We’re releasing eight simulated robotics environments and a Baselines implementation of Hindsight Experience Replay, all developed for our research over the past year. We’ve used these environments to train models which work on physical robots. We’re also releasing a set of requests for robotics research.
Come to OpenAI’s office in San Francisco’s Mission District for talks and a hackathon on Saturday, March 3rd.
We’re excited to welcome new donors to OpenAI.
We’ve co-authored a paper that forecasts how malicious actors could misuse AI technology, and potential ways we can prevent and mitigate these threats. This paper is the outcome of almost a year of sustained work with our colleagues at the Future of Humanity Institute, the Centre for the Study of Existential Risk, the Center for a New American Security, the Electronic Frontier Foundation, and othe
We’ve designed a method that encourages AIs to teach each other with examples that also make sense to humans. Our approach automatically selects the most informative examples to teach a concept—for instance, the best images to describe the concept of dogs—and experimentally we found our approach to be effective at teaching both AIs
Around 200 talented young people arrived at the University over two weekends in January and February to take part in the finals of a computational challenge held in the department.
We’ve built a system for automatically figuring out which object is meant by a word by having a neural network decide if the word belongs to each of about 100 automatically-discovered “types” (non-exclusive categories).
A new 3-minute video celebrating 60 years of impressive Computer Science by our department has been put together by the University.
To many, the Internet of Things (IoT) represents a new advancement in technology meant to streamline our lives even further but are there significant new security and privacy risks to using these forms of new technology?
Two of the department’s professors – Ursula Martin and Blanca Rodriguez – have been chosen to receive Impact Awards by the Mathematical, Physical, Engineering and Life Sciences (MPLS) Division at the University of Oxford.
After three years of collaboration between industry and academia and €4 million of funding from the European Commission's Horizon 2020 program, the ALIGNED project has successfully rolled out five trial platforms showcasing its success in applying the latest big-data research findings to solving the problems of industry.
We’re releasing a new batch of seven unsolved problems which have come up in the course of our research at OpenAI.
The Oxford Mathematics Institute and the British Society for the History of Mathematics will host a session on the ‘History of Computing beyond the Computer’ on 21-22 March.
There have been reports in the press this week of how the examination length for students taking examinations in the Department of Computer Science at the University of Oxford was extended in summer 2017. We would like to emphasise that the extension was applied to all students taking those examinations and was for academic reasons.
A team from the department led by Professor Cas Cremers has been working with Facebook software engineer Jon Millican to improve the security of group messaging.
A research team from the Department of Computer Science and the Department of Engineering has won ‘the outstanding student paper’ at one of the premier venues for publishing AI research: AAAI-18, the Association for the Advancement of Artificial Intelligence’s conference.
‘Privacy-preserving analytics in, or out of, the cloud’ is the title of the next Strachey Lecture, to be delivered by Professor Jon Crowcroft from the University of Cambridge, on 6 March
The fifth Oxbridge Women in Computer Science Conference will be held on 15 March 2018 in Cambridge, UK.
TheySay, a sentiment analysis company spun out from the Department of Computer Science, has been acquired by Aptean, a global provider of mission-critical enterprise software solutions.
Applications have opened for places on the University of Oxford’s free UNIQ summer school. Twenty young people will be chosen to participate in the Computer Science stream from 15 to 20 July 2018, alongside more than 30 other subjects on offer from the University.
The news media’s role in the flow of news content on Twitter during the 2017 UK General Election has been analysed by a team from the Reuters Institute for the Study of Journalism (RISJ) and the Department of Computer Science.
The winter 2017 edition of Inspired Research, the newsletter covering the department's latest news in research, teaching and social activities, is now online.
The University of Oxford is to showcase artificial intelligence (AI) research across the whole University with a one-day expo on 27 March 2018.
To help speed up code writing, many app developers use common libraries, but this can result in such libraries getting access to significant privileges on a device. DPhil student Vincent Taylor describes the problem and his research in the area
A new book on the code of ethics by Senior Researcher Paula Boddington has just been published.
The podcast of Professor Bjarne Stroustrup of Columbia University’s Strachey Lecture on ‘The Continuing Evolution of C++’ is now available on our website.
Embedded software in the transportation sector (railway, automotive and avionics) needs to meet high reliability requirements because errors may have severe consequences. Research since 2008 by Daniel Kroening’s group at Oxford has developed effective reasoning technology to provide assurance that key error types are eliminated from embedded software, and has created novel algorithms to prove its
A Marie Claire article about the potential risks from voice-activated smart speakers quotes the views of our Research Fellow Jason R.C. Nurse.
Professor Marina Jirotka and Senior Researcher Helena Webb consider what can be done by government and social media platforms to tackle the problem of fake news
Professor Michael Wooldridge describes his experience giving evidence to the House of Lords Select Committee on Artificial Intelligence
How people relate to intelligent machines is one of the complicated questions to solve in work to develop ethical codes for Artificial Intelligence (AI) research
The safety of self-driving cars relies on the accuracy of the deep neural networks that control them. To verify the performance of these networks, the department is developing model checking techniques
An important step in making sure that robots act safely and effectively in our society is to be able to formally express what trust is in a human-robot partnership
A clearer understanding of badgers' territorial habits give vital clues to the likely effectiveness of badger culls
A painting of alumna Anne-Marie Imafidon (2006, Mathematics and Computer Science) is among a new range of portraits that went on display at an exhibition in Oxford’s Weston Library on 24 November.
We’re releasing highly-optimized GPU kernels for an underexplored class of neural network architectures: networks with block-sparse weights. Depending on the chosen sparsity, these kernels can run orders of magnitude faster than cuBLAS or cuSPARSE. We’ve used them to attain state-of-the-art results in text sentiment analysis and generative modeling of text and images.
Professor Marina Jirotka was one of the contributors to a major report by academics that is calling for a digital environment that is ‘fit for childhood’
Jolie completed both her BA and DPhil at the Oxford University Computer Laboratory (the department’s former name). Her time armed her with the core skills for her current role in systematic trading in the quantitative finance industry
‘Where have all the cicadas gone?’ is the question addressed by Professor Alex Rogers in the latest podcast released by Oxford Sparks.
Three teams from the Department of Computer Science showed off their skills in a programming teamwork challenge on 24 November.
Professor Ursula Martin, Senior Research Fellow in Computer Science, was conferred with an honorary degree by the University of London’s Chancellor, Her Royal Highness The Princess Royal, in recognition of achievements in mathematics and computer science.
DPhil student Klaudia Krawiecka has won a national competition in Finland for the best information security thesis. She was the first woman to receive the award, which was presented at a ceremony on 24 October in Helsinki.
DPhil student Maarten Scholl won the Best Quantitative Finance MSc Thesis Award for his thesis entitled ‘The Role of Central Clearing Parties in Over-the-Counter Derivatives Markets’ on 3 November.
A free webinar on 22 November will provide insights into the latest research on high performance computational (HPC) simulations of cardiac electrophysiology using patient specific models of the whole heart.
Computer Science at Oxford came first in the 2018 UK university subject league tables compiled by the Complete University Guide.
Researchers from our department showcased their work in relation to Internet innovation, security and freedom at this year’s MozFest, an annual event that celebrates the open Internet movement
EPSRC has awarded Tim Muller of our security research group an EPSRC First Grant to help with the battle against fake reviews online.
Members of our Department have co-authored five papers that have been accepted by the top conference in programming languages, The ACM SIGPLAN Symposium on Principles of Programming Languages (POPL).
Graduate student Daniel Nichol, alongside four students from the Department of Mathematics, attended the 5th Heidelberg Laureate Forum (HLF) in late September.
Oxford entered nine teams of students in UKIEPC, a contest testing programming skills, problem solving and teamwork.
We’ve developed a hierarchical reinforcement learning algorithm that learns high-level actions useful for solving a range of tasks, allowing fast solving of tasks requiring thousands of timesteps. Our algorithm, when applied to a set of navigation problems, discovers a set of high-level actions for walking and crawling in different directions, which enables the agent to master new navigation tasks
Professor Boris Motik has been awarded a three-year fellowship extension by EPSRC for his work on intelligent information systems.
The final workshop of the EPSRC-funded project ‘Social Machines of Mathematics’ will focus on enabling mathematical cultures on 5-7 December at the Mathematics Institute
Our latest robotics techniques allow robot controllers, trained entirely in simulation and deployed on physical robots, to react to unplanned changes in the environment as they solve simple tasks. That is, we’ve used these techniques to build closed-loop systems rather than open-loop ones as before.
The Careers in Computing Fair will be held on Tuesday 24 October from 3-6pm in the Robert Hooke Building on Parks Road.
Computer Science at the University of Oxford has again topped the Times Higher Education rankings by subject for the UK.
The BBC has published a feature article describing the potential of quantum computing to transform the internet, with comments from Senior Research Fellow Jamie Vicary.
DPhil student Artem Kaznatcheev's blog has been nominated by Science Borealis as one of Canada's top 12 science blogs.
We show that for the task of simulated robot wrestling, a meta-learning agent can learn to quickly defeat a stronger non-meta-learning agent, and also show that the meta-learning agent can adapt to physical malfunction.
We’ve found that self-play allows simulated AIs to discover physical skills like tackling, ducking, faking, kicking, catching, and diving for the ball, without explicitly designing an environment with these skills in mind. Self-play ensures that the environment is always the right difficulty for an AI to improve. Taken alongside our Dota 2 self-play results, we have increasing confidence that self
Mariam Nouh has won third place in the ACM Student Research Competition (Graduate Category) at the Grace Hopper Conference in Orlando.
Martin Strohmeier’s DPhil dissertation titled ‘Security in Next Generation Air Traffic Communication Networks’ has received a commendation and been published on the BCS website.
Members of the Department’s Human Centred Computing group (HCC) took part in the Curiosity Carnival on 29 September.
We’re releasing an algorithm which accounts for the fact that other agents are learning too, and discovers self-interested yet collaborative strategies like tit-for-tat in the iterated prisoner’s dilemma. This algorithm, Learning with Opponent-Learning Awareness (LOLA), is a small step towards agents that model other minds.
We’re releasing two new OpenAI Baselines implementations: ACKTR and A2C. A2C is a synchronous, deterministic variant of Asynchronous Advantage Actor Critic (A3C) which we’ve found gives equal performance. ACKTR is a more sample-efficient reinforcement learning algorithm than TRPO and A2C, and requires only slightly more computation than A2C per update.
Our Dota 2 result shows that self-play can catapult the performance of machine learning systems from far below human level to superhuman, given sufficient compute. In the span of a month, our system went from barely matching a high-ranked player to beating the top pros and has continued to improve since then. Supervised deep learning systems can only be as good as their training datasets, but in s
Some freely available tools to detect threats to mobile device users from fake mobile phone towers, also called stingrays and International Mobile Subscriber Identity (IMSI) catchers, have been shown to have major shortcomings in a recently presented paper
RL-Teacher is an open-source implementation of our interface to train AIs via occasional human feedback rather than hand-crafted reward functions. The underlying technique was developed as a step towards safe AI systems, but also applies to reinforcement learning problems with rewards that are hard to specify.
9551 articles sourced historically · 100 per page