←
Science Works265 posts
Science Works avatar

Science Works

@sci_works

A policy, research, & software studio built to accelerate progress in British science and technology.

◷ Joined August 2022▧ 17 Photos and videos
Science Works avatar
Science Works@sci_works·Oct 29

We have a new Data Centre Covenant. But, until now, there was no detailed overview of what data centre capacity the UK plans to build: how much, where, and when? What signals are there about demand for UK compute? Our new report has the answers – which may surprise you! 🧵👇

0:06
Bia avatar
⟳ Science Works retweeted
Bia@BeatrizGietner·3h

I think the whole team at @sci_works is doing an amazing job when it comes to demonstrate and share the work we do in different ways, as @Lyan82 writes in the first edition of SW’s newsletter - which you can subscribe to at the bottom of the page here :) science.works

Post attachment

Researchers are finding new ways to share their work beyond the PDF: games, interactive explainers, digital editions, living papers, dataset sites. AI has made building a website almost free, and people are getting creative. Some favorites 🧵

Bia avatar
⟳ Science Works retweeted
Bia@BeatrizGietner·Nov 17

Did anyone say "mechanism design"? I'm working on it 🫡

Post attachment

I am sharing a new paper today with Dario Caldara: “Can Peer Review Survive AI-Generated Text?” sas.upenn.edu/~jesusfv/Peer_… Dario and I, as editors, are increasingly getting papers submitted to our journals that are 99 percent AI-generated. Even worse, we are also receiving many AI-generated referee reports, both as editors and, more frustratingly, as authors. And we cannot emphasize this enough: the real problem is not that the papers and reports were AI-generated; it is that no one had done the thinking behind them, and you can so easily tell. The papers were eminently forgettable, and the referee reports were disappointingly uninsightful. We have seen too many reports of the form: “The paper is well-written and addresses an important problem. However, the authors should consider additional baselines.” This deluge of AI-generated text puts peer review in jeopardy: journals will be flooded with low-quality submissions, economists will decline requests to referee papers or routinely use AI to referee them, and which papers get published will become less and less linked to their quality. We do not want to defend the status quo entirely. Peer review is slow, many reports are poor, and editors (including us!) make bad calls all the time. Still, well-executed peer review remains the best system we have for assessing each other’s work and advancing the profession. AI could improve peer review if used deliberately for that purpose, for instance, to check proofs and code. That is not what we are seeing. The difficulty is that readers, including editors and referees, cannot see how AI was used. As a result, in practice, the burden of verification shifts from the author to the reader. Since we reject this shift, Dario and I did what economists always do when something bothers us: we wrote a paper. So, what do we do in it? First, we check whether AI detectors such as Pangram get this right: is a referee report flagged as AI-generated indeed AI-generated? Or is Pangram just flagging proofreading and editing? No. We run an experiment with 25 referee reports from the pre-AI era. Proofreading and editing do not trigger AI-generated flags. Only when we ask the LLMs to substantially rewrite the referee report or to generate a new report do we get the AI-generated flags. Pretty compelling evidence, if I may say so. Second, we build a theoretical model to explore who should bear the burden of verification: the author or the reader? (Yours truly always enjoyed the courses on the rules of evidence in law school; yes, as Billy Wilder would say, nobody is perfect.) The answer is that it is much more efficient for the author to do it, because the author can verify at a much lower cost and only needs to do it once (rather than each reader repeatedly). Before AI, the authors more or less had to verify the results, even if only because they had to type them in. This was not perfect (and all papers ever published in economics probably have at least one mistake). But the problem is that, with AI, the author does not even need to read the text. Or can just skim it. Writing is thinking, and without writing there is usually not much thinking. So, a paper flagged as AI-generated is prima facie evidence that the author has not verified it. Should this lead to an automatic rejection? No. But there should be a zero-tolerance policy for serious mistakes. The first serious mistake we find in the paper, and it is out! In comparison, for a non-flagged paper, a serious mistake is something that the editor might determine it can be fixed in a revision. We are stricter with referee reports. There is absolutely no point in the referee generating a report using an LLM. As an editor, I can do that directly and save myself the trouble of asking you for it. Now, of course, you can use the LLM to help you with the report (if the journal allows it! Many journals do not allow you to upload the paper to an LLM, and you need to follow the rules). For example, you can use the LLM to search for existing results in the literature. But peer review is not about finding one sentence that is not entirely precise or a slightly sloppy mathematical proof. Referees are asked: “Is this idea important despite its flaws?” And, relatedly: “Is the paper a good fit for the journal in terms of quality and focus?” That judgment is the task for which you are selected as a subject expert. We also propose that journals require that, before a paper is submitted, the author submit it to an LLM using an approved prompt and include the LLM report and a discussion of what was and was not fixed. That way, the referee would not need to check (or at least not check so carefully) every technical detail, and they could focus on the big picture: Is this a good paper? Is this a good fit for this journal? What is fundamental for this paper to become publishable at this journal? Anyway. The paper has many other arguments, and we try to be nuanced. So, before ferociously typing the first complaint that crosses your mind, would you consider reading it? Chances are that we have already discussed your objection. We have been thinking about this for a long time. But by all means, leave good comments!

Post attachment
Alvin Djajadikerta avatar
⟳ Science Works retweeted
Alvin Djajadikerta@adjajadikerta·Nov 17

Warren Weaver and the Rockefeller foundation helped invent molecular biology. What are today’s philanthropic organisations inventing? Or - what do we need in place so that they CAN invent something? If you’re interested in this question space, this job might be for you…

Reminder this closes tomorrow 📨 The Medici, Henry Wellcome, John Rockefeller, Andrew Carnegie, Bill & Melinda Gates – philanthropy & patronage have shaped the history of science. What kinds of science could the 3rd wave of philanthropy make possible? Join & help us find out

Laura Ryan avatar
⟳ Science Works retweeted
Laura Ryan@Lyan82·Nov 17

Reminder this closes tomorrow 📨 The Medici, Henry Wellcome, John Rockefeller, Andrew Carnegie, Bill & Melinda Gates – philanthropy & patronage have shaped the history of science. What kinds of science could the 3rd wave of philanthropy make possible? Join & help us find out

🚨 come work with us! 🚨 We're looking for a Research Associate (London, UK): someone curious and diligent to investigate how philanthropy is allocated in science, and how it’s changing the way that science is organised and funded. Apply here by the 18th November science.works/work-with-us/r…

Dr Anton Howes avatar
⟳ Science Works retweeted
Dr Anton Howes@antonhowes·Nov 17

So many scientists are pushing for the UK to enter the EU's next research programme. But it will mean HUGE cuts to science funding, of £1.2–1.5 billion per year - ARIA’s entire budget and all of UKRI’s applicant-led research funding put together! science.works/writing/what-w…

Science Works avatar
⟳ Science Works retweeted
Science Works@sci_works·Nov 16

UK science sector bodies are pushing for Britain to join the EU’s next research programme (FP10) in full. But we've found there’s a £7.7 billion gap between the likely bill and what the government has budgeted for. This would mean devastating cuts to domestic research. 🧵

Will Stone avatar
⟳ Science Works retweeted
Will Stone@willecstone·Nov 16

Being fully part of the next research programme is a sacred cow. But it'll force government to make deep cuts to other research, and we need to reckon with that as soon as possible, not just wish the problem away.

UK science sector bodies are pushing for Britain to join the EU’s next research programme (FP10) in full. But we've found there’s a £7.7 billion gap between the likely bill and what the government has budgeted for. This would mean devastating cuts to domestic research. 🧵

Ben Johnson avatar
⟳ Science Works retweeted
Ben Johnson@ersatzben·Nov 16

Joining the EU's next research programme in full could leave a £7.7 billion hole in the UK science budget, new research by @sci_works finds. The government has budgeted for a programme roughly the size of today's Horizon Europe scheme. But the proposals being discussed and debated by the EU would result in a programme that's 65-109% bigger. A bigger programme means a bigger joining fee for the UK. In our central scenario (far from the most expensive), the UK would need to find an extra £1.2 billion in 2029-30. This is equivalent to three ARIAs, 1.4x UKRI’s entire response-mode grant portfolio, or 55% of QR.  There are no good ways to find this money. But if we leave the decision late, we’ll face little choice but to cancel live grant contracts or make savage cuts to QR – the block funding that's vital for financial sustainability of the research base. Our view: the UK should open negotiations by seeking association to the Excellent Science pillar only, where the UK has historically done best. This more targeted form of association would save us approx £2.4 billion a year. But if it has to be full-fat association, the UK should get improved terms (at least on a par with other countries), and decide now where the money is going to come from rather than leaving it until it's too late. International collaboration in research is a very good thing. The European programmes bring real benefits. But being publicly desperate to join a supersized programme at any price is a poor way to open a negotiation. Full report below from @willecstone and me:

UK science sector bodies are pushing for Britain to join the EU’s next research programme (FP10) in full. But we've found there’s a £7.7 billion gap between the likely bill and what the government has budgeted for. This would mean devastating cuts to domestic research. 🧵

Jack Aspinall avatar
⟳ Science Works retweeted
Jack Aspinall@jackaspi·Nov 16

Important piece of work. The EU have been a great funder of UK science, but it seems joining FP10 would necessitite meaningful cuts to domestic science. We mustn't sleepwalk into this!

UK science sector bodies are pushing for Britain to join the EU’s next research programme (FP10) in full. But we've found there’s a £7.7 billion gap between the likely bill and what the government has budgeted for. This would mean devastating cuts to domestic research. 🧵

Science Works avatar
Science Works@sci_works·Nov 16

"Government needs a plan for either finding the extra money or rethinking the terms of our involvement. Otherwise, we could face a nightmare scenario where we are forced to endure devastating cuts to UK universities and facilities that would permanently scar British science."

Multibillion-pound hole in the projected cost of joining the EU’s FP10 research programme should make country rethink whether full association is worth it, say researchers. @TWilliamsTHE reports [Link in comments]

Post attachment
Science Works avatar
Science Works@sci_works·Nov 16

What we’d do? Open negotiations by seeking the Excellent Science pillar only, where we do best. And secure better clawback terms (parity with other countries). Consider & plan for the costs now. Being publicly desperate to join at any price is a poor way to open a negotiation.

Science Works avatar
Science Works@sci_works·Nov 16

Some scenarios could be highly at odds with the government’s regional agenda. If we leave it late and have to fund association from block grants, this means redistributing money away from the North of England.

Post attachment
Science Works avatar
Science Works@sci_works·Nov 16

Not dealing with this up front makes the problem worse, as more £££ gets tied up. It’ll be tough to avoid cancelling live grants or deep cuts to block funding for unis. And – what we get back lags what we put in. By 2034-35, only 43p in the £ will have reached UK researchers.

Post attachment