̽��ֱ�� of Cambridge - New York ̽��ֱ��

New datasets will train AI models to think like scientists

sc604 — Mon, 02 Dec 2024 15:59:08 +0000

̽��ֱ��initiative, called Polymathic AI, uses technology like that powering large language models such as OpenAI’s ChatGPT or Google’s Gemini. But instead of ingesting text, the project’s models learn using scientific datasets from across astrophysics, biology, acoustics, chemistry, fluid dynamics and more, essentially giving the models cross-disciplinary scientific knowledge.

“These datasets are by far the most diverse large-scale collections of high-quality data for machine learning training ever assembled for these fields,” said team member Michael McCabe from the Flatiron Institute in New York City. “Curating these datasets is a critical step in creating multidisciplinary AI models that will enable new discoveries about our universe.”

On 2 December, the Polymathic AI team released two of its open-source training dataset collections to the public — a colossal 115 terabytes, from dozens of sources — for the scientific community to use to train AI models and enable new scientific discoveries. For comparison, GPT-3 used 45 terabytes of uncompressed, unformatted text for training, which ended up being around 0.5 terabytes after filtering.

̽��ֱ��full datasets are available to download for free on HuggingFace, a platform hosting AI models and datasets. ̽��ֱ��Polymathic AI team provides further information about the datasets in two papers accepted for presentation at the NeurIPS machine learning conference, to be held later this month in Vancouver, Canada.

“Just as LLMs such as ChatGPT learn to use common grammatical structure across languages, these new scientific foundation models might reveal deep connections across disciplines that we’ve never noticed before,” said Cambridge team lead Dr Miles Cranmer from Cambridge’s Institute of Astronomy. “We might uncover patterns that no human can see, simply because no one has ever had both this breadth of scientific knowledge and the ability to compress it into a single framework.”

AI tools such as machine learning are increasingly common in scientific research, and were recognised in two of this year’s Nobel Prizes. Still, such tools are typically purpose-built for a specific application and trained using data from that field. ̽��ֱ��Polymathic AI project instead aims to develop models that are truly polymathic, like people whose expert knowledge spans multiple areas. ̽��ֱ��project’s team reflects intellectual diversity, with physicists, astrophysicists, mathematicians, computer scientists and neuroscientists.

̽��ֱ��first of the two new training dataset collections focuses on astrophysics. Dubbed the Multimodal Universe, the dataset contains hundreds of millions of astronomical observations and measurements, such as portraits of galaxies taken by NASA’s James Webb Space Telescope and measurements of our galaxy’s stars made by the European Space Agency’s Gaia spacecraft.

̽��ֱ��other collection — called the Well — comprises over 15 terabytes of data from 16 diverse datasets. These datasets contain numerical simulations of biological systems, fluid dynamics, acoustic scattering, supernova explosions and other complicated processes. Cambridge researchers played a major role in developing both dataset collections, working alongside PolymathicAI and other international collaborators.

While these diverse datasets may seem disconnected at first, they all require the modelling of mathematical equations called partial differential equations. Such equations pop up in problems related to everything from quantum mechanics to embryo development and can be incredibly difficult to solve, even for supercomputers. One of the goals of the Well is to enable AI models to churn out approximate solutions to these equations quickly and accurately.

“By uniting these rich datasets, we can drive advancements in artificial intelligence not only for scientific discovery, but also for addressing similar problems in everyday life,” said Ben Boyd, PhD student in the Institute of Astronomy.

Gathering the data for those datasets posed a challenge, said team member Ruben Ohana from the Flatiron Institute. ̽��ֱ��team collaborated with scientists to gather and create data for the project. “ ̽��ֱ��creators of numerical simulations are sometimes sceptical of machine learning because of all the hype, but they’re curious about it and how it can benefit their research and accelerate scientific discovery,” he said.

̽��ֱ��Polymathic AI team is now using the datasets to train AI models. In the coming months, they will deploy these models on various tasks to see how successful these well-rounded, well-trained AIs are at tackling complex scientific problems.

“It will be exciting to see if the complexity of these datasets can push AI models to go beyond merely recognising patterns, encouraging them to reason and generalise across scientific domains,” said Dr Payel Mukhopadhyay from the Institute of Astronomy. “Such generalisation is essential if we ever want to build AI models that can truly assist in conducting meaningful science.”

“Until now, haven’t had a curated scientific-quality dataset cover such a wide variety of fields,” said Cranmer, who is also a member of Cambridge’s Department of Applied Mathematics and Theoretical Physics. “These datasets are opening the door to true generalist scientific foundation models for the first time. What new scientific principles might we discover? We're about to find out, and that's incredibly exciting.”

̽��ֱ��Polymathic AI project is run by researchers from the Simons Foundation and its Flatiron Institute, New York ̽��ֱ��, the ̽��ֱ�� of Cambridge, Princeton ̽��ֱ��, the French Centre National de la Recherche Scientifique and the Lawrence Berkeley National Laboratory.

Members of the Polymathic AI team from the ̽��ֱ�� of Cambridge include PhD students, postdoctoral researchers and faculty across four departments: the Department of Applied Mathematics and Theoretical Physics, the Department of Pure Mathematics and Mathematical Statistics, the Institute of Astronomy and the Kavli Institute for Cosmology.

What can exploding stars teach us about how blood flows through an artery? Or swimming bacteria about how the ocean’s layers mix? A collaboration of researchers, including from the ̽��ֱ�� of Cambridge, has reached a milestone toward training artificial intelligence models to find and use transferable knowledge between fields to drive scientific discovery.

Alex Meng, Aaron Watters and the Well Collaboration

A mosaic of simulations included in the Well collection of datasets

̽��ֱ��text in this work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Images, including our videos, are Copyright © ̽��ֱ�� of Cambridge and licensors/contributors as identified. All rights reserved. We make our image and video content available in a number of ways – on our main website under its Terms and conditions, and on a range of channels including social media that permit your use and sharing of our content under their respective Terms.

Yes

What’s going on in our brains when we plan?

sc604 — Tue, 11 Jun 2024 09:46:21 +0000

In pausing to think before making an important decision, we may imagine the potential outcomes of different choices we could make. While this ‘mental simulation’ is central to how we plan and make decisions in everyday life, how the brain works to accomplish this is not well understood.

An international team of scientists has now uncovered neural mechanisms used in planning. Their results, published in the journal Nature Neuroscience, suggest that an interplay between the brain’s prefrontal cortex and hippocampus allows us to imagine future outcomes to guide our decisions.

“ ̽��ֱ��prefrontal cortex acts as a ‘simulator,’ mentally testing out possible actions using a cognitive map stored in the hippocampus,” said co-author Marcelo Mattar from New York ̽��ֱ��. “This research sheds light on the neural and cognitive mechanisms of planning—a core component of human and animal intelligence. A deeper understanding of these brain mechanisms could ultimately improve the treatment of disorders affecting decision-making abilities.”

̽��ֱ��roles of both the prefrontal cortex—used in planning and decision-making—and hippocampus—used in memory formation and storage—have long been established. However, their specific duties in deliberative decision-making, which are the types of decisions that require us to think before acting, are less clear.

To illuminate the neural mechanisms of planning, Mattar and his colleagues—Kristopher Jensen from ̽��ֱ�� College London and Professor Guillaume Hennequin from Cambridge’s Department of Engineering —developed a computational model to predict brain activity during planning. They then analysed data from both humans and rats to confirm the validity of the model—a recurrent neural network (RNN), which learns patterns based on incoming information.

̽��ֱ��model took into account existing knowledge of planning and added new layers of complexity, including ‘imagined actions,’ thereby capturing how decision-making involves weighing the impact of potential choices—similar to how a chess player envisions sequences of moves before committing to one. These mental simulations of potential futures, modelled as interactions between the prefrontal cortex and hippocampus, enable us to rapidly adapt to new environments, such as taking a detour after finding a road is blocked.

̽��ֱ��scientists validated this computational model using both behavioural and neural data. To assess the model’s ability to predict behaviour, the scientists conducted an experiment measuring how humans navigated an online maze on a computer screen and how long they had to think before each step.

To validate the model’s predictions about the role of the hippocampus in planning, they analysed neural recordings from rodents navigating a physical maze configured in the same way as in the human experiment. By giving a similar task to humans and rats, the researchers could draw parallels between the behavioural and neural data—an innovative aspect of this research.

“Allowing neural networks to decide for themselves when to 'pause and think' was a great idea, and it was surprising to see that in situations where humans spend time pondering what to do next, so do these neural networks,” said Hennequin.

̽��ֱ��experimental results were consistent with the computational model, showing an intricate interaction between the prefrontal cortex and hippocampus. In the human experiments, participants’ brain activity reflected more time thinking before acting in navigating the maze. In the experiments with laboratory rats, the animals’ neural responses in moving through the maze resembled the model’s simulations.

“Overall, this work provides foundational knowledge on how these brain circuits enable us to think before we act in order to make better decisions,” said Mattar. “In addition, a method in which both human and animal experimental participants and RNNs were all trained to perform the same task offers an innovative and foundational way to gain insights into behaviours.”

“This new framework will enable systematic studies of thinking at the neural level,” said Hennequin. “This will require a concerted effort from neurophysiologists and theorists, and I'm excited about the discoveries that lie ahead.”

Reference:
Kristopher T. Jensen, Guillaume Hennequin & Marcelo G. Mattar. ‘A recurrent network model of planning explains hippocampal replay and human behavior.’ Nature Neuroscience (2024). DOI: 10.1038/s41593-024-01675-7

Adapted from an NYU press release.

Study uncovers how the brain simulates possible future actions by drawing from our stored memories.

Andriy Onufriyenko via Getty Images

Metaverse portrait

Yes

Scientists begin building AI for scientific discovery using tech behind ChatGPT

sc604 — Fri, 13 Oct 2023 10:55:15 +0000

While ChatGPT deals in words and sentences, the team’s AI will learn from numerical data and physics simulations from across scientific fields to aid scientists in modelling everything from supergiant stars to the Earth’s climate.

̽��ֱ��team launched the initiative, called Polymathic AI earlier this week, alongside the publication of a series of related scientific papers on the arXiv.org open access repository.

“This will completely change how people use AI and machine learning in science,” said Polymathic AI principal investigator Shirley Ho, a group leader at the Flatiron Institute’s Center for Computational Astrophysics in New York City.

̽��ֱ��idea behind Polymathic AI “is similar to how it’s easier to learn a new language when you already know five languages,” said Ho.

Starting with a large, pre-trained model, known as a foundation model, can be both faster and more accurate than building a scientific model from scratch. That can be true even if the training data isn’t obviously relevant to the problem at hand.

“It’s been difficult to carry out academic research on full-scale foundation models due to the scale of computing power required,” said co-investigator Miles Cranmer, from Cambridge’s Department of Applied Mathematics and Theoretical Physics and Institute of Astronomy. “Our collaboration with Simons Foundation has provided us with unique resources to start prototyping these models for use in basic science, which researchers around the world will be able to build from – it’s exciting.”

“Polymathic AI can show us commonalities and connections between different fields that might have been missed,” said co-investigator Siavash Golkar, a guest researcher at the Flatiron Institute’s Center for Computational Astrophysics. “In previous centuries, some of the most influential scientists were polymaths with a wide-ranging grasp of different fields. This allowed them to see connections that helped them get inspiration for their work. With each scientific domain becoming more and more specialised, it is increasingly challenging to stay at the forefront of multiple fields. I think this is a place where AI can help us by aggregating information from many disciplines.”

̽��ֱ��Polymathic AI team includes researchers from the Simons Foundation and its Flatiron Institute, New York ̽��ֱ��, the ̽��ֱ�� of Cambridge, Princeton ̽��ֱ�� and the Lawrence Berkeley National Laboratory. ̽��ֱ��team includes experts in physics, astrophysics, mathematics, artificial intelligence and neuroscience.

Scientists have used AI tools before, but they’ve primarily been purpose-built and trained using relevant data. “Despite rapid progress of machine learning in recent years in various scientific fields, in almost all cases, machine learning solutions are developed for specific use cases and trained on some very specific data,” said co-investigator Francois Lanusse, a cosmologist at the Centre national de la recherche scientifique (CNRS) in France. “This creates boundaries both within and between disciplines, meaning that scientists using AI for their research do not benefit from information that may exist, but in a different format, or in a different field entirely.”

Polymathic AI’s project will learn using data from diverse sources across physics and astrophysics (and eventually fields such as chemistry and genomics, its creators say) and apply that multidisciplinary savvy to a wide range of scientific problems. ̽��ֱ��project will “connect many seemingly disparate subfields into something greater than the sum of their parts,” said project member Mariel Pettee, a postdoctoral researcher at Lawrence Berkeley National Laboratory.

“How far we can make these jumps between disciplines is unclear,” said Ho. “That’s what we want to do — to try and make it happen.”

ChatGPT has well-known limitations when it comes to accuracy (for instance, the chatbot says 2,023 times 1,234 is 2,497,582 rather than the correct answer of 2,496,382). Polymathic AI’s project will avoid many of those pitfalls, Ho said, by treating numbers as actual numbers, not just characters on the same level as letters and punctuation. ̽��ֱ��training data will also use real scientific datasets that capture the physics underlying the cosmos.

Transparency and openness are a big part of the project, Ho said. “We want to make everything public. We want to democratise AI for science in such a way that, in a few years, we’ll be able to serve a pre-trained model to the community that can help improve scientific analyses across a wide variety of problems and domains.”

An international team of scientists, including from the ̽��ֱ�� of Cambridge, have launched a new research collaboration that will leverage the same technology behind ChatGPT to build an AI-powered tool for scientific discovery.

Yuichiro Chino via Getty Images

Network and data connection on a dark blue background.

̽��ֱ��text in this work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Images, including our videos, are Copyright © ̽��ֱ�� of Cambridge and licensors/contributors as identified. All rights reserved. We make our image and video content available in a number of ways – as here, on our main website under its Terms and conditions, and on a range of channels including social media that permit your use and sharing of our content under their respective Terms.

Yes

Companies’ zero-deforestation commitments have potential to halve cattle-driven deforestation in Brazilian Amazon

jg533 — Wed, 19 Apr 2023 23:01:02 +0000

Cattle-rearing is the biggest cause of tropical deforestation in the Amazon - and the world.

A study has found that some of the world’s largest slaughterhouses reduced cattle-driven deforestation in the Amazon by 15% - equivalent to sparing 7,000km2 of forest from clearance (4.5 times the size of London) - through their commitment to zero-deforestation policies between 2010 and 2018.

If these policies were fully implemented and adopted across all cattle companies operating in the Amazon, 24,000km2 of forest (an area larger than Wales) could have been spared over this time, effectively halving cattle-driven deforestation in Brazil.

Deforestation is the second largest contributor to global greenhouse gas emissions after fossil fuel use, and the Brazilian Amazon is the world’s deforestation hotspot.

Zero-deforestation commitments are supply chain policies aiming to ensure production of goods does not involve deforestation, by identifying and dropping suppliers who produce on areas recently cleared of natural vegetation. ̽��ֱ��commitments have been signed by many leading UK beef retailers, including the supermarkets Tesco, Sainsbury’s, and Waitrose.

Although the UK imported 60 million tonnes of Brazilian beef in 2017, according to the National Beef Association the UK is 75% self-sufficient for beef. Many British companies are increasingly turning away from Brazilian beef due to the perceived risk of deforestation. But the researchers argue that this is not the best approach.

“We can benefit the climate by eating less meat in general. But when it comes to deforestation, the solution is not to avoid beef from specific countries – because then we lose our power to make a difference in those places,” said Professor Rachael Garrett, Moran Professor of Conservation and Development at the ̽��ֱ�� of Cambridge Conservation Research Institute, senior author of the report.

She added: “If we do eat imported beef, we should buy it from retailers that are trying to improve cattle production systems in Brazil and elsewhere. If enough countries join the UK and EU in purchasing only deforestation-free beef it’s likely to have a positive impact on the planet by reducing deforestation.”

̽��ֱ��results are published today in the journal Global Environmental Change.

Due to the complexity of Brazilian supply chains and incomplete availability of public records, it has been challenging until now to determine how much of the cattle in any given region was being purchased by companies with zero-deforestation commitments. This impeded efforts to analyse the effectiveness of zero-deforestation policies linked to beef and leather goods - such as shoes and handbags.

̽��ֱ��researchers traced the links between farming regions, slaughterhouses and companies with zero-deforestation commitments in the Brazilian Amazon cattle sector, to see how these links influenced deforestation.

An agreement called G4 is the most widespread and strongly implemented zero-deforestation commitment for cattle in the Brazilian Amazon – accounting for over 99% of cattle exports. ̽��ֱ��study focused on companies that have adopted the G4 Agreement, and found they were associated with substantial reductions in deforestation.

“We’ve shown that zero-deforestation policies are having an important - and measurable – impact in protecting forests, and that with widespread adoption and rigorous implementation they could achieve a lot more,” said Garrett.

She added: “Even reducing deforestation by 15% is a huge amount. But this result shows that supply chain policies have significant limitations, and we need to couple them with more visionary approaches to help countries like Brazil improve their agricultural systems.”

̽��ֱ��researchers say a mix of interventions by the private and public sector is needed to improve cattle-rearing practices and help eliminate deforestation in countries like Brazil.

Public sector interventions could include support for alternative economic activities, and financial incentives or greater pressure to avoid deforestation from the Brazilian government.

“With this evidence, supermarkets can use their influence to help improve Brazilian cattle production. But more needs to be done to improve the rigour of corporate policies and the market coverage of policy adopters, even in relatively well-covered regions such as the Brazilian Amazon,” said Dr Sam Levy at ETH Zurich and New York ̽��ֱ��, lead author of the report.

Cattle production for beef and leather is the cause of over 70% of all deforestation in the Amazon – much of which is illegal. Zero-deforestation commitments cover 82% of beef exported from the Brazilian Amazon for trade internationally – but a large amount of beef production destined for Brazil’s domestic markets is not covered.

Deforestation causes the loss of diverse animal and plant life, threatens the livelihoods of indigenous groups, and increases inequality and conflict.

In 2021, the COP26 Glasgow Leaders’ Declaration on Forests and Land Use committed to halt and reverse deforestation by 2030. It was signed by over 100 countries, representing 85% of global forests.

̽��ֱ��research was funded by the National Science Foundation, Gordon and Betty Moore Foundation, and the European Research Council.

Reference

Levy, S A et al: ‘Deforestation in the Brazilian Amazon could be halved by scaling up implementation of zero-deforestation cattle commitments.’ Global Environmental Change, April 2023. DOI: 10.1016/j.gloenvcha.2023.102671

Study shows better adoption and implementation of company supply chain policies for Brazilian beef and leather could significantly reduce carbon emissions

If we do eat imported beef, we should buy it from retailers that are trying to improve cattle production systems in Brazil and elsewhere.

Rachael Garrett

Bruno Kelly / Greenpeace

Cattle in the Amazon

̽��ֱ��text in this work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Images, including our videos, are Copyright © ̽��ֱ�� of Cambridge and licensors/contributors as identified. All rights reserved. We make our image and video content available in a number of ways – as here, on our main website under its Terms and conditions, and on a range of channels including social media that permit your use and sharing of our content under their respective Terms.

Yes

Licence type:

Attribution-Noncommerical

Rewarding accuracy instead of partisan pandering reduces political divisions over the truth

fpjl2 — Mon, 06 Mar 2023 16:17:56 +0000

Offering a tiny cash reward for accuracy, or even briefly appealing to personal integrity, can increase people’s ability to tell the difference between misinformation and the truth, according to a new study.

̽��ֱ��findings suggest that fake news thrives on social media not only because people are tricked into believing it, but also due to a motivational imbalance: users have more incentive to get clicks and likes than to spread accurate content.

Social psychologists from the ̽��ֱ�� of Cambridge and New York ̽��ֱ�� argue that their study, published in the journal Nature Human Behaviour, highlights the “perverse incentives” driving shares on social media – particularly in “divisive political climates” such as the United States.

They say the psychological pull of pandering to one’s own “in-group” by attacking the other side of a social and political divide is a significant – and often neglected – factor for why so many believe and choose to spread misinformation, or disbelieve accurate news.

̽��ֱ��study involved four experiments with a total of over 3,300 people from the United States, with equal numbers of Democrats and Republicans. ̽��ֱ��researchers offered half of participants up to one US dollar if they correctly pointed out true or false headlines, and compared the results to those offered no incentive.

This tiny sum was enough to make people 31% better at discerning true from fake news. ̽��ֱ��best results came when participants were asked to identify accurate news that benefited the opposing political party.

In fact, the financial incentive reduced partisan division between Republican and Democrat over the truthfulness of news by around 30%. ̽��ֱ��majority of this shift occurred on the Republican side.

For example, the offer of up to a dollar made Republicans 49% more likely to report that the accurate Associated Press headline ‘Facebook removes Trump ads with symbols once used by Nazis’ was indeed true. A dollar made Democrats 20% more likely to report the Reuters headline 'Plant a trillion trees: U.S. Republicans offer fossil-fuel friendly climate fix' as accurate.

However, in another experiment, researchers inverted the set-up to “mirror the social media environment” by paying participants to identify the headlines likely to get the best reception from members of the same political party. ̽��ֱ��ability to spot misinformation reduced by 16%.

“This is not just about ignorance of facts among the public. It is about a social media business model that rewards the spread of divisive content regardless of accuracy,” said lead author Dr Steve Rathje, who conducted the work while he was a Gates Cambridge Scholar.

“By motivating people to be accurate instead of appealing to those in the same political group, we found greater levels of agreement between Republicans and Democrats about what is actually true.”

Previous research by the same team has shown that attacking political rivals is one of the most effective ways to go viral on Twitter and Facebook.

“Shifting the motivations to post on social media could help rebuild some of the shared reality lost to political polarisation in many nations, including the United States,” said senior author Prof Sander van der Linden, director of the ̽��ֱ�� of Cambridge’s Social Decision-Making Lab.

In one of the study’s experiments, half the participants were simply exposed to a short piece of text reminding them that people value truth, and falsehoods can hurt reputations. They were also told they would receive feedback on accuracy rates.

While this did not have the same effect as a small pay out, it still increased the perceived accuracy of true but politically inconvenient news by 25% compared to a control group.

“A short piece of text nudging users to consider the social value of truth could be deployed at scale by social media corporations,” said van der Linden.

Jay Van Bavel, Professor of Psychology at New York ̽��ֱ�� and co-author of the study, said: “It is not possible to pay everyone on the internet to share more accurate information. However, we can change aspects of social media platform design to help motivate people to share content they know to be accurate.”

Providing incentives improved the accuracy of news judgements across the political spectrum, but had a much stronger effect on Republican voters.

̽��ֱ��team point to previous research showing that Republicans tend to believe in and share more misinformation than Democrats. In the latest study, payment incentives brought Republicans far closer to the accuracy levels of Democrats – shrinking the political divide.

“Recent lawsuits have revealed that Fox News hosts shared false claims about ‘stolen’ elections to retain viewers, despite privately disavowing these conspiracy theories. Republican media ecosystems have proved more willing to harness misinformation for profit in recent years,” said Van der Linden, author of the new book Foolproof: why we fall for misinformation and how to build immunity.

Researchers argue that the findings hold lessons for social media companies and the “perverse incentives” driving political polarisation online.

Shifting the motivations to post on social media could help rebuild some of the shared reality lost to political polarisation

Sander van der Linden

̽��ֱ��text in this work is licensed under a Creative Commons Attribution 4.0 International License. Images, including our videos, are Copyright © ̽��ֱ�� of Cambridge and licensors/contributors as identified. All rights reserved. We make our image and video content available in a number of ways – as here, on our main website under its Terms and conditions, and on a range of channels including social media that permit your use and sharing of our content under their respective Terms.

Yes

How landscapes and landforms ‘remember’ or ‘forget’ their initial formations

sc604 — Fri, 27 Apr 2018 12:52:44 +0000

Crescent dunes and meandering rivers can ‘forget’ their initial shapes as they are carved and reshaped by wind and water while other landforms keep a memory of their past shape, suggests new research.

“Asking how these natural sculptures come to be is more than mere curiosity because locked in their shapes are clues to the history of an environment,” said Leif Ristroph from New York ̽��ֱ�� and the senior author of the paper, which is published in the journal Physical Review Fluids. “We found that some shapes keep a ‘memory’ of their starting conditions as they develop while others ‘forget’ the past entirely and take on new forms.” This understanding is important in geological dating and in understanding how landscapes form.

Shape ‘memory’ and its ‘loss’—or the retention of or departure from earlier formations—are key issues in geomorphology, the field of study that tries to explain landforms and the developing face of the Earth and other celestial surfaces. ̽��ֱ��morphology, or shape of a landscape, is the first and most direct clue into its history and serves as a scientific window for a range of questions—such as inferring flowing water on Mars in the past as well as present-day erosion channels and river islands.

“ ̽��ֱ��answer to the question ‘What’s in a shape?’ hinges on this memory property,” said first author Dr Megan Davies Wykes, a postdoctoral researcher in Cambridge’s Department of Applied Mathematics and Theoretical Physics, who completed the work while she was based at NYU.

To shed light on these phenomena, the researchers replicated nature’s dissolvable minerals—such as limestone—with a ready-made stand-in: pieces of hard candy. Specifically, they sought to understand how the candy dissolved to take different forms when placed in water.

To mimic different environmental conditions, they cast the candy into different initial shapes, which led to different flow conditions as the surface dissolved. Their results showed that when the candy dissolved most strongly from its lower surface, it tended to retain its overall shape—reflecting a near-perfect memory. By contrast, when dissolved from its upper surface, the candy tended to erase or ‘forget’ any given initial shape and form an upward spike structure.

̽��ֱ��key difference, the team found, is the type of water flow that ‘licks’ and reshapes the candy. Turbulent flows on the underside tend to dissolve the candy at a uniform rate and thus preserve the shape. ̽��ֱ��smooth flow on an upper surface, however, carries the dissolved material from one location to the next, which changes the dissolving rate and leads to changes in shape.

“Candy in water may seem like a far cry from geology, but there are in fact whole landscapes carved from minerals dissolving in water, their shapes revealed later when the water table recedes,” said Ristroph. “Caves, sinkholes, stone pillars and other types of craggy terrain are born this way.”

Reference:
Megan S. Davies Wykes et al. ‘Self-sculpting of a dissolvable body due to gravitational convection.’ Physical Review Fluids (2018). DOI: 10.1103/PhysRevFluids.3.043801

Adapted from an NYU press release.

Video: Side-view photograph of candy body (initially a sphere). ̽��ֱ��upper surface remains smooth while the undersurface becomes pitted and dissolves several times faster.

Laboratory findings point to what affects the development of nature’s shapes.

̽��ֱ��answer to the question ‘What’s in a shape?’ hinges on this memory property.

Megan Davies Wykes

Dissolving candy

Chiara Ferroni on Unsplash

Sea of dunes

̽��ֱ��text in this work is licensed under a Creative Commons Attribution 4.0 International License. For image use please see separate credits above.

Yes

AI crossword-solving application could make machines better at understanding language

sc604 — Mon, 07 Mar 2016 09:38:00 +0000

Researchers have designed a web-based platform which uses artificial neural networks to answer standard crossword clues better than existing commercial products specifically designed for the task. ̽��ֱ��system could help machines understand language more effectively.

In tests against commercial crossword-solving software, the system, designed by researchers from the UK, US and Canada, was more accurate at answering clues that were single words (e.g. ‘culpability’ – guilt), a short combination of words (e.g. ‘devil devotee’ – Satanist), or a longer sentence or phrase (e.g. ‘French poet and key figure in the development of Symbolism’ – Baudelaire). ̽��ֱ��system can also be used a ‘reverse dictionary’ in which the user describes a concept and the system returns possible words to describe that concept.

̽��ֱ��researchers used the definitions contained in six dictionaries, plus Wikipedia, to ‘train’ the system so that it could understand words, phrases and sentences – using the definitions as a bridge between words and sentences. Their results, published in the journal Transactions of the Association for Computational Linguistics, suggest that a similar approach may lead to improved output from more general language understanding and dialogue systems and information retrieval engines in general. All of the code and data behind the application has been made freely available for future research.

“Over the past few years, there’s been a mini-revolution in machine learning,” said Felix Hill of the ̽��ֱ�� of Cambridge’s Computer Laboratory, one of the paper’s authors. “We’re seeing a lot more usage of deep learning, which is especially useful for language perception and speech recognition.”

Deep learning refers to an approach in which artificial neural networks with little or no prior ‘knowledge’ are trained to recreate human abilities using massive amounts of data. For this particular application, the researchers used dictionaries – training the model on hundreds of thousands of definitions of English words, plus Wikipedia.

“Dictionaries contain just about enough examples to make deep learning viable, but we noticed that the models get better and better the more examples you give them,” said Hill. “Our experiments show that definitions contain a valuable signal for helping models to interpret and represent the meaning of phrases and sentences.”

Working with Anna Korhonen from the Cambridge’s Department of Theoretical and Applied Linguistics, and researchers from the Université de Montréal and New York ̽��ֱ��, Hill used the model as a way of bridging the gap between machines that understand the meanings of individual words and machines that can understand the meanings of phrases and sentences.

“Despite recent progress in AI, problems involving language understanding are particularly difficult, and our work suggests many possible applications of deep neural networks to language technology,” said Hill. “One of the biggest challenges in training computers to understand language is recreating the many rich and diverse information sources available to humans when they learn to speak and read.”

However, there is still a long way to go. For instance, when Hill’s system receives a query, the machine has no idea about the user’s intention or the wider context of why the question is being asked. Humans, on the other hand, can use their background knowledge and signals like body language to figure out the intent behind the query.

Hill describes recent progress in learning-based AI systems in terms of behaviourism and cognitivism: two movements in psychology that effect how one views learning and education. Behaviourism, as the name implies, looks at behaviour without looking at what the brain and neurons are doing, while cognitivism looks at the mental processes that underlie behaviour. Deep learning systems like the one built by Hill and his colleagues reflect a cognitivist approach, but for a system to have something approaching human intelligence, it would have to have a little of both.

“Our system can’t go too far beyond the dictionary data on which it was trained, but the ways in which it can are interesting, and make it a surprisingly robust question and answer system – and quite good at solving crossword puzzles,” said Hill. While it was not built with the purpose of solving crossword puzzles, the researchers found that it actually performed better than commercially-available products that are specifically engineered for the task.

Existing commercial crossword-answering applications function in a similar way to a Google search, with one system able to reference over 1100 dictionaries. While this approach has advantages if you want to look up a definition verbatim, it works less well when you input a question or query that the model has never seen in training. It also makes it incredibly ‘heavy’ in terms of the amount of memory it requires. “Traditional approaches are like lugging many heavy dictionaries around with you, whereas our neural system is incredibly light,” said Hill.

According to the researchers, the results show the effectiveness of definition-based training for developing models that understand phrases and sentences. They are currently looking at ways of enhancing their system, specifically by combining it with more behaviourist-style models of language learning and linguistic interaction.

Reference:
Hill, Felix et al. Learning to Understand Phrases by Embedding the Dictionary. Transactions of the Association for Computational Linguistics, [S.l.], v. 4, p. 17-30, feb. 2016. ISSN 2307-387X.

A web-based machine language system solves crossword puzzles far better than commercially-available products, and may help machines better understand language.

Despite recent progress in AI, problems involving language understanding are particularly difficult.

Felix Hill

Beth

Crossword

̽��ֱ��text in this work is licensed under a Creative Commons Attribution 4.0 International License. For image use please see separate credits above.

Yes

Licence type:

Attribution-Noncommerical