̽��ֱ�� of Cambridge - scientific knowledge

Opinion: the future of science is automation

Anonymous — Mon, 26 Feb 2024 13:02:43 +0000

Thanks to the widespread availability of food and medical care, the ability to travel, and many other scientific and technological developments, billions of people today are living better lives than kings of centuries past. It is deeply surprising to me how little appreciated this astonishing fact is.

Of course, despite all the progress we’ve made, the world faces many challenges in the 21st century: climate change, pandemics, poverty and cancer, to name just a few.

If all the countries in the world could join together to share technology and resources, we might be to deal with and overcome these challenges. However, history presents no example of such collaboration, and the current geopolitical situation does not offer much in the way of hope.

Our best hope of dealing with these challenges is to make science and technology more productive. ̽��ֱ��only feasible way to achieve this is through the integration of Artificial Intelligence (AI) and laboratory automation.

AI systems already possess superhuman scientific powers. They can remember massive volumes of facts and learn from huge datasets. They can execute flawless logical reasoning, and near optimal probabilistic reasoning. They are can read every scientific paper, indeed everything ever written. These powers are complimentary to human scientists.

When the scientific method was developed in the 17th century, one of the core insights was the need to conduct experiments in the physical world, not just to think.

Today, laboratory automation is steadily advancing, and robots can now carry out most of the laboratory tasks that humans can. We are also now seeing the emergence of the ‘Cloud Lab’ concept. ̽��ֱ��idea is to provide laboratory automation at scale and remotely, with scientists sending their samples to the cloud lab, using a computer interface to design and execute their experiments.

And then there are AI Scientists: AI systems integrated with laboratory automations that are capable of carrying out the closed-loop automation of scientific research (aka 'Robot Scientists', 'Self-driving Labs'). These systems automatically originate hypotheses to explain observations, devise experiments to test these hypotheses, physically run these experiments using laboratory robotics, interpret the results, and then repeat the cycle.

AI Scientists can work cheaper, faster, more accurately, and longer than humans. They can also be easily multiplied. As the experiments are conceived and executed automatically by computer, it’s possible to completely capture and digitally curate all aspects of the scientific process, making the science more reproducible. There are now around 100 AI Scientists around the world, working in areas from quantum mechanics to astronomy, from chemistry to medicine.

Within the last year or so the world has been stunned by the success of Large Language Models (LLMs) such as ChatGPT, which have achieved breakthrough performance on a wide range of conversation-based tasks. LLMs are surprisingly strong absorbers of technical knowledge, such as chemical reactions and logical expressions. LLMs, and more broadly Foundation Models, show great potential for super-charging AI Scientists. They can act both as a source of scientific knowledge, since they have read all the scientific literature, and a source of new scientific hypotheses.

One of the current problems with LLMs is their tendency to hallucinate, that is to output statements that are not true. While this is a serious problem in many applications, it is not necessarily so in science, where physical experiments are the arbiters of truth. Hallucinations are hypotheses.

AI has been used as a tool in the research behind tens of thousands of scientific papers. We believe this only a start. We believe that AI has the potential to transform the very process of science.

We believe that by harnessing the power of AI, we can propel humanity toward a future where groundbreaking achievements in science, even achievements worthy of a Nobel Prize, can be fully automated. Such advances could transform science and technology, and provide hope of dealing with the formidable challenges that face humankind in the 21st century

̽��ֱ��Nobel Turing Challenge aims to develop AI Scientists capable of making Nobel-quality scientific discoveries at a level comparable, and possibly superior to the best human scientists by 2050.

As well as being a potential transformative power for good, the application of AI to science has potential for harm. As a step towards preventing this harm, my colleagues and I have prepared the Stockholm Declaration on AI for Science. This commits the signees to the responsible and ethical development of AI for science. A copy of the declaration can be signed on ̽��ֱ��Stockholm Declaration on AI for Science website.

We urge all scientists working with AI to sign.

Professor Ross King from Cambridge's Department of Chemical Engineering and Biotechnology, who originated the idea of a 'Robot Scientist', discusses why he believes that AI-powered scientists could surpass the best human scientists by the middle of the century, but only if artificial intelligence for science is developed responsibly and ethically.

kynny via Getty Images

Robot arm handling test tubes.

̽��ֱ��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

Languages still a major barrier to global science, new research finds

fpjl2 — Thu, 29 Dec 2016 19:04:57 +0000

English is now considered the common language, or 'lingua franca', of global science. All major scientific journals seemingly publish in English, despite the fact that their pages contain research from across the globe.

However, a new study suggests that over a third of new scientific reports are published in languages other than English, which can result in these findings being overlooked - contributing to biases in our understanding.

As well as the international community missing important science, language hinders new findings getting through to practitioners in the field say researchers from the ̽��ֱ�� of Cambridge.

They argue that whenever science is only published in one language, including solely in English, barriers to the transfer of knowledge are created.

̽��ֱ��Cambridge researchers call on scientific journals to publish basic summaries of a study's key findings in multiple languages, and universities and funding bodies to encourage translations as part of their 'outreach' evaluation criteria.

"While we recognise the importance of a lingua franca, and the contribution of English to science, the scientific community should not assume that all important information is published in English," says Dr Tatsuya Amano from Cambridge's Department of Zoology.

"Language barriers continue to impede the global compilation and application of scientific knowledge."

̽��ֱ��researchers point out an imbalance in knowledge transfer in countries where English is not the mother tongue: "much scientific knowledge that has originated there and elsewhere is available only in English and not in their local languages."

This is a particular problem in subjects where both local expertise and implementation is vital - such as environmental sciences.

As part of the study, published today in the journal PLOS Biology, those in charge of Spain's protected natural areas were surveyed. Over half the respondents identified language as an obstacle to using the latest science for habitat management.

̽��ֱ��Cambridge team also conducted a litmus test of language use in science. They surveyed the web platform Google Scholar - one of the largest public repositories of scientific documents - in a total of 16 languages for studies relating to biodiversity conservation published during a single year, 2014.

Of the over 75,000 documents, including journal articles, books and theses, some 35.6% were not in English. Of these, the majority was in Spanish (12.6%) or Portuguese (10.3%). Simplified Chinese made up 6%, and 3% were in French.

̽��ֱ��researchers also found thousands of newly published conservation science documents in other languages, including several hundred each in Italian, German, Japanese, Korean and Swedish.

Random sampling showed that, on average, only around half of non-English documents also included titles or abstracts in English. This means that around 13,000 documents on conservation science published in 2014 are unsearchable using English keywords.

This can result in sweeps of current scientific knowledge - known as 'systematic reviews' - being biased towards evidence published in English, say the researchers. This, in turn, may lead to over-representation of results considered positive or 'statistically significant', and these are more likely to appear in English language journals deemed 'high-impact'.

In addition, information on areas specific to countries where English is not the mother tongue can be overlooked when searching only in English.

For environmental science, this means important knowledge relating to local species, habitats and ecosystems - but also applies to diseases and medical sciences. For example, documents reporting the infection of pigs with avian flu in China initially went unnoticed by international communities, including the WHO and the UN, due to publication in Chinese-language journals.

"Scientific knowledge generated in the field by non-native English speakers is inevitably under-represented, particularly in the dominant English-language academic journals. This potentially renders local and indigenous knowledge unavailable in English," says lead author Amano.

" ̽��ֱ��real problem of language barriers in science is that few people have tried to solve it. Native English speakers tend to assume that all the important information is available in English. But this is not true, as we show in our study.

"On the other hand, non-native English speakers, like myself, tend to think carrying out research in English is the first priority, often ending up ignoring non-English science and its communication.

"I believe the scientific community needs to start seriously tackling this issue."

Amano and colleagues say that, when conducting systematic reviews or developing databases at a global scale, speakers of a wide range of languages should be included in the discussion: "at least Spanish, Portuguese, Chinese and French, which, in theory, cover the vast majority of non-English scientific documents."

̽��ֱ��website conservationevidence.com, a repository for conservation science developed at Cambridge by some of the authors, has also established an international panel to extract the best non-English language papers, including Portuguese, Spanish and Chinese.

"Journals, funders, authors and institutions should be encouraged to supply translations of a summary of a scientific publication - regardless of the language it is originally published in," says Amano. ̽��ֱ��authors of the new study have provided a summary in Spanish, Portuguese, Chinese and French as well as Japanese.

"While outreach activities have recently been advocated in science, it is rare for such activities to involve communication across language barriers."

̽��ֱ��researchers suggest efforts to translate should be evaluated in a similar way to other outreach activities such as public engagement, particularly if the science covers issues at a global scale or regions where English is not the mother tongue.

Adds Amano: "We should see this as an opportunity as well as a challenge. Overcoming language barriers can help us achieve less biased knowledge and enhance the application of science globally."

Over a third of new conservation science documents published annually are in non-English languages, despite assumption of English as scientific ‘lingua franca’. Researchers find examples of important science missed at international level, and practitioners struggling to access new knowledge, as a result of language barriers.

Language barriers continue to impede the global compilation and application of scientific knowledge

Tatsuya Amano

Quinn Dombrowski

Atypical welcome

̽��ֱ��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-ShareAlike

Twenty top tips for interpreting scientific claims

sj387 — Thu, 21 Nov 2013 07:58:54 +0000

Scientists from the UK and Australia, concerned with the lack of scientific knowledge amongst key decision makers, have created 20 concepts to help those who interact regularly with science and scientists.

Recent issues such as nuclear power, bee declines, and the role of badgers in bovine tuberculosis have seen fierce debates and policy decisions being made without the support of the scientific community, something Professors William Sutherland, David Spiegelhalter and Mark Burgman have set out to change.

These scientists want to help people grasp the “imperfect nature of science” and enable policy-makers to interrogate their advisers and experts instead of simply accepting information as it is given. Though change will take time, it is their belief that “a wider understanding of these 20 concepts by society would be a marked step forward”, and could only lead to a better-informed future.

Professor Spiegelhalter said “These tips could be used as a checklist when confronted with scientific claims. Science is not just a body of facts – it’s important to have a grasp of the process by which conclusions are drawn, and the possible pitfalls on that path”

There is an obvious need to make sure that scientific policy is based on a sound understanding of science; this means making sure that policy-makers know the right things to ask, and how to interpret the answers they get. There have been many suggestions of how to increase the level of scientific knowledge in the political community, from encouraging more scientists to become politicians, to expanding the role of chief scientific advisors. However, none of these solutions fully address the fundamental issue of widespread “scientific ignorance” amongst those who have the ability to vote in parliament.

̽��ֱ��homogeneity of policy makers’ backgrounds shows just how far-reaching this problem is. No member of the current British cabinet has a scientific degree – the closest is Vince Cable, who initially read Natural Sciences at Cambridge before switching to Economics. Six cabinet members read Politics, Philosophy and Economics (PPE) at Oxford, including Ed Davey MP, the Secretary of State for Energy and Climate change. Of Britain’s 650 members of parliament, only Cambridgeshire MP Julian Huppert is a scientist; David Willets MP, the Minister of State for Universities and Science, read PPE.

This list of concepts will teach skills closely related to those that politicians already have, and will help people “understand the quality, limitations and biases of evidence”. This will, in turn, allow better interrogation of those communicating scientific information. By explaining the scientific process, these academics have helped to demystify science and make it accessible to those creating the country’s scientific policy. It is not a question of turning every policy-maker into a scientist, but of arming them with the tools to understand and question the scientific information they receive.

Some of the concepts seem common sense (“Scientists are human”), others less so (“Regression to the mean can mislead”, “Beware the base rate fallacy”). All contain practical advice and recommendations that, if followed, should help policy-makers better interact with science and scientists and understand the limitations of evidence. Though the authors acknowledge that improvements in policy will not happen instantaneously, and that uncertainty is inherent in the scientific method, they nonetheless feel that these concepts are the first step to take if we are to more closely integrate science into political decision-making.

Click here to read the Nature article in full

Aiming to improve policy-makers’ understanding of the imperfect nature of science, academics from the Universities of Cambridge and Melbourne have created a list of concepts that they believe should be part of the education of civil servants, politicians, policy advisers and journalists

Science is not just a body of facts – it’s important to have a grasp of the process by which conclusions are drawn, and the possible pitfalls on that path

Professor David Spiegelhalter

Sir Cam/ ̽��ֱ�� of Cambridge

Dr Pete Wothers giving a chemistry demonstration to an audience at the Cambridge Science Festival including David Willetts, Minister for Universities and Science.

This work is licensed under a Creative Commons Licence. If you use this content on your site please link back to this page.

Yes