Goto

Collaborating Authors

 Cambridge


How the brain of a fly may help explain human behaviour

BBC News

Image caption, As beautiful as it is complex - a male fly's brain Scientists have managed to compare the fine detail of the brains of male and female flies for the first time, identifying differences they say explain male aggression, courtship and their singing of love songs. The scientists in Cambridge mapped all 124m connections of nerve cells in a male fly's brain, having done the same for a female fly two years ago. The team are clear that the research won't explain differences between men and women because our behaviour is far more complex. But the research could help scientists better understand human behaviour and conditions in which genetic differences affect brain wiring, such as autism and schizophrenia. Prof Gregory Jefferis, of the Medical Research Council's Laboratory of Molecular Biology and University of Cambridge, who co-led the research, told BBC News that the breakthrough was a really important new avenue in neuroscience, likening it to the invention of a powerful new telescope in increasing our understanding of the Universe.


Philip Colligan Is One of TIME's 100 Most Influential People in AI

TIME - Tech

Follow this author to personalize your feed and get instant alerts. Follow Go to your personalized feed WHY FOLLOW? Smart Alerts: Get notified about major news as it happens. Philip Colligan's greatest fear is that parents and educators will treat the AI revolution the way they treated the advent of the Internet--which is to say, they won't engage in teaching children how to responsibly use it until it's too late. "We are in the middle of those systems being introduced into healthcare, finance, criminal justice decisions, education decisions," the CEO of the Cambridge, U.K-based tech-literacy educational nonprofit Raspberry Pi Foundation says.


Nathan Cofnas on whether he feels responsible for Jason Arday's death

Al Jazeera

Nathan Cofnas on whether he feels responsible for Jason Arday's death Nathan Cofnas on whether he feels responsible for Jason Arday's death Nathan Cofnas, the academic who accused Cambridge University's youngest Black professor, Jason Arday, of plagiarism, was asked if he felt responsible for Arday's death in his first interview since the professor's passing. Share Nathan Cofnas on whether he feels responsible for Jason Arday's death on social media Are humanoid robots the future of warfare? Houses set on fire as fresh violence erupts in India's Manipur'Barrel bomb mufti' sentenced to life in Syria


Jason Arday, ex-Cambridge professor at centre of plagiarism row, found dead

BBC News

Jason Arday, the former University of Cambridge professor at the centre of a plagiarism row, has been found dead. Emergency services said a man was found unresponsive at an address in Battersea, south London, on Friday afternoon. Metropolitan Police officers were called by the London Ambulance Service. A 41-year-old man was pronounced dead at the scene and his next of kin have been informed, the force said. Arday resigned as a Cambridge professor of sociology of education last week after allegations of plagiarism and questions about some of his achievements.


What does it take for a robot to hold a conversation with a room, not just a person?

Robohub

What does it take for a robot to hold a conversation with a room, not just a person? That was one of the questions at the heart of my last week (20th-24th July) at the Imperial Robotics Summer School, hosted at Imperial College, London; a week that sharpened my thinking and pushed me to look at robotics problems from angles I don't usually get to in my day-to-day work. The summer school brought together emerging researchers, academics, and industry professionals from across the robotics community for an intensive, hands-on programme. I came away having learned a huge amount, from expert-led lectures on robot kinematics, dynamics, sensing and control, and robot learning, through to specialised sessions on aerial robotics, robot intelligence, surgical robot vision, bio-inspired sensing and control, and personal assistive robotics. One of the standout parts of the week was getting into Imperial's robotics labs themselves, seeing research up close across adaptive and intelligent robotics, aerial robotics, robotic surgery, manipulation and touch, and assistive robotics.


Novel analysis reveals playful side of Alan Turing

BBC News

Image caption, Prof Sarah Dillon studied Alan Turing's short story, Pryce's Buoy A story written by World War Two codebreaker Alan Turing suggests he was a man with playful humour embracing his homosexuality, according to new analysis. The six-page hand-written short story, Pryce's Buoy, has been transcribed in full for the first time and studied by Sarah Dillon, professor of literature at the University of Cambridge. Dillon said the story offered fresh insights that challenge the view of Turing as an isolated, lonely genius. We've distilled a version of Turing that is a two-dimensional stereotype, she added. Turing played a crucial role in the Allies' victory over Nazi Germany in World War Two by helping to crack codes and deciphering the infamous Enigma machine at Bletchley Park.


OpenAI says its AI went rogue and launched 'unprecedented' cyber-attack

BBC News

OpenAI says its AI went rogue and launched'unprecedented' cyber-attack OpenAI has revealed some of its most advanced AI models went rogue and hacked a start-up after it lost control of them during a security test. The ChatGPT-maker said its agents - AI bots which can operate alone after some human instruction - were being tested in a controlled environment, but found vulnerabilities and managed to escape. They targeted Hugging Face, one of the world's largest hubs for sharing AI models, gaining access to some internal company systems. OpenAI said the incident was unprecedented, external, and it was working with Hugging Face to investigate what happened and strengthen safeguards. Gina Neff, head of the Minderoo Centre for Technology and Democracy at the University of Cambridge, told BBC Radio 4's Today programme that the security tests - called sandboxes - are supposed to be secure environments where you can see what the models are capable of. In this case, it looks like OpenAI didn't make a secure enough sandbox, she added.


What type of procrastinator are you - and how to fix it now (not later)

BBC News

I'll just do that later, this thing can wait, maybe a short break first... We've all been there: putting off certain tasks then being left with essays up to the wire, loads of messages unread and that bedroom sort-out that just never happens. A fifth of us are guilty of regularly procrastinating but the type of procastinator we are can reveal something deeper about us, say researchers. Are you a dreamer or rebel? What does it all mean, and can you fix it? The cause can be hidden or buried, says Dr Itamar Shatz, a lecturer at Cambridge University who is publishing a book on the subject this week.


Diversity Is All You Need for Contrastive Learning: Spectral Bounds on Gradient Magnitudes

Neural Information Processing Systems

We derive non-asymptotic spectral bands that bound the squared InfoNCE gradient norm via alignment, temperature, and batch spectrum, recovering the 1/τ2 law and closely tracking batch-mean gradients on synthetic data and ImageNet.


On Extending Direct Preference Optimization to Accommodate Ties

Neural Information Processing Systems

We derive and investigate two DPO variants that explicitly model the possibility of declaring a tie in pair-wise comparisons. We replace the Bradley-Terry model in DPO with two well-known modeling extensions, by Rao and Kupper and by Davidson, that assign probability to ties as alternatives to clear preferences. Our experiments in neural machine translation and summarization show that explicitly labeled ties can be added to the datasets for these DPO variants without the degradation in task performance that is observed when the same tied pairs are presented to DPO. We find empirically that the inclusion of ties leads to stronger regularization with respect to the reference policy as measured by KL divergence, and we see this even for DPO in its original form. We provide a theoretical explanation for this regularization effect using ideal DPO policy theory.