Cloud Hosting

Image Credits: Google DeepMind

Keeping up with an industry as fast-moving asAIis a tall order. So until an AI can do it for you, heres a handy roundup of recent stories in the world of machine learning, along with notable research and experiments we didnt cover on their own.

This week in AI, DeepMind, the Google-owned AI R&D lab, released a paper proposing a framework for evaluating the societal and ethical risks of AI systems.

The timing of the paper which calls for varying levels of involvement from AI developers, app developers and broader public stakeholders in evaluating and auditing AI isnt accidental.

Next week is the AI Safety Summit, a U.K.-government-sponsored event thatll bring together international governments, leading AI companies, civil society groups and experts in research to focus on how best to manage risks from the most recent advances in AI, including generative AI (e.g. ChatGPT, Stable Diffusion and so on). There, the U.K. is planning to introduce a global advisory group on AI loosely modeled on the U.N.s Intergovernmental Panel on Climate Change, comprising a rotating cast of academics who will write regular reports on cutting-edge developments in AI and their associated dangers.

DeepMind is airing its perspective, very visibly, ahead of on-the-ground policy talks at the two-day summit. And, to give credit where its due, the research lab makes a few reasonable (if obvious) points, such as calling for approaches to examine AI systems at the point of human interaction and the ways in which these systems might be used and embedded in society.

But in weighing DeepMinds proposals, its informative to look at how the labs parent company, Google, scores in a recent study released by Stanford researchers that ranks 10 major AI models on how openly they operate.

Rated on 100 criteria, including whether its maker disclosed the sources of its training data, information about the hardware it used, the labor involved in training and other details, PaLM 2, one of Googles flagship text-analyzing AI models, scores a measly 40%.

Now, DeepMind didnt develop PaLM 2 at least not directly. But the lab hasnt historically been consistently transparent about its own models, and the fact that its parent company falls short on key transparency measures suggests that theres not much top-down pressure for DeepMind to do better.

On the other hand, in addition to its public musings about policy, DeepMind appears to be taking steps to change the perception that its tight-lipped about its models architectures and inner workings. The lab, along with OpenAI and Anthropic, committed several months ago to providing the U.K. government early or priority access to its AI models to support research into evaluation and safety.

The question is, is this merely performative? No one would accuse DeepMind of philanthropy, after all the lab rakes in hundreds of millions of dollars in revenue each year, mainly by licensing its work internally to Google teams.

Perhaps the labs next big ethics test is Gemini, its forthcoming AI chatbot, which DeepMind CEO Demis Hassabis has repeatedly promised will rival OpenAIs ChatGPT in its capabilities. Should DeepMind wish to be taken seriously on the AI ethics front, itll have to fully and thoroughly detail Geminis weaknesses and limitations not just its strengths. Well certainly be watching closely to see how things play out over the coming months.

Here are some other AI stories of note from the past few days:

Machine learning models are constantly leading to advances in the biological sciences. AlphaFold and RoseTTAFold were examples of how a stubborn problem (protein folding) could be, in effect, trivialized by the right AI model. Now David Baker (creator of the latter model) and his labmates have expanded the prediction process to include more than just the structure of the relevant chains of amino acids. After all, proteins exist in a soup of other molecules and atoms, and predicting how theyll interact with stray compounds or elements in the body is essential to understanding their actual shape and activity. RoseTTAFold All-Atom is a big step forward for simulating biological systems.

Having a visual AI enhance lab work or act as a learning tool is also a great opportunity. The SmartEM project from MIT and Harvard put a computer vision system and ML control system inside a scanning electron microscope, which together drive the device to examine a specimen intelligently. It can avoid areas of low importance, focus on interesting or clear ones, and do smart labeling of the resulting image as well.

Using AI and other high tech tools for archaeological purposes never gets old (if you will) for me. Whether its lidar revealing Mayan cities and highways or filling in the gaps of incomplete ancient Greek texts, its always cool to see. And this reconstruction of a scroll thought destroyed in the volcanic eruption that leveled Pompeii is one of the most impressive yet.

University of Nebraska-Lincoln CS student Luke Farritor trained a machine learning model to amplify the subtle patterns on scans of the charred, rolled-up papyrus that are invisible to the naked eye. His was one of many methods being attempted in an international challenge to read the scrolls, and it could be refined to perform valuable academic work. Lots more info at Nature here. What was in the scroll, you ask? So far, just the word purple but even that has the papyrologists losing their minds.

Another academic victory for AI is in this system for vetting and suggesting citations on Wikipedia. Of course, the AI doesnt know what is true or factual, but it can gather from context what a high-quality Wikipedia article and citation looks like, and scrape the site and web for alternatives. No one is suggesting we let the robots run the famously user-driven online encyclopedia, but it could help shore up articles for which citations are lacking or editors are unsure.

Language models can be fine-tuned on many topics, and higher math is surprisingly one of them. Llemma is a new open model trained on mathematical proofs and papers that can solve fairly complex problems. Its not the first Google Researchs Minerva is working on similar capabilities but its success on similar problem sets and improved efficiency show that open models (for whatever the term is worth) are competitive in this space. Its not desirable that certain types of AI should be dominated by private models, so replication of their capabilities in the open is valuable even if it doesnt break new ground.

Troublingly, Meta is progressing in its own academic work toward reading minds but as with most studies in this area, the way its presented rather oversells the process. In a paper called Brain decoding: Toward real-time reconstruction of visual perception, it may seem a bit like theyre straight up reading minds.

But its a little more indirect than that. By studying what a high-frequency brain scan looks like when people are looking at images of certain things, like horses or airplanes, the researchers are able to then perform reconstructions in near real time of what they think the person is thinking of or looking at. Still, it seems likely that generative AI has a part to play here in how it can create a visual expression of something even if it doesnt correspond directly to scans.

Should we be using AI to read peoples minds, though, if it ever becomes possible? Ask DeepMind see above.

Last up, a project at LAION thats more aspirational than concrete right now, but laudable all the same. Multilingual Contrastive Learning for Audio Representation Acquisition, or CLARA, aims to give language models a better understanding of the nuances of human speech. You know how you can pick up on sarcasm or a fib from sub-verbal signals like tone or pronunciation? Machines are pretty bad at that, which is bad news for any human-AI interaction. CLARA uses a library of audio and text in multiple languages to identify some emotional states and other non-verbal speech understanding cues.

View original post here:
This week in AI: Can we trust DeepMind to be ethical? - TechCrunch

Deadly bioweapons, automated cybersecurity attacks, powerful AI models escaping human control. Those are just some of the potential threats posed by artificial intelligence, according to a new UK government report. It was released to help set the agenda for an international summit on AI safety to be hosted by the UK next week. The report was compiled with input from leading AI companies such as Googles DeepMind unit and multiple UK government departments, including intelligence agencies.

Joe White, the UKs technology envoy to the US, says the summit provides an opportunity to bring countries and leading AI companies together to better understand the risks posed by the technology. Managing the potential downsides of algorithms will require old-fashioned organic collaboration, says White, who helped plan next weeks summit. These arent machine-to-human challenges, White says. These are human-to-human challenges.

UK prime minister Rishi Sunak will make a speech tomorrow about how, while AI opens up opportunities to advance humanity, its important to be honest about the new risks it creates for future generations.

The UKs AI Safety Summit will take place on November 1 and 2 and will mostly focus on the ways people can misuse or lose control of advanced forms of AI. Some AI experts and executives in the UK have criticized the events focus, saying the government should prioritize more near-term concerns, such as helping the UK compete with global AI leaders like the US and China.

Some AI experts have warned that a recent uptick in discussion about far-off AI scenarios, including the possibility of human extinction, could distract regulators and the public from more immediate problems, such as biased algorithms or AI technology strengthening already dominant companies.

The UK report released today considers the national security implications of large language models, the AI technology behind ChatGPT. White says UK intelligence agencies are working with the Frontier AI Task Force, a UK government expert group, to explore scenarios like what could happen if bad actors combined a large language model with secret government documents. One doomy possibility discussed in the report suggests a large language model that accelerates scientific discovery could also boost projects trying to create biological weapons.

This July, Dario Amodei, CEO of AI startup Anthropic, told members of the US Senate that within the next two or three years it could be possible for a language model to suggest how to carry out large-scale biological weapons attacks. But White says the report is a high-level document that is not intended to serve as a shopping list of all the bad things that can be done.

The UK report also discusses how AI could escape human control. If people become used to handing over important decisions to algorithms it becomes increasingly difficult for humans to take control back, the report says. But the likelihood of these risks remains controversial, with many experts thinking the likelihood is very low and some arguing a focus on risk distracts from present harms.

In addition to government agencies, the report released today was reviewed by a panel including policy and ethics experts from Googles DeepMind AI lab, which began as a London AI startup and was acquired by the search company in 2014, and Hugging Face, a startup developing open source AI software.

Yoshua Bengio, one of three godfathers of AI who won the highest award in computing, the Turing Award, for machine-learning techniques central to the current AI boom was also consulted. Bengio recently said his optimism about the technology he helped foster has soured and that a new humanity defense organization is needed to help keep AI in check.

View original post here:
The UK Lists Top Nightmare AI Scenarios Ahead of Its Big Tech ... - WIRED