UCL's Academic Centre of Excellence in Cyber Security Research presents: 2022 Open Day

About this event

Since 2012, UCL has been recognised by EPSRC/UKRI and GCHQ/NCSC as a Centre of Excellence in Cyber Security Research (ACE-CSR).

After more than two years, we are finally ready to welcome all our partners and collaborators back in person at UCL. We will be officially welcoming the latest researcher to join our Centre, Professor Lorenzo Cavallaro, and host an afternoon of talks and discussions, followed by a drinks reception.

Prospects and Pitfalls for Machine Learning in Software Security

Large language models (LLMs) can now write code with reasonable fluency and functionality, even matching human performance in competitive programming competitions. What implications does this new capability hold for software security?

In this keynote, we will examine this question from two angles:

First, we will look at the security of code produced by large language models, and whether programmers using such models tend to write more or less secure code.

Second, we will look at several ways in which the coding capabilities of LLMs might be employed to help solve difficult problems in software security including reverse engineering, vulnerability repair, and bug finding. Along the way, we'll discuss some traps and pitfalls that have made some prior research in this area unreliable, and suggest ways to mitigate these issues in future research.

This is a hybrid event taking place at Goodenough College and will also be livestreamed. Please select the correct ticket for your mode of attendance.

The link to the livestream is included in the registration confirmation email. A reminder will be sent closer to the event.

Panellists

Brendan Dolan-Gavitt

Brendan Dolan-Gavitt is an Assistant Professor in the Computer Science and Engineering Department at the NYU Tandon School of Engineering, and is a member of the NYU Center for Cybersecurity (CCS).

He leads the MESS Lab, which focuses on software security, reverse engineering, embedded systems, and machine learning security. He is the recipient of a 2022 NSF CAREER award, and can be found posting bad jokes as @moyix on Twitter.

Fred Langford

Fred is Director of Trust & Safety Technology at UK Communications Regulator Ofcom joining in November 2020 and is a highly experienced technology leader. His work predominantly focuses on directing Ofcom's technical requirements in relation to the new UK Online Safety regulatory regime.

Prior to joining Ofcom Fred was Deputy CEO and Chief Technology Officer at Internet Watch Foundation (IWF) and President of INHOPE (International CSEA Hotline Association). Fred has 30 years’ experience across a variety of sectors (Commercial, CSO’s, Military and Regulatory) primarily focusing on the internet, its technology, governance, security, safety and regulation.

Fred is a Board Member of Video Standards Council which sets PEGI ratings in the UK, Former Chair of the UK Council for Internet Safety, Technical Working Group (UKCIS TWG), Board Member of the UK Home Office Child Abuse Image Database Strategic Group, Member of the National Crime Agency (NCA) Prevent Strategic Board, a founding Director of the UK Safer Internet Centre and is an expert advisor to UK and other Governments, Parliamentarians, The Commonwealth, Police and NGOs.

Jen Persson

Jen Persson is Director and founder of the NGO Defend Digital Me that campaigns for children’s privacy in the field of UK education and wider public sector. She is a current contributor to the Council of Europe digital citizenship working group on Artificial Intelligence (AI) in Education.

Jen supported the Committee of Convention 108 to draft Council of Europe Guidelines for Data Protection in Education Settings adopted in 2020. Defend Digital Me published a report mapping education data in England including 'safeguarding-in-schools' technology in the report The State of Data 2020.

Their most recent report is a snapshot of the application and growing uses of biometric data, The State of Biometrics 2022: A review of policy and practice in UK Education. @TheABB https://defenddigitalme.org/

Richard Pursey

Richard is a founding director of the Online Safety Tech Industry Association, the international industry body for organisations operating in the field of online safety. He is also founder and Chairman of SafeToNet, a British safety tech company that safeguards children online in over 130 countries around the world.

Richard is the pioneer behind SafeToWatch, AI software that detects and filters harmful content in live-stream video and is an authority in using technology to balance a child’s privacy rights with the need to keep them safe.

Richard sits on the United Nations advisory Board along with the Interregional Crime and Justice Research Institute (UNICRI) for using AI to make children safer online. As a previous Non-executive Director of a National Health Service Trust, Richard also has first-hand knowledge of the mental health issues and societal problems surrounding child abuse.