that you’re running a profitable business, and that part of your sales strategy is to send the occasional mass email to people who’ve signed up to be on your mailing list. For a while, this approach leads to a reliable flow of new sales, but then one day, that abruptly stops. What happened?

You pour over logs, looking for an explanation, but it turns out that the problem wasn’t with your software; it was with your data. Maybe the new intern accidentally added a character to every email address in your dataset, or shuffled the names on your mailing list so that Christina got a message addressed to “John”, or vice-versa. Versions of this story happen surprisingly often, and when they happen, the cost can be significant: lost revenue, disappointed customers, or worse — an irreversible loss of trust.

Today, entire products are being built on top of datasets that aren’t monitored properly for critical failures — and an increasing number of those products are operating in high-stakes situations. That’s why data observability is so important: the ability to  track the origin, transformations and characteristics of mission-critical data to detect problems before they lead to downstream harm.

There once was a time when AI researchers could expect to read every new paper published in the field on the arXiv, but today, that’s no longer the case. The recent explosion of research activity in AI has turned keeping up to date with new developments into a full-time job.

Fortunately, people like YouTuber, ML PhD and sunglasses enthusiast Yannic Kilcher make it their business to distill ML news and papers into a digestible form for mortals like you and me to consume. I highly recommend his channel to any TDS podcast listeners who are interested in ML research — it’s a fantastic resource, and literally the way I finally managed to understand the Attention is All You Need paper back in the day.

Yannic is joined me to talk about what he’s learned from years of following, reporting and doing AI research, including the trends, the challenges and the opportunities that he expects are going to shape the course of AI history in coming years.

--- 

Intro music:

➞ Artist: Ron Gelinas

➞ Track Title: Daybreak Chill Blend (original mix)

It’s no secret that governments around the world are struggling to come up with effective policies to address the risks and opportunities that AI presents. And there are many reasons why that’s happening: many people — including technical people — think they understand what frontier AI looks like, but very few actually do, and even fewer are interested in applying their understanding in a government context, where salaries are low and stock compensation doesn’t even exist.

So there’s a critical policy-technical gap that needs bridging, and failing to address that gap isn’t really an option: it would mean flying blind through the most important test of technological governance the world has ever faced. Unfortunately, policymakers have had to move ahead with regulating and legislating with that dangerous knowledge gap in place, and the result has been less-than-stellar: widely criticized definitions of privacy and explainability, and definitions of AI that create exploitable loopholes are among some of the more concerning results.

AI ethics is often treated as a dry, abstract academic subject. It doesn’t have the kinds of consistent, unifying principles that you might expect from a quantitative discipline like computer science or physics.

But somehow, the ethics rubber has to meet the AI road, and where that happens — where real developers have to deal with real users and apply concrete ethical principles — is where you find some of the most interesting, practical thinking on the topic.

That’s why I wanted to speak with Wendy Foster, the Director of Engineering and Data Science at Shopify. Wendy’s approach to AI ethics is refreshingly concrete and actionable. And unlike more abstract approaches, it’s based on clear principles like user empowerment: the idea that you should avoid forcing users to make particular decisions, and instead design user interfaces that frame AI-recommended actions as suggestions that can be ignored or acted on.

Wendy joined me to discuss her practical perspective on AI ethics, the importance of user experience design for AI products, and how responsible AI gets baked into product at Shopify on this episode of the TDS podcast.

---

Intro music:

- Artist: Ron Gelinas

- Track Title: Daybreak Chill Blend (original mix)

Over the last two years, the capabilities of AI systems have exploded. AlphaFold2, MuZero, CLIP, DALLE, GPT-3 and many other models have extended the reach of AI to new problem classes. There’s a lot to be excited about.

But as we’ve seen in other episodes of the podcast, there’s a lot more to getting value from an AI system than jacking up its capabilities. And increasingly, one of these additional missing factors is becoming trust. You can make all the powerful AIs you want, but if no one trusts their output — or if people trust it when they shouldn’t — you can end up doing more harm than good.

That’s why we invited Ayanna Howard on the podcast. Ayanna is a roboticist, entrepreneur and Dean of the College of Engineering at Ohio State University, where she focuses her research on human-machine interactions and the factors that go into building human trust in AI systems. She joined me to talk about her research, its applications in medicine and education, and the future of human-machine trust.

---

Intro music:

- Artist: Ron Gelinas

- Track Title: Daybreak Chill Blend (original mix)

On the face of it, there’s no obvious limit to the reinforcement learning paradigm: you put an agent in an environment and reward it for taking good actions until it masters a task.

And by last year, RL had achieved some amazing things, including mastering Go, various Atari games, Starcraft II and so on. But the holy grail of AI isn’t to master specific games, but rather to generalize — to make agents that can perform well on new games that they haven’t been trained on before.

Bias gets a bad rap in machine learning. And yet, the whole point of a machine learning model is that it biases certain inputs to certain outputs — a picture of a cat to a label that says “cat”, for example. Machine learning is bias-generation.

So removing bias from AI isn’t an option. Rather, we need to think about which biases are acceptable to us, and how extreme they can be. These are questions that call for a mix of technical and philosophical insight that’s hard to find. Luckily, I’ve managed to do just that by inviting onto the podcast none other than Margaret Mitchell, a former Senior Research Scientist in Google’s Research and Machine Intelligence Group, whose work has been focused on practical AI ethics. And by practical, I really do mean the nuts and bolts of how AI ethics can be baked into real systems, and navigating the complex moral issues that come up when the AI rubber meets the road.

*** 

Intro music:

➞ Artist: Ron Gelinas

➞ Track Title: Daybreak Chill Blend (original mix)

As impressive as they are, language models like GPT-3 and BERT all have the same problem: they’re trained on reams of internet data to imitate human writing. And human writing is often wrong, biased, or both, which means language models are trying to emulate an imperfect target.

Language models often babble, or make up answers to questions they don’t understand. And it can make them unreliable sources of truth. Which is why there’s been increased interest in alternative ways to retrieve information from large datasets — approaches that include knowledge graphs.

Knowledge graphs encode entities like people, places and objects into nodes, which are then connected to other entities via edges, which specify the nature of the relationship between the two. For example, a knowledge graph might contain a node for Mark Zuckerberg, linked to another node for Facebook, via an edge that indicates that Zuck is Facebook’s CEO. Both of these nodes might in turn be connected to dozens, or even thousands of others, depending on the scale of the graph.

Corporate governance of AI doesn’t sound like a sexy topic, but it’s rapidly becoming one of the most important challenges for big companies that rely on machine learning models to deliver value for their customers. More and more, they’re expected to develop and implement governance strategies to reduce the incidence of bias, and increase the transparency of their AI systems and development processes. Those expectations have historically come from consumers, but governments are starting impose hard requirements, too.

The more powerful our AIs become, the more we’ll have to ensure that they’re doing exactly what we want. If we don’t, we risk building AIs that use dangerously creative solutions that have side-effects that could be undesirable, or downright dangerous. Even a slight misalignment between the motives of a sufficiently advanced AI and human values could be hazardous.

AI research is often framed as a kind of human-versus-machine rivalry that will inevitably lead to the defeat — and even wholesale replacement of — human beings by artificial superintelligences that have their own sense of agency, and their own goals.

Divya Siddarth disagrees with this framing. Instead, she argues, this perspective leads us to focus on applications of AI that are neither as profitable as they could be, nor safe enough to prevent us from potentially catastrophic consequences of dangerous AI systems in the long run. And she ought to know: Divya is an associate political economist and social technologist in the Office of the CTO at Microsoft.

Cruise is a self-driving car startup founded in 2013 — at a time when most people thought of self-driving cars as the stuff of science fiction. And yet, just three years later, the company was acquired by GM for over a billion dollars, having shown itself to be a genuine player in the race to make autonomous driving a reality. Along the way, the company has had to navigate and adapt to a rapidly changing technological landscape, mixing and matching old ideas from robotics and software engineering with cutting edge techniques like deep learning.

How can you know that a super-intelligent AI is trying to do what you asked it to do?

Automated weapons mean fewer casualties, faster reaction times, and more precise strikes. They’re a clear win for any country that deploys them. You can see the appeal.